US20020176009A1

Movatterモバイル変換

Info

Publication number: US20020176009A1
Application number: US10/107,892
Authority: US
Inventors: Sandra Johnson; Shih-Chung Chao; Nadi Itani; Caiyi Wang; Brannon Harris; Ash Prabala; Douglas Holberg; Alan Hansford; Syed Azim; David Welland
Original assignee: Individual
Current assignee: Individual
Priority date: 1998-05-08
Filing date: 2002-03-27
Publication date: 2002-11-28
Also published as: US20060197847A1; US7289145B2

Abstract

An image processor system for a charge coupled device (CCD) or CMOS imaging system includes a correlated double sample and variable gain (CDSVGA) circuit for receiving data from a CCD system and an automatic gain control (AGC) circuit which first controls gain by adjusting said CCD system and then for yet a higher gain level makes gain adjustments in said CDSVGA circuit AND a digital gain circuit to produce a combined target gain level. A processing system for an imager device includes a camera system for producing an imager signal, a correlated double sample (CDS) circuit for receiving data from an imager, a variable gain amplifier (VGA), an analog-to-digital converter (ADC) coupled to said CDS circuit, a digital gain circuit (DGC) coupled to said ADC, and an automatic gain control (AGC) circuit coupled to said DGC for controlling the CDS circuit and the DGC. The processing circuitry includes an analog front end and a digital signal processing system for capturing full motion video and outputting a CCIR 601 4:2:2 YCrCb video data output for presentation on a user selected display.

Description

CROSS REFERENCE TO RELATED APPLICATIONS

This application is related to patent application Ser. Nos. ______, ______, ______, ______, and ______, respectively entitled “[0001]Digital Camera Signal Processor and Method” having inventors Syed Khalid Azim, Shih-Chung Chao, Brannon Craig Harris, and Ash Prabala;“High Voltage Input Pad System and Method” having inventors Douglas R. Holberg, Nadi Rafik Itani, and David R. Welland; “Pipelined Analog-to-Digital Converter (ADC) Systems, Methods, and Computer Program Products” having inventors Sandra Marie Johnson and David R. Welland;“Histogram-Based Automatic Gain Control Method and System for Video Applications” having inventors Nadi Rafik Itani, Caiyi Wang, and David R. Welland; and“Selectable Threshold Multimode Gain Control Apparatus and Method for Setting Mutually Continuous Analog, Digital, and Shutter Gain Levels” having inventors Nadi Rafik Itani, Caiyi Wang, and David R. Welland; each of these applications filed on even date herewith and incorporated herein by reference in their entirety.

BACKGROUND OF THE INVENTION

This invention relates to analog and digital processors and methods, and more particularly to processing of full motion video derived from charge coupled device (CCD) cameras and CMOS imagers.[0002]

FIELD OF THE INVENTION

Charge coupled device (CCD) cameras are configured to capture full motion video according to many different CCD output formats and pixel rates. One such CCD camera includes a sensor array, a high voltage CCD driver for driving the sensor array, and a DC-DC converter. Such a camera produces a four-color mosaic CCD output, including a stream of cyan, magenta, yellow, and green color samples in various combinations.[0003]

These color samples are unfortunately not in a format which can directly be interpreted as images by a viewer using a display. In particular, the color samples are output from the CCD display in analog form representative of pixel information. Many conventional displays require a digital input representative of the image content to be displayed. Therefore, there is a need for a processor to convert the four-color mosaic CCD output into digital formatted data to enable user controlled hardware or software data processing leading to production of viewable digital images.[0004]

There is a further need for modularization of analog and digital subsystems that convert analog data into intermediate digital and final display-ready digital forms.[0005]

There is a further need to separately control gain over the analog and digital subsystems.[0006]

SUMMARY OF THE INVENTION

According to the present invention, a processing system for an imager device includes a camera system for producing an imager signal, a correlated double sample (CDS) circuit for receiving data from an imager, a variable gain amplifier (VGA), an analog-to-digital converter (ADC) coupled to said CDS circuit, a digital gain circuit (DGC) coupled to said ADC, and an automatic gain control (AGC) circuit coupled to said DGC for controlling the VGA circuit and said DGC.[0007]

According to one embodiment of the present invention, a correlated double sampler and variable gain amplifier (CDSVGA) circuit for receiving CCD data, includes a first fixed capacitor for receiving CCD data, and a first amplifier connected to the first fixed capacitor for amplifying CCD data. The first amplifier is connected to the first fixed capacitor, a first variable capacitor connected in parallel with said first amplifier, and a first switch connect in parallel with said first variable capacitor. The first switch is clocked at a first clock phase. The CDSVGA according to one embodiment of the present invention further includes a second variable capacitor connected to the first amplifier, a second amplifier connected to the second variable amplifier, a second fixed capacitor connected in parallel with the second amplifier, and a second switch connected in parallel with the second fixed amplifier. The second switch is clocked at a second clock phase.[0008]

According to one embodiment of the invention, a distributed gain control circuit (DGCC) includes an imager signal source, a timing circuit for controlling the production of signals to the imager signal source, an amplifier system for receiving imager signals from the imager signal source, an analog to digital converter connected to the amplifier for receiving an amplified imager signal stream from the amplifier and converting the amplified imager signal stream into digital form, a digital gain circuit connected to said analog to digital converter, and an automatic gain control (AGC) circuit for receiving an output digital level from the digital gain circuit for controlling the gain of the amplifier system and the digital gain circuit subject to a predetermined gain function (PGF).[0009]

According to one embodiment of the present invention, a signal processing system (SPS) on an integrated substrate for a video camera includes analog front-end (AFE) circuitry, and digital signal processing system (DSPS) circuitry connected to the analog front-end (AFE) circuitry.[0010]

According to the present invention, a signal processing system (SPS) for an imager device includes a camera system for producing an imager signal, a correlated double sample (CDS) circuit for receiving data from an imager, a variable gain amplifier (VGA), an analog-to-digital converter (ADC) coupled to the VGA circuit, a digital gain circuit (DGC) coupled to the ADC, and an automatic gain control (AGC) circuit for controlling the VGA circuit and the DGC.[0011]

According to the present invention, gain control of a signal processing system for an imager device includes a correlated double sampler, and variable gain amplifier circuit (CDSVGA) for receiving data from the imaging system, and an automatic gain control (AGC) circuit for controlling the level of gain provided to the camera system according to the present invention.[0012]

According to the present invention, signal processing circuitry for a video camera has first and second data processing subsystems including an analog front-end (AFE) and a digital signal processing system (DSPS) connected to the analog front-end (AFE). The signal processing system according to the present invention performs signal processing functions for a low cost CCD or CMOS imaging camera capable of capturing full motion video. A complete digital video camera according to one embodiment of the present invention includes in part a sensor array, a high voltage CCD driver, a DC-DC converter, the AFE and the DSPS. The AFE data processing subsystem according to the present invention receives a mosaic CCD output (a stream combining cyan, magenta, yellow and green color samples) from the CCD camera, performs analog signal processing, and produces a digital output which can be converted to a form suitable for image display. The DSPS data processing subsystem according to the present invention accepts the digital output of the AFE data processing subsystem, performs digital processing on the received digital output of the AFE, and according to one embodiment of the present invention outputs a CCIR 601 4:2:2 YCrCb video data product suitable for presentation on a user selected display.[0013]

The AFE and DSPS data processing subsystems can be implemented as a cooperative chipset according to the present invention. The respective data processing subsystems include registers which are configured according to the present invention to share a common address space of the respective subsystems. To an external controller, the combined AFE and DSPS subsystems are operable as a unitary data processing system. Additionally, the separate AFE and the DSPS of the respective subsystems according to the present invention can be used as stand-alone units which can be addressed and controlled directly through respective I2C interfaces.[0014]

The SPS according to the present invention is thus partitionable into physically separate subsystems which can individually be fabricated on separate semiconductor substrates to enable the combined installation of the AFE and its analog functions in an integrated camera package jointly operable with an imager. This enables digital data processing to be accomplished either in the camera package itself or separately in a personal computer or other data processing system. This processing transforms the digital analog pixel samples into display-ready digital data format. Accordingly, digital and analog signal processing functions are adaptively localizable and delocalizable in accordance with application package requirements. By establishing distributable analog and digital functionalities, the entire imager signal processing functionality can be localized and fabricated in silicon. Alternatively, the analog functions can be fabricated in silicon while the digital functions are software implemented. The development of separate analog and digital modules accordingly permits convenient system retrofitting with advanced analog or digital designs. Further, the analog and digital subsystems are operable at different data rates. Thus, the output of an analog to digital (A/D) converter within the AFE subsystem can be 2× lower than the 4:2:2 YCrCb format output from the digital section according to one embodiment of the present invention. By partitioning the subsystems into separate modules, the data transmitted from the analog module subsystem to the digital module subsystem can be provided at a reduced interfacing load between the analog and digital signal processing domains creating processing efficiencies. According to the present invention, the AFE subsystem provides an imager interface which is scalable for a plurality of selected imager output formats and pixel rates. In particular, the timing signals and clocks such as horizontal and vertical shift register clocks, the applicable imager output sampling pulses, and the number of horizontal and vertical pixels per frame are user programmable according to the present invention.[0015]

Further according to the present invention, independent and dependent (through the DSPS) register addressing is enabled and timing generator parameters are programmable. An automatic gain control circuit allows for up to 98 dB of gain range including shutter exposure with an option for flickerless operation that is enabled with a hysteresis method according to the present invention.[0016]

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a block diagram of a CCD camera system according to the present invention;[0017]

FIG. 2 is a block diagram of a digital signal processing (DSP) system according to the present invention;[0018]

FIG. 3 is a block diagram of an analog data processing subsystem in accordance with one embodiment of the present invention;[0019]

FIG. 4 is a diagram of a common output waveform of a selected imager, which is processed by an analog data processing subsystem in accordance with one embodiment of the present invention;[0020]

FIG. 5 is a block diagram of a correlated double sampling variable gain amplifier (CDS/VGA) for an analog data processing subsystem according to the present invention;[0021]

FIG. 6 is a timing diagram of the operation of a correlated double sampling variable gain amplifier (CDS/VGA) operating with a two phase clock according to an embodiment of the present invention;[0022]

FIG. 7 is a block diagram of a black level adjustment circuit (BLAC) for an analog data processing subsystem according to the present invention;[0023]

FIG. 8 is a block diagram of a digital gain circuit according to the present invention;[0024]

FIG. 9 is a block diagram of a gain control system for analog data processing subsystem according to one embodiment of the present invention;[0025]

FIG. 10 is a diagram of gain control versus light intensity for shutter gain and chip gain according to the present invention;[0026]

FIG. 11 is a block diagram of an automatic gain control circuit for an analog data processing subsystem according to another embodiment of the present invention;[0027]

FIG. 12 is a histogram used in an automatic gain control circuit according to the present invention of one frame of pixel data produced by an imager system;[0028]

FIG. 13 is a diagram of gain with a flickerless setting having hysteresis in accordance with operation according to the present invention;[0029]

FIG. 14 is a timing diagram of output data as a function of clock signals for the analog subsystem according to the present invention;[0030]

FIG. 15 is a block diagram of a frequency synthesizer phase locked loop (FSPLL) circuit according to the present invention;[0031]

FIG. 16 is a timing diagram of imager signals for selected pixel operations;[0032]

FIG. 17 is a timing diagram of pixel clocks in an analog timing circuit for an analog data processing subsystem according to the present invention;[0033]

FIG. 18 is a block diagram of a timing circuit for an analog data processing subsystem according to the present invention;[0034]

FIG. 19 is a diagram of vertical timing signal operation according to one embodiment of the present invention;[0035]

FIG. 20 is a timing diagram of horizontal and clamp signals in CCD operation according to the present invention;[0036]

FIG. 21 is a timing diagram of vertical shift register signals in operation according to the present invention;[0037]

FIG. 22 is a diagram of an imager horizontal pixel row according to the present invention;[0038]

FIG. 23 is a timing diagram of vertical shift register signals for charge readout phase in operation according to the present invention;[0039]

FIG. 24 is a timing diagram of selected charge read-out signals in operation according to the present invention;[0040]

FIG. 25 is a timing diagram of selected electronic shutter control signals according to the present invention;[0041]

FIG. 26 is an I2C operation figure according to the present invention;[0042]

FIG. 27 is a diagram of the transfer curve of a luma high pass filter, according to the present invention;[0043]

FIG. 28 is a diagram of a transfer curve for a chroma low pass filter according to the present invention;[0044]

FIG. 29 shows a color separation and filter system according to the present invention;[0045]

FIG. 30 is a diagram of a white balancing circuit according to the present invention;[0046]

FIG. 31 shows a gamma correction curve according to the present invention;[0047]

FIG. 32 is a diagram of an RGB to YUV converter[0048]90 according to the present invention;

FIG. 33 is a diagram of an interpolation waveform according to the present invention;[0049]

FIG. 34 is a diagram showing a horizontal scaling pattern according to the present invention;[0050]

FIG. 35 is a block diagram of a horizontal interpolation system according to the present invention;[0051]

FIG. 36 is a diagram of a video output data format, according to the present invention, for selected portions of the horizontal video line;[0052]

FIG. 37 is a timing diagram for video data out and clock signals according to the present invention;[0053]

FIG. 38 is a diagram, according to the present invention, of the connection of selected I2C bus devices; and[0054]

FIG. 39 is a block diagram of primary and secondary I2C bus systems according to the present invention.[0055]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

As shown in FIG. 1, a video camera system (VCS)[0056]13 according to the present invention includes the following integrated circuit (IC) components: aCCD array sensor14, avertical driver circuit15 for thevideo camera system13, a crystal oscillator16, first and second signal processing subsystems (SPS)17 and18, and a DC toDC converter19 including aregulator20.First SPS17 is an analog signal processing (ASP) front end (AFE) which receives and processes video samples from thesensor14 and generates timing clocks and pulses required by the first and second SPSs17 and18, thesensor14, the DC toDC converter19, andvertical driver circuit15. Thevertical driver circuit15 generates high voltage vertical shift register clock signals and an overflow drain bias voltage provided toCCD array sensor14. The video output of thesensor14 is directly connected to the input of thefirst SPS17. DC toDC converter19 receives unregulated 5 volts DC and produces first and second regulated output voltages at 5 and 15 volts, subject to a clock signal received from the firstsignal processing subsystem17. Firstsignal processing subsystem17 produces a control signal provided to saidvertical driver circuit15.CCD array sensor14 receives an input voltage bias signal, input clock signals, and reset device bias voltages.

Camera frame formats, according to the present invention, are subject to a user selected CCD format for data input to[0057]

first SPS

17 and an output video frame format produced bysecond SPS18. CCDs, according to the present invention, include interlaced odd and even frame format cameras, each containing half the vertical lines, as well as progressive scan CCD cameras. According to particular embodiments of the present invention, a variety of frame formats can be used.

According to one embodiment of the present invention, the horizontal CCD line time can be set at 63.5 microsecond, and the pixel clock frequency can be set to match the number of horizontal pixels produced by a selected CCD. For instance, for a 512×492 CCD producing 624 total pixels/line, a clock frequency of 9.82 can be set. According to the present invention, scaling and interpolation of frames from[0058]

CCD array sensor

Data is formatted according to the present invention to accommodate multiple selected transmission channel requirements by producing a generalized output digital video signal. According to one embodiment of the present invention, first or second digital video outputs are selectable: first, a 4:2:2 YCrCb output multiplexed onto a 10 bit bus at twice the pixel sampling rate; and second, 20 bit parallel Y and CrCb outputs provided at a user selected pixel rate.[0059]

A block diagram of a signal processing system, a digital signal processing system (DSPS)[0060]18, according to one embodiment to the present invention, is shown in FIG. 2.DSPS18 in particular includes adeformatter31, acolor separation module32, a YCrCb toRGB converter33, awhite balance module34,gamma correction module35, an RGB toYUV converter36, acolor saturation circuit37, aninterpolation module48, and anoutput formatter module49.DSPS18 further includes anAWB control module54, aregister module50, a simplified I2Cinterface master module51, a I2Cinterface slave module52, and aclock driver53.Color saturation module37 includes aluma gain module38, aU saturation module39, and aV saturation module40. According to the present invention,deformatter31 is coupled to color separation andfilter module32 which in turn is connected to YCrCb to RGB (YR)converter33.YR converter33 is connected towhite balance module34 which in turn is connected togamma correction module35.White balance module34 is controlled byAWB module54.Gamma correction module35 is connected to RGB to YUV (RY)module36 which in turn is connected tocolor saturation module37. The U and V components received from RGB toYUV module36 are multiplied by user-programmable coefficient values. These values are written into respective blue saturation (for U) and the red saturation (for V) registers. These respective values can vary from 0 to 2. The luma gain value is written into theluma gain register38. The gain can vary, according to the present invention, from 0 to 31/8 using the four 1sb bits in increments of 1/8.Color saturation module37 is connected tointerpolation module48 which in turn is connected tooutput formatter module49. Simplified I2Cinterface master module51 is connected to I2Cinterface slave module52.Register module50 is connected to both I2Cinterface slave module52 andAWB control module54. DSPS30 accepts a digital output from a first signal processing system (SPS), performs selected digital signal processing according to the present invention, as will be discussed in detail below, and outputs a 10 bit 4:2:2 YCrCb data block, according to one embodiment. Input data (DIN) is first processed bydeformatter31, which is effective for extracting HREF and VREF signals from the input data, according to one embodiment of the present invention and also for subtracting therefrom the black level code. The resulting data output fromdeformatter31 is provided tocolor separation circuit32. The HREF and VREF signals are also made available on output pins from DSPS30. The color separation andfilter block circuit32 accepts input data fromdeformatter31 and generates luma samples and chroma difference samples therefrom. Each input sample color separation andfilter module32 represents a pair of complementary color pixels, as shown below. The order of red and blue lines and the color of the first pixel is programmable according to the present invention. The following respective tables, i.e., Table 1 and Table 2 provide odd and even field examples for the successive generation of red and blue lines according to one embodiment of the present invention.

TABLE 1


EVEN FIELD

Sample->	0	1	2	3	4	.

Hz line #0	Mg	G	Mg	G	Mg	.	Generates
	+	+	+	+	+		RED color
	Ye	Cy	Ye	Cy	Ye		difference
Hz line #1	G	MG	G	Mg	G	.	Generates
	+	+	+	+	+		BLUE color
	Ye	Cy	Ye	Cy	Ye		difference
. . .

[0061]

TABLE 2


ODD FIELD

Sample->	0	1	2	3	4	.

Hz line #0	Ye	Cy	Ye	Cy	Ye	.	Generates
	+	+	+	+	+		BLUE color
	G	Mg	G	Mg	G		difference
Hz line #1	Ye	Cy	Ye	Cy	Ye	.	Generates
	+	+	+	+	+		RED color
	Mg	G	Mg	G	Mg		difference
. . .

The functions performed by the color separation and[0062]

filter block

32 further include generation of luma samples; the generation of chroma samples; and the performance of luma filtering. In particular, luma samples are generated by color separation andfilter module32 by determining the sum of a current input and a next input sample, i.e., Y(I)=x(I)+x(I+1). This produces for a first sample, Y(O)=(Mg+Ye)+(G+Cy)=3G+2R+2B. Other samples result in the same mix of R, G, and B. Chroma color difference samples are generated according to the present invention by subtracting two consecutive input samples. Each video line provides only one color difference sample (either red or blue). Absent color difference samples on each line are provided by copying the previous line of color difference samples. Examples of color difference calculations are:

Crd(0) forline 0 of even field=(Mg+Ye)_[0]−(G+Cy)_[1]=2R−G;

Crd(1) forline 0 of even field=(Mg+Ye)_[2]−(G+Cy)_[1]=2R−G;and

Cbd(0) forline 1 of even field=(Mg+Cy)_[1]−(G+Ye)_[0]=2B−G.

The first red color difference on a selected line is generated by subtracting a next sample from a current sample, and the blue color difference is generated by subtracting a current sample from a next sample. Luma samples are processed with a high pass filter with a peaked high frequency response.[0063]

A pin diagram of an analog image processor system (AIPS)[0064]103 serving as an analog front end (AFE) in accordance with one embodiment of the present invention calls for AIPS103 receiving image data atpin17 according to one embodiment of the present invention. Data output is provided from AIPS103 on pins1-3,35-38, and42-44. Additional pins and functions of AIPS103 are detailed in Table 3 which follows:

TABLE 3


PIN LIST FOR ANALOG IMAGE PROCESSOR SYSTEM

Seq.	Pin	Pin	Interface
No.	Name	Type	Type	Description ofFunction

1	DOUT7	O	CMOS	This pin provides a
			12mA	digitized mosaic data
				output.
2	DOUT8	O	CMOS	This pin provides a
			12mA	digitized mosaic data
				output.
3	DOUT9	O	CMOS	This pin provides a
			12mA	digitized mosaic data
				output.
4	CLKO	O	CMOS	This pin provides a data
			12mA	clock output.
5	VDDD		Supply	This pin provides a supply
				for digital components.
6	GNDD		Ground	This pin provides a ground
				fordigital components
7	HSYNC	O	CMOS	This pin provides a hori-
			4mA	zontal sync signal.
8	SDA	B	N-CH	This pin provides a I2C
			O.D.	data pin (bidirectional).
9	SCL	I	CMOS	This pin provides a I2C bus
				control.
10	VRST	I	CMOS	Vertical sync reset.
11	INTERP	I	CMOS	Interpolate mode pin
12	DIAG0	B	CMOS	Analog test I/O
			analog
13	DIAG1/	B	CMOS	Analog test I/O freq. of
	F_CLKIN		analog	input clock
14	RST	I	CMOS	Reset pin; may be connect-
				ed to external power-on-
				reset-circuit
15	SEAL_SUB		Analog
			substrate
16	ACSUB		Analog
			substrate
17	AIN	I	CMOS	Video data input from CCD
			analog
			input
18	GNDAC		Ground	Ground for analog
19	VDDAC		Supply	Supply for analog
20	FR	I/O	CMOS	Reset gate clock pulse for
			8mA	CCD
21	VDDAM		Supply	Supply for analog
22	GNDAM		Ground	Ground for analog
23	H1	I/O	CMOS	Horizontal shift reg clock to
			28mA	CCD
24	H2	I/O	CMOS	Horizontal shift reg clock to
			28mA	CCD
25	V4X	I/O	CMOS	Vertical shift register clock
			4mA	to V-driver
26	VH3X	I/O	CMOS	Charge read pulse to
			4mA	V-driver
27	V3X	I/O	CMOS	Vertical shift register clock
			4mA	to V-driver
28	VH1X	I/O	CMOS	Charge read pulse to
			4mA	V-driver
29	V1X	I/O	CMOS	Vertical shift register clock
			4mA	to V-driver
30	V2X	I/O	CMOS	Vertical shift register clock
			4mA	to V-driver
31	OFDX	O	CMOS	Charge sweep out pulse for
			4mA	shutter control-input to
				vertical driver
32	XTAL_IN	I	CMOS	Chip input clock-2x pixel
				clock
33	XTAL_OUT	O	CMOS	Oscillator output to crystal
34	VCLK	I/O	N-CH	Hz line freq. open-drain
			O.D.4mA	clock output
35	DOUT0	O	CMOS	Digitized mosaic data out-
			12mA	put,LSB
36	DOUT1	O	CMOS	Digitized mosaic data out-
			12mA	put
37	DOUT2	O	CMOS	Digitized mosaic data out-
			12mA	put
38	DOUT3	O	CMOS	Digitized mosaic data out-
			12mA	put
39	CLK2XO	I/O	CMOS	2x pixel clock or 2*5/4
			12mA	pixel clock (when interpola-
				tion is turned on)
40	GND_CORE	Ground	Ground	Ground for pad ring1
			for pad
			ring1
41	VDD_CORE	Supply	Supply for	Supply for pad ring1
			pad ring1
42	DOUT4	O	CMOS	Digitized mosaic data out-
			12mA	put
43	DOUT5	O	CMOS	Digitized mosaic data out-
			12mA	put
44	DOUT6	O	CMOS	Digitized mosaic data out-
			12mA	put

FIG. 3 is a block diagram of an analog image processor subsystem (AIPS)[0065]113, referred to generally as an analog front end (AFE) in accordance with one embodiment of the present invention.AIPS13 includes correlated double sampler and variable gain amplifier (CDSVGA)circuit114 receiving data in the form of an input voltage (VIN) from an image acquisition device (or imager), such as are conventionally known, an analog-to-digital converter (ADC)116 connected toCDSVGA circuit114, a black level adjustment circuit (BLAC)115 feeding back to theADC116 input, adigital gain circuit117, aformatter circuit118, and anAGC controller circuit119 for controllingCDSVGA circuit114 anddigital gain circuit117.Digital gain circuit117 is connected at its input toADC116 and at its output toformatter118.AIPS113 additionally includes an analogclock generator circuit120, atiming generator circuit121, a phase lock loop (PLL)circuit122, and an I2Cbus interface circuit123.AGC controller circuit119 controlsdigital gain circuit117 andCDSVGA circuit114 and timing for OFD pulses (shutter gain).PLL circuit122 contributes to control of analogclock generator circuit120.Timing generator circuit121 provides timing signals toformatter circuit118.

gain amplifier circuit

114 parameters according to the present invention are expressed in Table 4 as follows:

TABLE 4


CDSVGA PARAMETERS

Symbol	Parameter	Conditions	Min	Typ	Max	Units

Vin_VGA	Input				1.65	V_0-p
	Voltage
	Range
Vout_VGA	Output		1.65			V_0-p
	Voltage
	Range
A_VGA	Total Gain	Minimum			0		dB
		Setting
A_VGA	Total Gain	Maximum			20		dB
		Setting
T_SS	Setting Time	Sampling				17	ns
	with	for both
	9 bits of	stages
	resolution
T_SS	Setting Time	Clearing				17	ns
	with	for both
	9 bits of	stages
	resolution
Vn_VGA	Input	Maximum			0.2	mV
	Referred	Gain
	Noise (rms)	Setting
A_step	Gain Step	Over Full	39.2	78.4	117.6	mdB
		Gain
		Range
PW_VGA	Power			78		mW
	Dissipation

Referring now to FIG. 4, there is shown a diagram of a common output waveform of a selected image used in connection with the present invention. Correlated double sampling according to the present invention is accomplished by receiving imager output data including reset noise, thermal noise, and 1/f noise, which are generated by the imager. The noise received degrades the S/N ratio and is canceled by correlated double sampling according to the present invention. Noise received during the active video portion of the CCD signal is assumed to be correlated with the noise originating during the feed-through portion of the signal. This noise is canceled by subtracting the feed-through level from the video level with correlated double sampling according to the present invention. The active video signal is the difference between feed-through and video levels according to the present invention. The active video signal varies according to light conditions. In order to insure that the full dynamic range of the[0067]

ADC

116 is utilized even under low light conditions, the imager output is amplified using a variable gain amplifier (VGA). Gain levels are controlled, according to one embodiment of the present invention, by an 8 bit control word.CDSVGA circuit114 according to the present invention includes first and second stages, each having a gain range of 0-10 dB. Thus, a total gain range of both stages of 0-20 dB is accomplished according to one embodiment of the present invention.

Referring now to FIG. 5, there is shown a block diagram of first and second CDSVGA circuit stages according to the present invention.[0068]

CDSVGA circuit

114 particularly includes first and second

respective stages

131 and132 according to one embodiment of the present invention.First stage131 includes a first capacitor C1,133, which is fixed; afirst amplifier134 connected tofirst capacitor133; a variable capacitor135 in parallel withfirst amplifier134; and afirst switch136 alternating between open and closed states in accordance with a clock φ1 in parallel withfirst amplifier134.Second stage132 includes a variable capacitor C3,143 connected to the output offirst amplifier134; asecond amplifier144 connected tovariable capacitor143; a second fixed capacitor C4,145 in parallel withsecond amplifier144; and asecond switch146 alternating between open and closed states in accordance with a clock φ2.Second switch146 is connected in parallel with secondfixed capacitor145. The total gain of theCDSVGA circuit114 according to the present invention is A=(C1/C4)* (C3/C2) and is adjustable according to the present invention by varying C2 and C3.CDSVGA circuit114 according to the present invention uses a two phase non-overlapping clock to perform the indicated CDS functions. The two phase clock according to the present invention also allows video signals to be passed to the output while maintaining a positive polarity signal.First stage131 performs correlated double sampling CDS as follows. When clock φ1 is high, the feed-through level is sampled across first capacitor C1, and the output of the first stage is forced to a predetermined reference voltage level. When clock φ1 falls, the output voltage Vo1 offirst amplifier134 follows the input gain up by −C1/C2 so that when the video level arrives, this output will be at (Vvid−Vft)* (−C1/C2) above the reference voltage.Second stage132 operates similarly, except that its switch is controlled by the second phase of the two phase non-overlapping clock. This adds a half clock delay, which is effective to maintain a positive output voltage with respect to the reference level.

Referring now to FIG. 6, there is shown a timing diagram of the two phase clock of[0069]

CDSVGA circuit

114 and the imager signal, and the output signals of the first and second stages. In particular, the leading edge of the imager input signal and the leading edge of clock φ1 are temporally coincident. Clock φ2 is out of phase with clock φ1. The output ofstage 2 rises as the output ofstage 1 falls.

Referring now to FIG. 7, there is shown a block diagram of a black level adjustment circuit (BLAC)[0070]115 according to the present invention.BLAC115 includes according to one embodiment of the present invention afirst summer element151 which receives a VGA output signal VGAOUT from CDSVGA114 (FIG. 1B), and an analog to digital converter (ADC)153.BLAC115 further includes aclip circuit155, anamplifier157 with a gain factor K, asecond summer element159, a firstunit delay element161 having a transfer function Z⁻¹, a binary-to-thermometer converter163, a secondunit delay element165 having a transfer function Z⁻¹, and a digital-to-analog converter167.ADC153 converts output signals fromfirst summer element151 into digital form. A 10 bit analog-to-digital converter (ADC) is employed according to the present invention, using a pipeline architecture with redundancy. After power-up,ADC153 runs a self-calibration routine, either after power-up or automatically between every frame or on register demand, depending on a selection bit stored in an associated register.Clip circuit155 receives a 12 bit signal fromADC153 and clips it to a 7 bit signal according to one embodiment.Amplifier157 amplifies the output signal fromclip circuit155 by a factor of K, which is a selected amplification factor.Second summer element159 receives the amplified output ofamplifier157. Firstunit delay element161 receives the output ofsecond summer element159. Binary-to-thermometer circuit163 andsecond summer element159 are connected to the output of firstunit delay element161. Secondunit delay element165 receives the output of binary-to-thermometer circuit163 at a selected line frequency and is connected at its output toDAC167, which in turn is connected tofirst summer element151. In order to maintain a constant reference level for black pixels according to the present invention, a feedback loop is implemented to set a selected black level value at the output ofADC153 with a 10 bit digital code64. This loop is active when optically black pixels are output during a portion of the horizontal blanking period. The presence of black pixels in the CCD output is indicated by a CLAMP pulse. The open loop transfer function of the black level adjustment loop according to the present invention is H(z)=Kn/(z−1) where K=1/256 and “n” is the number of black pixels during a selected clamp period. The 8 most significant bits (MSBs) of the black level accumulator are read or written through the associated register. If written, the least significant bits (LSBs) are set to 0. The clamp pulse used by the black level adjustment block is delayed according to the present invention from the actual clamp pulse, because of the latency of theADC153. The black level adjust loop can be disabled using an associated register.

Referring now to FIG. 8, there is shown a block diagram of a[0071]

digital gain circuit

180 according to one embodiment of the present invention.Digital gain circuit117 according to one embodiment includes a read only memory (ROM)171, amultiplier172, and aclipper circuit173.ROM171 stores selected gain values for use as a multiplicand bymultiplier172.ROM171 is connected tomultiplier172 which in turn is connected toclipper circuit173. The digital gain is used in conjunction with the analog gain provided through CDSVGA114 (FIG. 1B) to supply a substantial controllable range of automatic gain adjustment. According to one embodiment of the present invention, the digital gain portion of the AGC is engaged only after the analog gain has been employed and an additional level of gain is desired. According to one embodiment of the present invention, the digital gain is engaged after the entire analog gain has been deployed. Digital gain provides an additional 0 to 18 dB of gain at 0.074 dB gain steps.

Referring now to FIG. 9, there is shown a block diagram of a[0072]

camera system

180 connected to an analog image processing system (AIPS)113 which is configured according to the present invention. The imager signal is subject to external gain applied through the shutter speed setting of thecamera system180, through an analog internal chip gain provided byCDSVGA114, and throughdigital gain circuit117. A selected image is input tocamera system180, where a 60 dB gain adjustment potential can be exercised by changing the shutter speed ofcamera system180. This 60 dB range is applied when there is too much light. The gain brings the video level to a good operating range. The input signal fromCCD system180 goes to a correlated double sample/variable gain amplifier (CDSVGA)circuit114 where low frequency noise is removed and a selected analog gain level is applied. The signal is then subject to a chip gain in a range up to approximately 38 dB in theCDSVGA circuit114 anddigital gain circuit117. This gain is used to boost dark images to a proper signal level. After a received signal is digitized by ADC176, it is applied through anAGC control circuit119 to set a selected level of gain according to the present invention, in analog and/or digital gain portions. The total digital gain range applicable bydigital gain circuit117 is about 18 dB according to one embodiment, which is enough to cover a substantial range of lighting conditions indoors and outdoors, as well as physical and hardware variations.Formatter circuit118 according to the present invention takes the ADC output, clips received data to a range from binary “0000 0001 00” to binary “111 1110 11” and adds special end-of-video (EAV) and start-of-video (SAV) codes to each video line according to the present invention. The output offormatter circuit118 is available at the pins DOUT<9:0>, causing transitions to be made at the falling edges of the pixel rate clock CLKO. Timingblock circuit181 causesCCD system180 to shift data out during successive horizontal line periods. The data provided is shifted from the horizontal shift register ofCCD system180 at the imager output pin, one pixel at a time.Timing circuit181 creates the required driving signals to control the timing operations ofCCD system180. The timing signals particularly enable shifting data out ofCCD system180 are H1, H2 [=not(H1)], and FR.

Referring now to FIG. 10, there is shown a diagram of gain control versus light intensity for shutter gain and chip gain according to the present invention. In particular, a graph of an automatic gain control method according to the present invention shows a predetermined level of gain is applied by[0073]

AGC control circuit

119 acting selectively throughtiming block circuit181 uponCCD system180,CDSVGA circuit114, anddigital gain circuit117, to establish selected gain levels at a desired low intensity level, first with shutter gain to the extent possible, in incremental steps, for example, and then with chip gain applied throughCDSVGA circuit114 anddigital gain117. According to the present invention, gain is applied by first and second gain steps by using up the gain in the block closest to the input first (shutter gain) and then proceeding to a next gain block (the chip gain) once the complete shutter gain has already been applied. This improves the signal to noise ratio (SNR) according to the present invention. According to one embodiment of the present invention, the gain is split between both blocks seamlessly to ensure that the end of one gain region coincides with the beginning of the next gain region, and according to one embodiment each gain block has approximately equal gain steps.

Referring now to FIG. 11, there is shown a block diagram of an automatic gain control (AGC)[0074]

circuit

119 according to one embodiment of the present invention. In particular,AGC circuit119 includes a multiplexer (MUX)191 configured to receive input mosaic pixel values and luminance pixel values subject to control line settings of a signal line MOSAIC, permitting selection of whether mosaic pixel values or luminance pixel values are to be provided tohistogram circuit192.AGC circuit119 further includes ahistogram circuit192 subject to control signals TARGET and AGC_WIN, anerror circuit193 connected tohistogram circuit192, asummation element194 connected toerror circuit193, aclip circuit195 connected tosummation element194, aunit delay element196 connected to theclip circuit195, and asplitter circuit197 connected to theunit delay element196 effective for producing a shutter gain signal and a chip gain signal to control relative gain settings according to the present invention. The output ofunit delay element196 provides a selected gain slope subject to a gain input provided fromclip circuit195 subject to explicitly written override values AGC_GAIN_WR.AGC circuit119 receives pixel values of either mosaic or luminance style and generates a histogram of the received data for successive full frames according to the present invention. Based on the contents of the histogram, image brightness levels to be selected are determined, causing an AGC value to be incremented, decremented, or left unchanged for each frame.Histogram circuit192 stores frame data into histogram bins as discussed further below.Error circuit193 takes the histogram information and generates an error code that either increments or decrements or does not change the output AGC gain value. Summingelement194 accumulates the AGC value in view of an error signal fromerror circuit193, andclip circuit195 clips the result to insure that it is within a predetermined gain range.Splitter circuit197 takes received gain values and distributes them to an appropriate gain block including shutter and chip gain circuits.

Referring now to FIG. 12, there is shown a diagram of a histogram according to one embodiment of the present invention. Data from each frame captured by the camera system is categorized into particular bins of the histogram according to brightness level. Six explicit bins and one implicit bin are included. The applicable fixed value needed to increment each bin is shown below the chart. The target level a bin needs to exceed for a particular output of the histogram block is programmable through a target level register. The output of[0075]

histogram circuit

192 is a 7 bit word, where only one bit is high, indicating the highest level bin that exceeds a target level. According to one example, a count of 0000100 indicatesbin 1 is the highest level bin above the threshold.

Table 5 below is a diagram showing error signal generation by[0076]

error circuit

193 according to the present invention. A 7 bit code is produced fromhistogram circuit192, corresponding to one of the seven bins provided according to one embodiment of the present invention. From this, an appropriate error code is chosen and multiplied by a speed factor. The value of slew and speed are programmable. The slew value establishes the recovery speed from a very bright picture that saturated the output of the ADC.

TABLE 5


ERROR CODE GENERATION

Bin	Error	Error Output

Bin5	SLEW	Error x (Speed Multiplier)
Bin4	−2
Bin3	−1
Bin2	0
Bin1	1
Bin0	4
BinX	16

One of three AGC windows can be selected through associated register according to one embodiment of the present invention. In particular, a full AGC window, a 1/4 full AGC window, and a 1/16 full AGC window can be selected. Changing the area upon which AGC adjustments are applied permits better scene selection according to the present invention. Maximum gain, minimum chip gain, and maximum shutter gain are programmable according to the present invention. The user selects maximum gain to cause a scene to go dark at a certain low light level rather than gaining up to a noisy image. A minimum chip gain level prevents the output of the camera system from becoming saturated by the time the shutter gain is supposed to be active. If the output of the imager saturates, the shutter gain will never be engaged and particular bright scenes will be lost.[0077]

Referring now to FIG. 13, there is shown a diagram of gain as a function of code or light intensity with a flickerless setting established by hysteresis. Flickerless modes are included according to the present invention to enable indoor operation with fluorescent lights. If the fluorescent light flickers at twice the frequency of the power supply frequency, it is averaged upon receipt by the camera system over an integer number of cycles to avoid flicker in the resulting video to be displayed. There are two possible flickerless settings for particular exposure times. One setting averages one cycle of the fluorescent lights and another averages two cycles of the fluorescent light. A hysteresis loop is used according to the present invention to prevent variations in gain from causing the shutter speed to jump back and forth between the one and two cycle settings. Such flickering would produce undesirable effects, since analog gain is difficult to set to match a 2× gain step exactly. Flickerless modes are possible for combinations of camera type (PAL or NTSC) and operation environments (PAL or NTSC).[0078]

Referring now to FIG. 14, there is shown a timing diagram for a non-interpolated system according to the present invention, in which the falling edges of clock CLKO lag the falling edges of CLK[0079]2XO by 4 to 8 nanoseconds and both clocks have approximately 50% duty cycles. Data <9:0> is output at the falling edges of CLKO.

An output digital data format according to the present invention for the AIPS[0080]103 particularly includes embedded end of active video (EAV) code, blank codes, and start active video (SAV) code. The timing reference signals are contiguous with the video data and continue through the vertical blanking interval. Each timing reference signal consists of the four-word sequence. In addition to the embedded EAV and SAV timing signals, the CS7665 provides individual synchronization output signals which are employed by many video encoder circuits. These synchronization signals are typically used to interface the H.656 digital video stream to other components and subsystems. The individual synchronization signals include HREFOUT and VREFOUT. HREFOUT is an active-high signal indicating when active pixel data is being transmitted on DO_A[0-9] or DO_B[0-9]. HREFOUT is low when non-active picture data is being transmitted during horizontal blanking. Depending on the mode of operation, the HREFOUT signal follows either the HREFIN signal or the HREF defined by the EAV and SAV code. VREFOUT is an output signal that is active high when the CS7665 is putting out active video lines. The active-low portion of this signal defines the vertical blanking period. Alternately, when the ZV mode bit in register 06h is set, this output behaves as a VSYNC signal appropriate for ZV ports. The VSYNC signal is active-high during the first six horizontal line period of every field. The transition in VSYNC signal lags the HREF signal's rising edge during odd field and leads the rising edge of HREF during even field. The CS7665 delivers 4:2:2 component digital video output data in YCrCb format. The digital outputs can be configured for 10-bit interleaved Y and CrCb data, or for 20-bit parallel operation. The IN-TERL bit of the Operational Control Register 06h determines which output format is active.Logic 0 places the CS7665 in interleave mode with output data on channel “A”.Logic 1 places the CS7665 in non-interleaved mode where luminance data is output on channel “A” and chrominance data is output on channel “B”. According to one embodiment of the present invention,second SPS18 accepts 10 bit digital video data samples fromfirst SPS17 in a CCIR 656 format. Such data samples include cyan, magenta, green and yellow pixel values, with a black reference level set at code 64. Binary codes “0000 0000.xx” and “1111 1111.xx” are reserved for synchronization information. Horizontal and vertical synchronization information is encoded in the data stream, according to one embodiment of the present invention. A diagram of an input video data waveform signal tofirst SPS17 fromCCD array sensor14 begins at a horizontal reference level (HREF), enters a blanking period, and then completes with presentation of active video content. The blanking period includes transmission of EAV code, blank code, and SAV code. “T” is the output sample period. The blanking period includes three subperiods, the first being 4T sampling periods long, the second being N_bT sampling periods long, and the third being 4T sampling periods long. EAV code is transmitted in the first subperiod of the blanking period, comprising four sample periods, respectively 0-3, according to the following Table 6. For this table, “f” is the field bit, “0” is the first field, “1” is the second field, “v” is the vertical blanking bit, “0” represents active video lines, “1” represents vertical blanking, and P₃P₂P₁P₀are respective error protection bits as per CCIR standard 656.

TABLE 6


EAV code

Smpl.	Word

0	Hex FF
1	Hex 00
2	Hex 00
3	Bin 1fv1 P₃P₂P₁P₀

Blank code is transmitted in the second subperiod of the blanking period, comprising multiple sample periods including sample periods 4-7 et seq., according to the following Table 7:[0081]

TABLE 7


Blank code

Smpl.	Word

4	Hex 80 (U)
5	Hex 10 (Y)
6	Hex 00 (V)
7	Hex 10 (Y)
	repeat above 4 words

SAV code is transmitted in the third subperiod of the blanking period, comprising four sample periods including sample periods 0-3, according to the following Table 8:[0082]

TABLE 8


SAV code

Smpl.	Word

0	Hex FF
1	Hex 00
2	Hex 00
3	Bin 1fv0 P₃P₂P₁P₀

During the active video period, active video samples of mosaic data are transmitted, except during VREF, when U=V=80 (hex) and Y=10 (hex).[0083]

Referring now to FIG. 15, there is shown a block diagram of a frequency synthesizer phase locked loop (FSPLL)[0084]122 according to the present invention. In particular,FSPLL122 includes a first divide by 2 (DIV2)circuit211, afirst multiplexer212, asecond multiplexer213, and second divide by 2 (DIV2)circuit214, a divide by 4 (DIV4)circuit215, aphase detector circuit216, alow pass filter217, acharge pump circuit218, and a voltage controlled oscillator (VCO)circuit219.FSPLL122 receives an input clock signal CLKIN which is provided to first and

second multiplexers

212 and213 as well as to first divide-by-2circuit211. The output of first divide-by-2circuit211 is provided tomultiplexer212, which is subject to an output select signal F_CLKIN.Second multiplexer213 is subject to a select signal BYPASS. The output offirst multiplexer212 is provided tophase detector216 which in turn is connected tolow pass circuit217. Lowpass filter circuit217 is connected in turn to charge pump218 which in turn is connected toVCO circuit219, which then is connected to second divide by 2circuit214. The output of divide-by-2circuit214 is provided to divide-by-4circuit215 and tomultiplexer213. Divide-by-4circuit215 is connected to phasedetector216.Multiplexer213 provides a ×4 multiplied output signal fromfrequency synthesizer122 when select signal, BYPASS, is 0. Thus,FSPLL122 provides 4× the input pixel clock frequency according to one embodiment of the present invention. Table 9 below provides parameters according to one embodiment of the present invention for FSPLL122:

TABLE 9


FREQUENCY SYNTHESIZER PARAMETERS

Symbol	Parameter	Min	Typ	Max	Units

F^XTALIN	CLKIN Frequency	6.75	27	MHZ
		(F_CLKIN = 1)	(F_CLKIN = 0)
D^XTALIN	CLKIN Duty Cycle	20	80	%
F^X4CLK	X4CLK Frequency	27	54	MHZ
	Output Jitter
		200	ps
	Duty Cycle
		50	%
	PLL Acquisition
		20	us
	Time

Referring now to FIG. 16, there is shown a timing diagram of imager signals during pixel sampling operation. The timing signals are generated from clock X4CLK, as shown in FIG. 17. For every pixel clock cycle, one pixel is read-out from[0085]

CCD system

180 of FIG. 7. Pixel timings are based on clock X4CLK running at a factor of 4 or 5 (when horizontal interpolation is used for the 624 active pixel case) times the pixel clock. Circuit timings are referenced from a reset gate pulse FR. The pixel timings change with the total number of output pixels per horizontal line. The horizontal line period is fixed at 63.5 microseconds for NTSC systems. The rising edge of H1 in FIG. 16 is adjusted relative to the rising edge of clock φ1. The rising edge of FR is adjusted relative to the falling edge of clock φ2, and the falling edge of FR is adjusted relative to apoint 1/4 pixel cycle after the falling edge of clock φ2 for the non-interpolation mode and 1/5 pixel cycle after the falling edge of clock φ2 for the interpolation mode. Each edge is controlled by two bits in an associated register, and the relative phase delays are listed in Table 10 below, where Δt is about 1 to 2 nanoseconds:

TABLE 10


H1 AND FR PHASE ADJUSTMENTS

H1

FR

rising edge	falling edge	rising edge	falling edge	Phase Delay
bits <7:6>	bits <5:4>	bits <3:2>	bits <1:0>	(ns)

00	00	00	00	0
01	01	01	01	Δt
10	10	10	20	2*Δt
11	11	11	11	3*Δt

Referring now to FIG. 17, there is shown a timing diagram of pixel clocks from timing block circuit[0086]181 (FIG. 8) according to the present invention. Timingblock circuit181 provides driving signals to control CCD system180 (FIG. 8) and controls the embedding of active video start and end information in the data stream produced byCCD system180. The timing signals produced are particularly shown in FIG. 17 and are discussed in the associated text. The timing signals are tailored according to the present invention to selected operating phases ofCCD system180. Charge read-out timing signals apply when once every field, the accumulated charge generated by the photo gates ofCCD system180 is transferred to the vertical shift register. This transfer takes place during a specific horizontal line. The timing signals involved in this operation are V1X, V2X, V3X, V4X, VH1X, and VH3X. Vertical transfer timing signals apply when on every horizontal line period, one line of data from the vertical shift register is shifted down into the horizontal shift register. The timing signals involved in this transfer are V1X, V2X, V3X, and V4X.

[0087]

Timing generator circuit

121 includes ahorizontal timing generator231 and avertical timing generator232, as shown in FIG. 18. Thehorizontal timing generator231, time_hor, generates timing signals required to control events on a per horizontal line basis. These timing parameters include the following signals:

1. HREF—horizontal reference signal. It stays active high during the active video portion of each line.[0088]

2. HREF (−4)—HREF advanced by 4 pixel clock periods.[0089]

3. HENB—horizontal shift register clock enable signal. Enables H[0090]1 and H2 out of analog timing.

4. V[0091]1X, V2X, V3X, and V4X—vertical register shift clock. Used both during vertical transfer and charge read-out.

5. VH[0092]1X, VH3X—CCD charge read-out pulse.

6. VCLK—signal used by the DC/DC converter. At the normal mode, it is the same as HREF; at the faster mode, it operates at about 1/16 of the horizontal line frequency and gets reset at the beginning of HREF.[0093]

7. HSYNC—horizontal sync signal.[0094]

8. CLAMP([0095]14)—black clamp signal provided to the ADC. CLAMP(14) is this clamp signal delayed by the pipeline delay of the ADC.

9. OFDX—overflow drain control clock. This signal sets the electronic shutter speed.[0096]

Referring now to FIG. 19, there is shown a diagram of vertical timing signal operation according to one embodiment of the present invention. The vertical timing generator generates timing signals to control events on a per field basis. These timing signals include the following:[0097]

1. OFDEN—enable signal to control the number of lines the OFDO clock is active.[0098]

2. XFRLIN[[0099]3]—indicates lines on which CCD charge read-out is performed.

3. VRST—Vertical field reset signal.[0100]

4. FLD—FLD=1 indicates the odd field of CCD data.[0101]

5. VREF—vertical reference signal. It is active high during the active video lines.[0102]

The CCD system includes an array which is read out alternately as odd and even fields with interlaced horizontal lines. Thus, each field has half the total number of horizontal rows.[0103]

Referring now to FIG. 20, there is shown a timing diagram of horizontal and clamp signals, according to the present invention. In particular, FIG. 20 shows timing information for HREF, HSYNC, CLAMP, and HENB signals according to the present invention. The waveforms for these signals are repeated on every line. The horizontal shift register clocks H[0104]1 and H2 operate at the CLKO frequency and are active throughout the horizontal line period except when HENB is high.

Referring now to FIG. 21, there is shown a timing diagram of vertical shift register signals in operation according to the present invention.[0105]

Referring now to FIG. 22, there is shown a diagram of an imager horizontal pixel row according to one embodiment of the present invention. Each horizontal row of a CCD system is divided into regions corresponding to a particular type of pixel. Different camera systems have different numbers of pixels in each region, and the timing signals established according to the present invention are adaptive to this requirement. Horizontal timing for[0106]

CCD system

180 according to the present invention is particularly based on maintaining a fixed 63.5 microsecond horizontal line time.

Referring now to FIG. 23, there is shown a timing diagram of selected vertical shift register signals according to the present invention. FIG. 23 particularly shows the timings for the V[0107]1X through V4X signals. The specified waveforms repeat on every horizontal line except during charge transfer lines. Signals VH1X and VH3X are required during charge read-out, as shown in FIG. 24.

Referring now to FIG. 24, there is shown a timing diagram of charge read-out signals according to the present invention. OFD signal is used to control the electronic shutter timing of the camera system. Overflow drain control signals are shown in FIG. 25.[0108]

Referring now to FIG. 25, there is shown a timing diagram of electronic shutter control signals according to the present invention.[0109]

Referring now to FIG. 26, there is shown a timing diagram for I2C interfacing according to the present invention. The AIPS I2C interface can be accessed two ways. One, an external controller can read and write to all the registers of AIPS[0110]103. Two, the AIPS103 can be accessed through the DSPS which controls the AIPS register reads and writes. Control of the I2C interface is through two external pins, a bidirectional data pin (SDA) and a serial input clock (SCL) according to I2C protocol specifications. SDA is a bidirectional line, connected via an external pull-up resistor to a positive supply voltage as shown in detail in the related application which is incorporated herein by reference. When the bus is free, both lines are HIGH. The output stages of devices connected to the bus have an open drain in order to perform a wired-AND function according to the present invention. Data on the I2C bus can be transferred at a rate of up to 400 kbit/sec, according to one embodiment. The write format includes a three byte packet. The first byte is the station address with the direction bit set to “0” in order to indicate a “write.” The second byte is the device register address (0 . . . 255). The third byte is the register data. A sequential write can be done if instead of the stop bit at the end, an acknowledge bit is sent and the sequential address is written with the next data byte. The address set format includes a two byte packet. The first byte is the device address with the data direction bit set to “0” in order to indicate a a “write.” The second byte is the register address. The address set format is the same as a write format; however, the register data is not sent. It is used to set the address for the read format. The read format includes two bytes. The first byte is the station address with the direction bit set to “1” in order to indicate a “read.” The AIPS103 then sends a byte back on the bus, from the register which was addressed by the last write format or address set format. A sequential read is done by using an acknowledge bit permitting data from the sequential address to be read. Writing a 1 to the reset registers triggers a software reset in AIPS103. Software reset resets selected digital blocks except for I2C and ADC calibration logic elements.

Referring now to FIG. 27, there is shown a diagram of the transfer curve of a luma high pass filter, according to the present invention. High frequency components are emphasized in transfer curve to sharpen the selected images. A luma filter equation, according to one embodiment of the present invention, is given by the relationships[0111]

H(z)=[−Z⁻¹−2Z⁻²+14Z⁻³−2Z⁻⁴−Z⁻⁵].

The transfer curve for a chroma low pass filter according to the present invention is shown in FIG. 28. The chroma filter according to the present invention is an eight tap FIR filter which accepts chroma color difference samples. Since CCDs are a sampled system, the input image is be pre-filtered optically to limit pass band to 6.75 MHZ with fs=13.5 MHZ. However, once the image is converted into an CCD output signal, chroma is limited to 3.375 MHZ, according to one embodiment of the present invention, because of the need to sub-sample chroma at 6.75 MHZ according to a 4:2:2 format. The chroma filter equation is given by H(Z)=[0305*Z[0112]⁻¹+0.0903*Z⁻²+0.1664*Z⁻³+0.2205*Z⁻⁴+0.2205*Z⁻⁵+0.1664*Z⁻⁶+0.0903*Z⁻⁷+0.0305*Z⁻⁸], with coefficients quantized as follows:

TABLE 11


QUANTIZED CHROMA LPF COEFFICIENTS

	Floating Point	Quantized	Implementation

0.0305	32/1024	32
0.903	92/1024	64 + 32 + 4
0.1664	168/1024	128 + 32 + 8
0.2205	224/1024	128 + 64 + 32

Word sizes at selected internal nodes of the color separation and filter are shown in FIG. 29.[0113]

Filter system

344 particularly includes aluma calculation module345, aluma filter346, a divide by 8module347, apix difference module348, achroma filter349, and a divide by 8module50. The output of deformatter31 (FIG. 1B) is connected to the input ofluma calculation module345 andpix difference module348. The outputs ofluma calculation module345 and pix difference module are respectively provided toluma filter346 andchroma filter349. The respective outputs ofluma filter346 andchroma filter349 are respectively connected to divide by 8

modules

347 and350, to produce respectively a luma output and a chroma output which is provided to YCrCb to RGB converter33 (FIG. 1B). The luma filter output may contain negative numbers because of the negative coefficients inluma filter346. These negative luma values are maintained until after the conversion to RGB in YCrCb toRGB converter33. Any negative numbers in RGB are then be clipped to zero. Color separation andfilter module32 according to the present invention further include a programmable red line/blue line signal, for determining whether a current CCD line is to provide blue or red color difference pixels. This signal is generated in a control block (not shown) which is programmable as to whether the first CCD line is blue or red. Further, according to the present invention, a programmable color difference operation signal (POSPIX) is produced. This signal determines whether the first pixel of the first line from CCD is treated as positive or negative for taking color difference.

The output (luma, blue difference, red difference) of the color separation and[0114]filter module32 is transformed into red, green and blue components by the YCrCb2RGB converter323. The conversion involves multiplication with a 3×3 matrix, according to one embodiment of the present invention, as shown below: $\langle \begin{matrix} R \\ G \\ B \end{matrix} \rangle = \langle \begin{matrix} K_yR \\ K_yG \\ K_yB \end{matrix}  \begin{matrix} K_CrR \\ K_CrG \\ KCrB \end{matrix} | \begin{matrix} K_CbR \\ K_CbG \\ K_CbB \end{matrix} \rangle \langle \begin{matrix} Y \\ Cr \\ Cb \end{matrix} \rangle$
The above coefficient values as specified below are settable, according to the present invention, as follows, for example:[0115]
K_yR=1; K_CrR=31/16; K_CbR=−1/2;
K_yG+1; K_CrG+−7/16; K_CbG=−9/16;
K_yB=1; K_CrB=−5/16; K_CbB=31/16;
Referring now to FIG. 30, there is shown a diagram of a white balancing circuit according to the present invention.[0116]White balancing circuit370 includes a whitebalance adjustment circuit371 and awhite clipping circuit372. Whitebalance adjustment circuit371 is connected towhite clipping circuit372 with red and blue adjustment value lines, respectively R2 and B2. Whitebalance adjustment circuit371 receives input red, green, and blue signal value levels, as well as Kr and Kb gain factor values. Whitebalance adjustment circuit371 produces output signals R2, B2, RMG, and BMG.White clipping circuit372 receives input signals green, adjusted red, and adjusted blue, respectively G, R2 and B2. White clipping circuit further produces output clipped red, green and blue signal values, respectively R′, G′, and B′. Ambient white light may, accordingly, be shifted towards bluish or reddish hue based on the characteristics of the light source. This relative shift is often characterized by the color temperature of the light. A low temperature indicates a shift towards red and high temperature indicates a shift towards blue. The objective of the color balancing function is to balance the blue and red components by adjusting their relative gains so that a white object indeed appears as white. The steps required for color adjustment, according to the present invention, include adjusting the red and blue components of the input signal using first and second gain factors values, Kr and Kb, respectively. These factors are computed in a control section (not shown) and vary between 0.5 and 2.5. In particular, R2=(1+Kr)*R1, and B2=(1+Kb)*B1. Further, two color components RMG=(red-green) and BMG=(blue-green) are computed, according to the present invention. These are used by the control block (not shown) in computing Kr and Kb. Further, according to one embodiment of the present invention, the black restored red and blue values are clipped, as well as green, to a maximum binary equivalent of 0.92 (i.e., 235 for an 8 bit sample). The generation of Kr and Kb is accomplished by computing average values of RMG and BMG over a 1/4 window in the middle of a selected screen. Kr and Kb registers are then initialized to hex 80, and at the end of each field the value is adjusted up or down based on the average RMG and BMG per field. A final value is determined over eight fields through a process of successive approximation. These registers can also be directly written through the I2C interface.
Referring now to FIG. 31, there is shown a gamma correction curve according to the present invention. Because of the non-linear characteristics (input voltage versus light intensity) of a target CRT display, the CCD output is gamma corrected so that the light intensities of the camera input and CRT output are linearly related. Gamma correction, according to one embodiment of the present invention, is implemented using look-up RAM tables. These tables are user-programmed to contain any desired correction curve.[0117]
Referring now to FIG. 32, there is shown a diagram of an RGB to[0118]YUV converter36 according to the present invention. The RGB to YUV converter performs a matrix multiplication operation to convert the RGB color components into YUV components.
Referring now to FIG. 33, there is shown a diagram of an interpolation waveform according to the present invention. The YUV output from the[0119]color saturation module37 is interpolated, according to the present invention, to increase the number of pixels by a factor of 5/4. The interpolation operation is selectable by the user through a register bit and is used for generating a 640×480 (active pixels) VGA output from a 512×480 (active pixels) CCD. This requires a 10 MHZ sampling clock before interpolation and a 12.5 MHZ clock after interpolation.
Referring now to FIG. 34, there is shown a diagram showing a horizontal scaling pattern according to the present invention. According to the pattern, the pixel values of a first set,[0120]410, of four registers, respectively410A-410D, are used to generate pixel values for a second set,411, of five registers, respectively411A-411E. In particular, according to the present invention, the values of the first and last registers of thefirst set410 are entered into corresponding first and last of thesecond set411. Thus, the value ofregister410A becomes the value ofregister411A; and the value ofregister410D becomes the value ofregister411E. Further, the value ofregister411B is generated by adding 1/4 of the value ofregister410A to 3/4 of the value ofregister410B. The value ofregister411C is generated by adding 1/2 of the value ofregister410B to 1/2 of the value ofregister410C. Finally, the value ofregister411D is generated by adding 3/4 of the value ofregister410C to 1/4 of the value ofregister410D. Since the interpolation requires data to go from a 10 MHZ sampling rate to a 12.5 MHZ sampling rate, synchronization, according to the present invention, is employed.
Referring now to FIG. 35, there is shown a block diagram of a[0121]horizontal interpolation system48 according to the present invention. Thesystem48, according to the present invention, insures that the 12.5 MHZ pixel latch always picks a valid, synchronized data. The select logic of interpolation system338 is reset by the HREF signal according to the present invention. In particular, interpolation system420 includes a successively interconnected series of first through fourth respective synchronizing registers421A-421D. First synchronizingregister421A receives an input HREF signal at 10 MHZ and is clocked at 10 MHZ, outputting the HREF signal at clock time tosecond synchronizing register421B. Second, third and fourth synchronizing registers, respectively421B-421D, are clocked at 12.5 MHZ, with the output of thesecond synchronizing register421B being applied to the input of thethird synchronizing register421C and the output of thethird synchronizing register421C being applied to thefourth synchronizing register421D. The output of the fourth synchronizing register is applied to selectlogic422 which controlsmultiplexer423. Interpolation system420 further includes a fourpixel buffer425 clocked at 10 MHZ, which receives four input pixel values at a 10 MHZ rate. These four values are provided tointerpolation module426 which provides applicable values to fivepixel buffer427.Multiplexer423 receives the five pixel values frombuffer427 and provides selected ones thereof topixel output buffer424 which is clocked out at 12.5 MHZ.
Referring now to FIG. 36, there is shown a diagram of a data format, according to the present invention, for portions of the horizontal video line. Values for particular parameters in FIG. 39 are given in Table 12 which follows:[0122]
TABLE 12
Horizontal line size (pixels)
Parameter Description 429 624 780 858 Units
1/T Output clock 13.5 19.6 25 27 MHZ
frequency
N_b HREF samples -8T 146 216 272 268 samples
AV Active video 704 1024 1280 1440 samples
samples
Referring again to FIG. 37, there is shown a timing diagram for data out and clock signals according to the present invention.[0123]
Referring again to FIG. 26, there is shown a timing diagram for complete data transfer on an I2C bus, according to the present invention, to permit parallel operation of multiple CCD cameras. In order to allow multiple cameras to be connected together, according to the present invention, the video data output bus can be put in Hi-Z condition through the setting of a predetermined register bit.[0124]
Referring now to FIG. 38, there is shown a diagram, according to the present invention, of the connection of selected I2C bus devices. The[0125]DSPS18, according to the present invention, provides a secondary I2C interface for accessing the internal control and status registers. External pins are a bidirectional data pin (SDA) and a serial input clock (SCL). The protocol follows I2C specifications. These specifications are well known. See, for example, Philips Semiconductors “Desktop Video Data Handbook.” The secondary I2C interface works, according to the present invention, in slave mode. In addition, theDSPS18, according to the present invention, has a simplified master I2C interface for communicating with AFE. Both the serial data line (SDA) and the serial clock line (SCL) are connected via external pull-up resistors Rp (472 and473) to a positive supply voltage VDD. When the I2C bus is free, both lines are HIGH. The output stages of devices connected to the bus have an open-drain or open-collector, according to the present invention, in order to perform a wired-AND function. Data on the I2C bus is transferred at a rate of up to 400 kbit/sec in fast mode according to the present invention. The number of interfaces to the bus is dependent, in part, on, according to the present invention, limiting bus capacitance of 400 pF.
Referring now to FIG. 39, there is shown a block diagram of primary and secondary I2C bus systems according to the present invention.[0126]DSPS18 has two 12C interfaces: one is a primary interface486 that is used by an external controller to read or write data fromDSPS18; the second is a simplified master interface which is used byDSPS18 to communicate with other slave devices such as the camera analog chip (i.e., an AFE), a video encoder chip, etc. This arrangement, according to the present invention, is based on two I2C buses—asecondary bus484 for the camera system, and a primary bus486 to be shared with multiple cameras controlled by a single I2C primary controller. In a single camera application, the external camera may connect directly to all the devices. The primary I2C interface ofDSPS18 would not be used in such an application. TheEPROM481 is provided for initializing all the registers on all the devices onsecondary I2C bus484 on reset in a stand-alone application. This includes loading a new station address into each of the devices, provided it can be done by writing to a register location. In a I2C arrangement according to the present invention, communications between external controller and any of the devices on the camera take place throughDSPS18 according to the present invention. The external I2C controller accesses registers on any of selected devices throughDSPS18 according to the present invention. The external controller sends, for example, a 4 byte command toDSPS18 including a DSPS station address, according to the present invention, destination device station address, and a desired register address of the destination device followed by data, if it is a write operation.DSPS18 then completes the required transaction with the destined device. According to the present invention, the external controller sends 4 bytes, with theDSPS18 astation address repeated in the first 2 bytes, even when accessing registers onDSPS18. In applications where theDSPS18 is not used as an I2C gateway for slave devices,DSPS18 can be set to a mode where it needs only 3 bytes (theDSPS18 station address, its register address, and data) to access registers on theDSPS18. The descriptions below assumeDSPS18 is in a 4-byte mode. In a 3-byte mode, the destination station address is not required.
In a diagram of an EPROM map for initialization of DSPS registers according to the present invention, an end of reset triggers a DSPS to download register values in any of selected attached devices from the EPROM. In particular,[0127]DSPS18 goes out and fetches 3 bytes from theEPROM481. These 3 bytes represent destination station address, register address, and data.DSPS18 then writes the data into the specified register of the destined station. After completing this process,DSPS18 goes out and reads the next 3 bytes fromEPROM481. The number of register data to be read from theEPROM481 is loaded into a 2-byte count register inDSPS18. The top 6 bytes in the EPROM specify the number of triple bytes to be read fromEPROM481. During thetime DSPS18 is doing initialization fromEPROM481, an INITACT bit is set in the status register ofDSPS18. Attempts to write to DSPS registers by an external controller are ignored during this time. The only exception to the above rule is when accessing data for the gamma RAM onDSPS18.
Write operations from an external controller to any secondary device via[0128]DSPS18 require the following operations. The external controller sends a 4-byte write command toDSPS18.DSPS18 initiates a write operation to a selected destination slave device and sets the I2CBUSY bit in the appropriate status register. The external controller polls the status register to check ifDSPS18 has completed the command.
[0129]DSPS18 has a command buffer which allows an external controller to queue an additional command while the current command is still being executed. If more than one command is sent before the I2CBUSY bit is cleared,DSPS18 saves only the last command and executes it after the current one is completed. Commands that involve writing or reading only toDSPS18 registers are not put in the queue and are executed immediately without affecting any transactions going on in the master I2C interface. Any attempt to write data to one of theDSPS18 registers from an external controller whileDSPS18 is busy initializing fromEPROM481 will be ignored. However, reads fromDSPS18 are allowed. If during a read or write operation to aslave device DSPS18 fails to receive an acknowledge bit, the execution of the command is aborted and the NODEV bit in the status register is set. This bit remains set unless it is explicitly cleared by writing to it or a new command is written toDSPS18.
Read operations from the external controller via[0130]DSPS18 require a 4-byte command, according to the present invention, (when 4BYTEMODE pin is tied low) similar to the write operation described above. The behavior of I2CBUSY bit is also similar to what it is during write command. The read format consists of two bytes. The first byte is the device address with the direction bit set to “1” in order to indicate a “read.” The DSPS chip then sends one byte back on the bus, from the register which was addressed by the last write format or address set format.
The address set format consists of a 2 byte packet. The first byte is the device address with the data direction bit set to “0” in order to indicate a “write.” The second byte is the register address. The address set format is the same as a write format, however, the register data is not sent. It is used to set the address for the read format.[0131]
The I2C station addresses for both[0132]AFE482 andDSPS18 can be changed, according to the present invention, through the I2C interface. Location FF, for example, holds the station address forDSPS18 and FE holds the station address forAFE482. Once a station address is changed, all subsequent I2C accesses use the new station address.
The status register (01h) on[0133]DSPS18 contains 3 bits related to the I2C interfaces:
(1) bit[0134]0: NODEV—If DSPS fails to get an acknowledge from a slave device, the NODEV bit is set. It is cleared by writing to the register or if DSPS starts a new read/write transaction with a slave device.
(2) bit[0135]1: I2CBUSY—This bit is set whenDSPS18′ master I2C interface is active talking with a selected of the slave devices. It is cleared when, according to the present invention, a transaction is completed orDSPS18 fails to receive an acknowledge.
(3) bit[0136]2: INITACT—Upon reset,DSPS18 starts initializing all its own registers as well as the ones on slave devices. The INITACT bit is set during this process and is cleared when initialization is done.

[0137]

DSPS

18 outputs data compatible with ZV port specifications according to one embodiment of the present invention. This mode is enabled by a register control bit called ZVMODE at register hex 06. The VREF output pin normally outputs the vertical blanking signal, but when ZVMODE bit is set, it outputs a VSYNC signal. The following register list shows registers for both theAFE482 andDSPS18 chips.

TABLE 13


Register	Register		Default
(hex)	name	Access	Value	Comments

00h	Reset	W	00	bit 0 = 1 for Argus reset
				bit 4 = 1 for Gorgon reset
				Bit is automatically cleared
01h	Status	R	00h
02h	Reserved	R	00h
03h	Digital	R/W	08h	1
	gain
04h	Brightness	R	NA
05h	Feature	R/W	00h
	control
06h	Operation	R/W	0Dh
	control
07h	Reserved
08h	Red	RAV	80h	1
	balance
09h	Blue	R/W	80h	1
	balance
0Ah	Red	R/W	80h	1
	saturation
0Bh	Blue	R/W	80h	1
	saturation
0Ch	Gamma	R/W	01h
	correction
0Dh	Reserved	R	00h
0Eh	Test	R/W	00h
	control
	register A
0Fh	Test	R/W	00h
	control
	register B
10h	K_yR	R/W	80h	2
11h	K_CrR	R/W	7Ch	31/16
12h	K_CrR	R/W	E0h	−8
13h	K_yG	R/W	80h	2
14h	K_CrG	R/W	E4h	−7/16
15h	K_CbG	R/W	DCh	−9/16
16h	K_yB	R/W	80h	2
17h	K_CrB	R/W	ECh	−5/16
18h	K_CbB	R/W	7Ch	31/16
19h	Slave data	R	00h	Slave data register
	register			cannot be written by
				external controller
1Ah	EPROM	R/W	FFh	Indicates how many
	count -			triple bytes must be read
	low byte			from EPROM upon reset
1B	EPROM	R/W	FFh
	count -
	high byte
1C	Version	R	FFh	FF is the code for
	(major			revision A
1D	Version	R	00h	00 is the code for revision A
	(minor)
1E-1F	Reserved	R	00h
20h	Power	R/W	00	bit 0 = 1 for Argus power
	down			down
				bit 4 = 1 for Gorgon power
				down
21h	Test	R/W	00	bit 0 = 1 for Argus test enable
	enable			bit 4 = 1 for Gorgon test
				enable
22h	Test	R/W	00
	control
	register
23	Version	R	00
24h	Operation-	R/W	00
	al control
25h	Reserved	R	00
26h	Analog	R/W	00
	gain
27h	Shutter	R/W	00
28h	Brightness	R	00
29h-2Fh	Reserved	R	00h
30h-3Fh	Timing	R/W
	control
40h-4Fh	Timing	R/W
	control
50h-55h	Timing	R/W
	control
56h-FDh	AFE	R/W	54h
	station
	address
FEh	Gorgon's	R/W	54h
	station
	address
FFh	Argus'	R/W	34
	station
	address

Reset-00h. Writing a 1 to[0138]bit0 at location 00h triggers a software reset inDSPS18. Similarly, writing a 1 tobit4 at location 00h triggers a software reset inAFE482 chip. The bit automatically clears. The reset stays active for 32 cycles of CLKIN.
TABLE 14
Bit Name Access Default Description
0 Reset_dig R/W 0 Writing a 1 resets the part.
Behaves identical to reset pin.
1-3 Reserved R 0
4 Reset_ana R/W 0 Writing a 1 resets the part.
5-7 Reserved R 0
The following actions take place on either the register reset or the pin reset:[0139]
1. All registers go back to default values.[0140]
2. If an external EPROM is hooked with the local 12C bus, the EPROM data is downloaded into specified device registers at the end of reset.[0141]

Register 01h—status. This register is read only.[0142]

TABLE 15


Bit	Name	Access	Default	Description

0	EVNFLD	R	—	1 indicates even field.
1	NODEV	R		0	1 indicates that the addressed slave
				device on local I2C bus did not
				respond.
2	12CBUSY	R		0	1 indicates that DSPS master I2C
				is still busy accessing the
				addressed slave device.
3	INTACT	R		0	1 indicates that DSPS master I2C
				is busy initializing registers from
				EPROM.
4-7	Reserved

Bit[0143]0 (even field) is provided as a means of synchronizing to the vertical camera rate. The bit is read as a “1” during even fields and as a “0” during odd fields.
Register 1Ch—Version (major). This register is read-only. Data=FF for revision A.[0144]
TABLE 16
Bit Name Access Default Description
0-3 R F
4-7 R F
Register 1Dh—Version (minor). This register is read-only. Data=00 for revision A.[0145]
TABLE 17
Bit Name Access Default Description
0-3 R 0
4-7 R 0
Digital gain—03h. This register controls the digital gain applied to the luma signal after the RGB2YUV block. The gain value varies from 0 to 31/8 in increments of 1/8.[0146]
TABLE 18
Bit Name Access Default Description
0-4 DIG_GAIN R/W 00h 00h = 0 (no brightness), 08h =
1, 1Fh = 31/8
5-7 Reserved
Brightness—04h. [not used] This register contains the maximum value of the low-pass filtered luminance signal. It changes at field rate. When AGC is enabled, the AGC circuit attempts to keep this value at 0EBh ([0147]235).

Feature control—05h.[0148]

TABLE 19


Bit	Name	Access	Default	Description

0	AGCEN	R/W	0	Automatic gain control enable
				for DSPS' AGC if implemented.
1	AWBCTL	R/W	0	Automatic white balance con-
				trol. This bit is set by writing
				a “1”. Writing a “0” has no
				effect. This bit is read as a “1”
				while AWB is in progress (it
				takes about 160 msec.). While
				AWB is in progress, AGC is
				disabled if implemented. This bit
				is read as a “0” when AWB is
				not in progress.
2	GAMMAON	R/W	0	Default is gamma correction
				disabled.
3	LUMAOFF	R/W	0	Default is high pass luma filter
				enabled.
4	CHRMOFF	R/W	0	Default is low pass chroma filter
				enabled.
5	AGCWIN	R/W	0	Select full window when set to 1
				for AGC peak.Select center 1/4
				window by default. Applies to
				devices where AGC is
				implemented.
6-7	Reserved	R		0

DSPS operational control—06h[0149]

TABLE 20


Bit	Name	Access	Default	Description

0	OBLUE-	R/W	1	Blue line/red line indicator for
	LINE			odd field. This bit determines if
				the first line after VREF pro-
				vides blue line color difference
				pixels (“1”) or if it provides red
				line color difference (“0”).
1	EBLUE-	R/W	0	Blue line/red line indicator for
	LINE			even field. This bit determines if
				the first line after VREF pro-
				vides blue line color difference
				pixels (“1”) or if it provides red
				line color difference pixels (“)”).
2	POSPIX	R/W	1	Color difference operation signal
				for odd field. This bit determines
				if the first pixel of the first line
				from the CCD is treated as a
				positive signal or if it is treated
				as the negative pixel in the color
				separation block.
3	HIZDOUT	R/W	1	Puts DOUT bus in Hi-Z state for
				parallel operation of multiple
				cameras. Hi-Z enabled when set
				to 1 (see notes).
4	INREF	R/W	0	HREF and VREF become input
				pins. EAV and SAV codes in
				data stream are ignored.
5	DVCI	R/W	0	Data output is in 8 bit DVCI
				mode when set.
6	ZVMODE	R/W	0	ZV mode output enabled when
				set.
7	Reserved	R		0

Note on HIZDOUT: the HIZDOUT bit and the HIZENB pin both affect the output pin tri-state conditions in the following manner. The 12C interface pins never go into Hi-Z mode.[0150]
TABLE 21
Effect on Output Pins
HIZDOUT bit HIZENB pin (XOR)
0 0 Active
0 1 Hi-Z enabled
1 0 Hi-Z enabled
1 1 Active
Red balance—08h. This register controls the red contribution to the R-Y chrominance signal. When it is 00h, the red contribution is minimized. When it is FFh, the red contribution is maximized. When AWB is in progress, this value is adjusted so that the absolute magnitude of the R-Y is minimized.[0151]
TABLE 22
Bit Name Access Default Description
0-7 RBAL R/W 00h This is the gain factor for the R-Y
component used for white
balancing.
Blue balance—09h. This register controls the blue contribution to the B-Y chrominance signal. When it is 00h, the blue contribution is minimized. When it is FFh, the blue contribution is maximized. When AWB is in progress, this value is adjusted so that the absolute magnitude of the B-Y is minimized.[0152]
TABLE 23
Bit Name Access Default Description
0-7 BBAL R/W 00h This is the gain factor for the B-Y
component used for white
balancing.
This register controls the amplitude of the final R-Y chrominance signal. At 00h, the amplitude of the R-Y signal is minimized. When it is FFh, the amplitude of the R-Y signal is maximized. A value of 80h sets the saturation to 1.[0153]
TABLE 24
Bit Name Access Default Description
0-7 RSAT R/W 80h This is the gain factor for the R-Y
component in the final output
signal.
Blue saturation—0Bh. This register controls the amplitude of the B-Y chrominance signal. At 00h, the amplitude of the B-Y signal is minimized. When it is FFh, the amplitude of the B-Y signal is maximized. A value of 80h sets the saturation to 1.[0154]
TABLE 25
Bit Name Access Default Description
0-7 BSAT R/W 80h This is the gain factor for the B-Y
component in the final output
signal.

Gamma Correction—0Ch. The gamma correction register provides access to the R, G and B gamma look-up table RAMs. The procedure for write and read access to all the 255 locations on each of the gamma RAM is shown below. The diagram shows the mode where register access through 12C requires three bytes (station address, register address, data). 4-byte mode access requires an additional byte in the beginning where the station address is repeated.[0155]

TABLE 26


Bit	Name	Access	Default	Description

0	GAMMA	R/W	1 h	Selects blue gamma RAM for
	BLUE			subsequent RAM accesses.
1	GAMMA	R/W	0 h	Selects green gamma RAM for
	GREEN			subsequent RAM accesses.
2	GAMMA	R/W	0 h	Selects red gamma ram for
	BLUE			subsequent ram accesses.
3-7	Reserved

Writing to the gamma register (OC) selects the R, G and/or B RAM. Continuing data writes without sending a stop bit after the register write results in writes to RAM location starting from location 00h. Reads from register 0Ch function in a similar way. All three gamma RAMs may be selected for simultaneously writing to them. Reading of the RAMs, however, should be done one at a time.[0156]

Test register A—0Eh. Select test modes for Argus.[0157]

TABLE 27


Bit	Name	Access	Default	Description

0	INCONT	R/W	0	Configures all input pins for
				continuity test.
1	OCONT	R/W	0	Configures all bidirectional pins
				for input continuity test.
2	RAMWR	R/W	0	Enables writing of COL_SEP
				block input data into FIFO RAM.
3-7	Reserved

Test register B—0Fh. Specifies which test point in DSPS is observable on the output port of DSPS.[0158]

TABLE 28


TEST REGISTER B (OF hex)

	Bits [3-0]	Observation Point

	0000	Normal Y, U, V output
	0001	FIFO RAM output
	0010	Color separation Y output
	0011	Color separation Cr output
	0100	Color separation Cb output
	0101	Gamma corrected R output
	0110	Gamma corrected G output
	0111	Gamma corrected B output
	1000	R gamma RAM output
	1001	G gamma RAM output
	1010	B gamma RAM output
	1011-1111	Unused

Color conversion coefficients—10h-18h. These represent the nine coefficients for the 3×3 matrix used to convert from YCrCb to RGB.[0159]
TABLE 29
Bit Name Access Default Description
0 PDN_DIG R/W 0 Enable power down mode of
digital chip.
1-3 Reserved R 0
4 PDN_ANA R/W 0 Enable power down mode of
analog chip.
5-7 Reserved R 0
Test enable—21h.[0160]
TABLE 30
Bit Name Access Default Description
0 TSTEN_DIGT R/W 0 Enable test mode of digital
chip.
1-3 Reserved R 0
4 TSTEN_ANA R/W 0 Enable test mode of analog
chip.
5-7 Reserved R 0
Test modes. The DSPS system, according to the present invention, has several test modes which allow controllability and observability of internal circuits. Most of the tests are enabled by TESTPINB (pin) and TESTENB (register bit). In addition, a specific test bit corresponding to the desired test must be enabled by setting the appropriate test register bit. However, two of the tests, (I) parallel access to 12C registers, and (ii) parallel access to gamma RAM, are enabled only through pins. This also allows direct loading of the registers and gamma RAM for a non-PC, stand-alone-type application. The test modes are listed below:[0161]
1. Ability to observe selected nodes (through test register 0Fh) on DOUT bus.[0162]
2. Ability to write and read all 12C registers and gamma RAM through a parallel port composed of DIN pins operated in bidirectional mode.[0163]
3. Ability to write to FIFO RAM data that is input to COL SEP block and also directly observe output from RAM.[0164]
4. Input and output pins can be configured for continuity test through[0165]bits0 and1 of register 0Eh.

Parallel access of 12C registers and gamma RAM. Argus allows parallel access of I2C registers and gamma RAM for testing purposes.[0166]

TABLE 31


PARALLEL ACCESS OF REGISTERS IN TEST MODE

Normal	Redefined
pin	pin
function	function	Type	Description

DIN [7-0]	ADDRS [7-	1	Register address
	0]
DIN8	SELREG		1	Select register address
			space
DIN9	SELRAM
	1	Select gamma RAM address
			space
DOUT9	RDB
	1	Parallel mode readstrobe
DOUT8	WRB
	1	Parallel mode write
			strobe
DOUT [7-	TDAT	B	parallel data
0]
TPARENB	TPARENB		1	Enables parallel access.
			TSTPINB must also be
			enabled.

Data is written into the I2C registers by the rising edge of a WRB provided address which is already valid. Valid output data appears on TDAT corresponding to the input address when RDB becomes active low.[0167]

TABLE 32


PIN LIST FOR DSPS

		Signal	Interface
Pin #	Signal Name	Type	Type	Description

1	4BYTMODE	1	CMOS	Sets theArgus slave
				12C interface in
				4 bytes per write
				transaction mode.
2	DOUT19
3	DOUT18
4	DOUT17
5	DOUT16
6	DOUT15
7	DOUT14	0	CMOS 16 mA	4:2:2 YUV data
				output.
8	DOUT13
9	DOUT12
10	GND2
11	VDD2
12	DOUT11
13	DOUT10
14	DOUT9
15	DOUT8
16	DOUT7
17	DOUT6
18	DOUT5
19	DOUT4
20	DOUT3
21	GND3
22	VDD3
23	DOUT2
24	DOUT1
25	DOUT0
26	VDD4		Supply
27	GND4		Ground
28	SDAS	B	CMOS O.D.	Slave12C data pin
			8 mA	(bidirectional)
29	SCLS	1	CMOS	Slave	12C bus
				control
30	HREFOB	0	CMOS 8 mA	Output HREF signal
31	VREFOB	0	CMOS 8 mA	Output VREF signal
32	HREFIB	1		Input HREF signal
33	VREFIB	1		Input VREF signal
34	RSTB	1	CMOS Schmitt	Reset pin; may be
				connected to exter-
				nal power-on-
				reset circuit.
35	DIN9	B	CMOS	4 mA	Digital video input
36	DIN8			data from Gorgon.
37	DIN7			Data is in mosaic
38	DIN6			form encoded in
39	DIN5			CCIR 656 like
				format.
				Also used as parallel
				I/O bus for access-
				ing 12C registers
				and gamma RAM
				during test mode.
40	GND5		Ground
41	VDD5		Supply
42	DIN4	B	CMOS	4 mA	Digital video input
43	DIN3			data from AFE.
44	DIN2			Data is in mosaic
45	DIN1			form encoded in
46	DIN0			CCIR 656 like
				format.
				Also used as parallel
				I/O bus for access-
				ing 12C registers
				and gamma RAM
				during test mode.
47	SDA	B	CMOS O.D.	Master12C data pin
			8 mA	(bidirectional)
48	SCL	0	CMOS 8mA	Master	12C bus
				control

49		N.C.		Unused - connect to
				ground
50		N.C.		Unused - connect to
				ground
51		N.C.		Unused - connect to
				ground
52		N.C.		Unused - connect to
				ground
53	SCANMODE	1	CMOS	Is tied high during
				scan test. Puts chip
				in scan mode. For
				example, bypasses
				clock buffer and
				reset.
54	INTERP	1	CMOS	Selects interpolation
				mode
55	CLKIN	1	CMOS	Chip input clock =
				2 × pixel clock
56	CLKIN2X	1	CMOS	Chip clock = 2 × or
				2 × 5/4 pixel clock
57	GND1	1	Ground
58	VDD1		Supply
59	CLKOUT	0	CMOS 15mA	2 × or 2 × 5/4 out-
				put data clock.
60	TESTPINB	1	CMOS	Test pin for enabl-
				ing test mode
61	TPARENB	1	CMOS	Enables exterior
				parallel mode access
				to registers and
				gamma RAM.
62	FIELD	1	CMOS	Field indication.
				Changes on the
				fourth horizontal
				line in each field.
				Odd field: FIELD =
				0
				Even field: FIELD =
				1
63	HIZENB	1	CMOS	Test pin for putting
				all output pins in
				Hi-Z mode. Works
				in conjunction with
				HIZDOUT bit.
64	SCENBL	1	CMOS	Controlled by
				ATPG during scan
				test.

In summary, according to the present invention, a signal processing system (SPS) for a video camera includes first and second subsystem modules, an analog front-end (AFE), and a digital signal processing system (DSPS) connected to the analog front-end (AFE). The signal processing system according to the present invention performs signal processing for a CCD camera capable of capturing full motion video. A complete digital video camera according to the present invention additionally includes a CCD array, a high voltage CCD driver, and a DC-DC converter. The AFE according to the present invention receives mosaic CCD output (a stream of cyan, magenta, yellow and green color samples) from a CCD camera, performs analog signal processing, and produces a 10 bit digital output. The DSPS according to the present invention accepts the digital output of the AFE, performs digital processing on the received digital output of the AFE, and, in turn, outputs a CCIR 601 like 4:2:2 YCrCb video data product according to one embodiment.[0168]
The AFE and the DSPS are a cooperative system according to the present invention. Their registers are configured according to the present invention to share the same address spaces, and to an external controller, the combined AFE and DSPS modules appear as a single device. Nonetheless, the AFE and the DSPS according to the present invention can be used as stand-alone systems which can be addressed directly through their respective I2C interfaces.[0169]
The combined signal processing system, according to the present invention, is partitioned into physically separate subsystems, respectively including an AFE and a DSPS, respectively fabricated on separate semiconductor substrates to enable the installation of the AFE and its analog functions with a CCD camera package jointly with the CCD imager, and to enable digital data processing to be accomplished either in the camera package itself or separately in a personal computer or other data processing system. Thus, digital and analog signal processing functions are adaptively localizable and delocalizable in accordance with application package requirements. By establishing distributable analog and digital functionalities, the entire CCD signal processing functionality can be localized and fabricated in silicon. Alternatively, the analog functions can be fabricated in silicon while the digital functions can be software implemented. The development of a separate analog and digital modules, according to the present invention, permits convenient retrofitting with advanced analog or digital designs. Further, the analog and digital subsystems are operable at different data rates. Thus, the output of an analog ADC is more than 2× lower than the 4:2:2 YCrCb format output from the digital section. By partitioning into separate modules, the data transmitted from the analog module to the digital module can be provided at a reduced interfacing rate between the analog and digital signal processing domains. According to the present invention, an analog front end provides a CCD interface which is scalable for a plurality of selected CCD output formats and pixel rates. In particular, the timing signals and clocks such as horizontal and vertical shift register clocks, applicable CCD output sampling pulses, and the number of horizontal and vertical pixels per frame are programmable. Further, according to the present invention, the starting pixel type is programmable in a predetermined register.[0170]
According to the present invention, a signal processing system (SPS) for an imager device includes a camera system for producing an imager signal, a correlated double sample (CDS) circuit for receiving data from an imager, a variable gain amplifier (VGA), an analog-to-digital converter (ADC) coupled to the VGA circuit, a digital gain circuit (DGC) coupled to the ADC, and an automatic gain control (AGC) circuit for controlling the CDS circuit and the DGC.[0171]
According to the present invention, gain control of a signal processing system for an imager device includes a correlated double sampler, a variable gain amplifier circuit for receiving data from the imaging system, and an automatic gain control (AGC) circuit for controlling the level of gain provided to the camera system and a correlated double sampler and variable gain amplifier (CDSVGA) circuit according to the present invention.[0172]
According to the present invention, signal processing circuitry for a video camera has first and second data processing subsystems including an analog front-end (AFE) and a digital signal processing system (DSPS) connected to the analog front-end (AFE). The signal processing system according to the present invention performs signal processing function for a low cost CCD or CMOS imaging camera capable of capturing full motion video. A complete digital video camera according to one embodiment of the present invention includes in part a sensor array, a high voltage CCD driver, and a DC-DC converter. The AFE data processing subsystem according to the present invention receives a mosaic CCD output (a stream combining cyan, magenta, yellow and green color samples) from the CCD camera, performs analog signal processing, and produces a digital output which can be converted to a form suitable for image display. The DSPS data processing subsystem according to the present invention accepts the digital output of the AFE data processing subsystem, performs digital processing on the received digital output of the AFE, and according to one embodiment of the present invention outputs a CCIR 601 4:2:2 YCrCb video data product suitable for presentation on a user selected display.[0173]
The AFE and DSPS data processing subsystems can be implemented as a cooperative chipset according to the present invention. The respective data processing subsystems include registers which are configured according to the present invention to share a common address space of the respective subsystems. To an external controller, the combined AFE and DSPS subsystems are operable as a unitary data processing system. Additionally, the separate AFE and the DSPS of the respective subsystems according to the present invention can be used as stand-alone units which can be addressed and controlled directly through respective I2C interfaces.[0174]
The SPS according to the present invention is thus partitionable into physically separate subsystems which can individually be fabricated on separate semiconductor substrates to enable the combined installation of the AFE and its analog functions in an integrated camera package jointly operable with an imager. This enables digital data processing to be accomplished either in the camera package itself or separately in a personal computer or other data processing system to transform the analog pixel sample into display-ready digital data format. Accordingly, digital and analog signal processing functions are adaptively localizable and delocalizable in accordance with application package requirements. By establishing distributable analog and digital functionalities, the entire imager signal processing functionality can be localized and fabricated in silicon. Alternatively, the analog functions can be fabricated in silicon while the digital functions are software implemented. The development of separate analog and digital modules accordingly permits convenient system retrofitting with advanced analog or digital designs. Further, the analog and digital subsystems are operable at different data rates. Thus, the output of an analog to digital (A/D) converter within the AFE subsystem can be 2× lower than the 4:2:2 YCrCb format output from the digital section according to one embodiment of the present invention. By partitioning the subsystems into separate modules, the data transmitted from the analog module subsystem to the digital module subsystem can be provided at a reduced interfacing load between the analog and digital signal processing domains creating processing efficiencies. According to the present invention, the AFE subsystem provides an imager interface which is scalable for a plurality of selected imager output formats and pixel rates. In particular, the timing signals and clocks such as horizontal and vertical shift register clocks, the applicable imager output sampling pulses, and the number of horizontal and vertical pixels per frame are user programmable according to the present invention.[0175]
Further according to the present invention, independent and dependent (through the DSPS) register addressing is enabled and timing generator parameters are programmable. An automatic gain control circuit allows for up to 98 dB of gain range including shutter exposure with an option for flickerless operation that is enabled with a hysteresis method according to the present invention.[0176]