CN102624978A - Rotatable multi-camera mobile phone and three-dimensional camera implementation method based on sound source recognition and positioning - Google Patents

Abstract

The invention discloses a rotatable multi-camera mobile phone, which comprises a mobile phone body, a baseband chip, at least two cameras, a voice control identification and positioning module, a camera rotation driving module, a three-dimensional image processing module and a display module, wherein the baseband chip, the at least two cameras, the voice control identification and positioning module, the camera rotation driving module, the three-dimensional image processing module and the display module are arranged in the mobile phone body, the voice control identification and positioning module is used for positioning a sound source coordinate, at least two image signals acquired by the at least two cameras are integrated and processed by the three-dimensional image processing module to generate a simulated three-dimensional image signal and are transmitted to the display. The invention can realize three-dimensional shooting, three-dimensional video call and three-dimensional self-shooting at any time and any place, has the function of using a mobile phone screen as a mirror, and can realize speaker orientation in video conferences and on-site interviews through sound control.

Description

Rotatable multi-cam mobile phone and the three-dimensional camera shooting implementation method of locating based on identification of sound source

(1) technical field:

The present invention relates to the mobile communication equipment technical field, the three-dimensional camera shooting implementation method that is specifically related to rotatable multi-cam mobile phone and locatees based on identification of sound source.

(2) background technology:

Along with continuous progress in science and technology, the development of electronic industry is like a raging fire, particularly the mobile communication equipment field.Mobile phone not merely is a kind of simple communication tool, only need realize the ability What for, but will satisfy the ultimate attainment user experience of consumer.Now, most mobile phone all has only a camera, can't realize so 3D takes.Simultaneously most of mobile phones also all be an embedded camera at the mobile phone back side, and this camera can not rotate to the mobile phone front, and video calling also just can't be realized like this.Mobile phone also need overturn during auto heterodyne so that camera is aimed at oneself, and this has brought inconvenience in operation.Once more, in occasions such as some spot coverage, videoconference, video conferences, the mobile phone that needs constantly to overturn switches localizing objects, except that operation inconvenience, also possibly reveal like this and clap some timely information.Except these, existing a lot of women like working as mirror to mobile phone screen to be made, and can have a few money mobile phones to accomplish such effect, the way that What is more has proposed on mobile phone, to inlay mirror, the weight of the increase mobile phone that this meeting is meaningless actually.

Drafting (Image Based Rendering IBR) based on image is emerging over past ten a years subject.It can utilize a series of known reference pictures to synthesize the image of new viewpoint.With the traditional drafting based on three-dimensional modeling (3D Model-Based Rendering MBR) compared with techniques, it has does not need complicated three-dimensional scenic modeling, and render speed is fast, to advantages such as computer are less demanding.But utilize known image to synthesize new images merely, required image data processed is very huge, for example the light field rendering technique among the IBR.And the image re-projection that grew up in recent years technology has reduced the data volume of handling owing to introduced inside and outside parameter of video camera and depth information, has accelerated render speed, is a special kind of skill more rising among the IBR.

(3) summary of the invention:

In order to overcome the deficiency of prior art; The objective of the invention is to: but the multi-cam multi-functional mobile phone and the three-dimensional camera shooting implementation method thereof that provide a kind of built-in acoustic control cam to rotate; Realize three-dimensional the shooting; 3 D video conversation and three-dimensional whenever and wherever possible the auto heterodyne have the function that mobile phone screen uses when mirror, can be through purposes such as person's orientations in acoustic control realization video conference, the spot coverage.

The present invention realizes through following technical scheme:

Rotatable multi-cam mobile phone comprises mobile phone body, also comprises the baseband chip of being located in this mobile phone body; At least two cameras, acoustic control identification positioning module, camera rotation driver module; 3-D view processing module and display module; Said baseband chip rotates the driver module control connection with acoustic control identification positioning module and camera respectively, and this acoustic control identification positioning module is used for the localization of sound source coordinate, and the sound source coordinate signal is sent to camera rotation driver module through baseband chip; Camera rotation driver module is used to drive camera and rotates to Sounnd source direction; And obtain the picture signal of Sounnd source direction, at least two picture signals that at least two cameras obtain generate the 3-D view signal of simulation via 3-D view processing module integration processing; And being sent to display module, this display module is used to show the 3-D view signal that receives.

As the improvement of technique scheme, further technical scheme of the present invention is following:

Further; Above-mentioned camera rotation driver module further comprises first magnetic coil of being located on the camera and second magnetic coil that acts on this first magnetic coil; Described camera rotatably is located in the mobile phone body, and this first magnetic coil drives the camera rotation under the magnetic drive of second magnetic coil.

Outside camera, be wound with first magnetic coil, energising back camera just is equivalent to a magnet.When first magnetic coil passed to set directional current, according to right-hand screw rule, four referred to bend towards the sense of current of first magnetic coil, and that end of thumb indication is exactly the N extreme direction of first magnetic coil.The N utmost point of supposing camera is positioned at camera end.

Just be equivalent to another magnet after the energising of second magnetic coil.When second magnetic coil passes to the first direction electric current, utilize right-hand screw rule: the right hand is held second magnetic coil, and four refer to bend towards the sense of current of second magnetic coil, and that end of thumb indication is exactly the N extreme direction of second magnetic coil.The N utmost point of supposing second magnetic coil is positioned at mobile phone protecgulum end.In like manner, if second magnetic coil passes to the second direction electric current, then the N utmost point of second magnetic coil is positioned at the cell phone rear cover end.

All can produce magnetic field after first magnetic coil and the energising of second magnetic coil, according to the principle that the same sex is repelled each other, there is a natural attraction between the sexes, the N utmost point of camera can point to the S utmost point of second magnetic coil, and the S utmost point of camera can point to the N utmost point of second magnetic coil.

Further, above-mentioned acoustic control identification positioning module comprises sound collection means and sound source identification positioning unit, and said sound collection means is the built-in mobile phone microphone.

This identification of sound source locating module adopts the auditory localization technology based on microphone array, and the time difference that arrives two microphones in the microphone array according to sound is carried out DOA estimation, thus the localization of sound source position.

The auditory localization algorithm of difference mainly comprises two steps based on the time of advent:

(1) advanced line delay is estimated, therefrom obtain in the microphone array two array elements between sound arrive time delay.

(2) utilize time-delay to estimate to carry out DOA estimation.

The structure flow process of this algorithm is following:

(1) at first obtains speaker's voice signal by first microphone and second microphone; Again through A/D sampling (transferring analog signal to digital signal) and low pass filter; Obtain voice signal to be imported at last, can be designated as X1 (n) and X2 (n) respectively, n is the time.

(2) bandwidth of voice signal is 0.3～3.4kHz, thereby needs the noise outside band pass filter filtering speech signal bandwidth of design.Promptly keep linear phase for the signal phase of handling is not changed, need to adopt the FIR filter.Behind the FIR band pass filter, can obtain X1w (n) and X2w (n) with half overlapping Hamming window to X1 (n) and X2 (n) windowing, through the FFT Fourier transform, convert signal into frequency domain from time domain then, the crosspower spectrum that can try to achieve X1 (n) and X2 (n) then does

P₁₂(k)＝X_1w(k)×X_2w(k)

In order to weaken the influence of noise and reverberation, can carry out frequency domain weighting.Power spectrum does after the weighting

C₁₂(k)＝ψ₁₂(k)×P₁₂(k)

(3) be further outstanding peak value, behind frequency domain weighting, can be level and smooth to the crosspower spectrum between microphone signal, obtain

$a_{12}^m=\left\{\begin{array}{c}{\mathrm{<figure-callout\; label="C"\; filenames="HDA0000139602850000111.png,HDA0000139602850000121.png"\; state="\{\{state\}\}">C</figure-callout>}}_{12}^1,m=1\\ a_{12}^{m-1}+{\mathrm{<figure-callout\; label="C"\; filenames="HDA0000139602850000111.png,HDA0000139602850000121.png"\; state="\{\{state\}\}">C</figure-callout>}}_{12}^1,m>1\end{array}\right.$

Wherein, m representes the level and smooth frame number that adds up.

(4) inversefouriertransform is asked cross-correlation function, and its peak value is exactly the time delay between microphone 1 and 2

(5) according to the angular distance positioning mode, obtain sound source position.The horizontal angle of the relative initial point of sound source does

$\mathrm{\&theta;}={\cos }^{-1}\left(\frac{d}{2a}\right)$

The elevation angle of the relative initial point of sound source does

$\mathrm{\&phi;}={\cos }^{-1}\left(\frac{d}{a}\right)$

Wherein, a be microphone to spacing, d is that sound source arrives the right range difference of microphone.

The d=sound source is to the aerial propagation velocity of time delay * sound of microphone 1 and 2.

Based on the three-dimensional camera shooting implementation method of identification of sound source location, this method comprises the steps:

(1) two camera is taken Same Scene in different points of view and is obtained two width of cloth reference pictures;

(2) utilize the corresponding camera inner parameter matrix of first width of cloth reference picture and the spin matrix of target image to calculate homography matrix from first width of cloth reference picture to target image;

(3) utilize homography matrix that first width of cloth reference picture is projected to target image, thus the part of synthetic target image;

(4) order of definite traversal second width of cloth reference picture pixel;

(5) traversal order of confirming according to step (4) travels through the pixel of second width of cloth reference picture; Utilize 3-D view transformation equation group that these pixels are projected to target image; If this pixel has been projected to the target image zone partly that step (3) is synthesized, then projection image's vegetarian refreshments is left intact, otherwise directly copies this pixel point value in second width of cloth reference picture to the corresponding projection image of target image vegetarian refreshments; Obtain the remainder of target image, thus synthetic target image.

Beneficial effect of the present invention is:

1. the present invention adopts the auditory localization technology to respond to the sound source; Automatically the synchronous rotation of control dual camera; Can be applied to some occasions such as video conference, videoconference, spot coverage easily,, make it just turn to speaker's direction at dialect through auditory localization technology control camera; Thereby each picture that need take under can shooting has promptly and accurately saved the unnecessary trouble that the upset mobile phone is located photographic subjects.

2. the present invention adopts dual camera to make the 3D shooting be achieved at hardware aspect.

3. when the present invention is through acoustic control Rotating Double camera to mobile phone front, can realize that 3D autodynes whenever and wherever possible, the 3D video calling lets communication become and is full of enjoyment, and is on the spot in person, do not have distance.Also realized simultaneously working as the function that mirror makes to mobile phone screen, saved the women and gone out, brought many facilities for people's life with the trouble of mirror.

(4) description of drawings:

Fig. 1 is the partial structurtes sketch map of front of the present invention dual camera;

Fig. 2 is the partial structurtes sketch map of the back side of the present invention dual camera;

Fig. 3 is the partial structurtes sketch map of side press-key of the present invention;

Fig. 4 is that the module of acoustic control dual camera rotation of the present invention connects sketch map;

Fig. 5 is the flow chart of acoustic control dual camera rotation of the present invention;

Fig. 6 is a 3D implementation structure sketch map of the present invention;

Fig. 7 is the rotational structure sketch map of camera first angle of the present invention;

Fig. 8 is the rotational structure sketch map of camera second angle of the present invention;

Fig. 9 is the rotational structure sketch map of camera angular of the present invention;

Figure 10 is the algorithm structure flow chart of acoustic control of the present invention location identification module;

Figure 11 is two width of cloth reference pictures of two cameras shootings and the projection relation sketch map of target image, and wherein I1 and I2 are two width of cloth reference pictures, and Is is a target image;

Figure 12 is the building-up process sketch map of first width of cloth reference picture to target image;

Figure 13 is the building-up process sketch map of second width of cloth reference picture to target image;

Figure 14 is the hole-filling sketch map of target image.

(5) embodiment:

Below in conjunction with accompanying drawing the present invention is done detailed explanation, but not as to qualification of the present invention:

Fig. 1, Fig. 2, Fig. 3 are present embodiment design structurally.The present invention structurally mainly is made up of mobile phone body 1, the first camera 2-1, the second camera 2-2, the button 3 of closing/opening the acoustic control function.

Fig. 4, Fig. 5, Fig. 6 are the structure chart of present embodiment on function.Present embodiment main acoustic control cam of realizing on function rotates, and 3D takes, autodyne, and video calling, screen is when functions such as mirror make.

It is as shown in Figure 4 that the module of acoustic control dual camera rotation connects sketch map, and acoustic control dual camera rotary module mainly is made up of sound collection means, identification of sound source locating module, built-in mobile phone baseband chip, camera rotation driver module and mobile phone dual camera.Wherein said sound collection means is the built-in mobile phone microphone.Said dual camera is an embedded in mobile phone formula camera lens.Said baseband chip is connected with camera rotation driver module with acoustic control identification positioning module respectively.

The flow chart of acoustic control dual camera rotation is as shown in Figure 5.At first with mobile phone power-on and adjust mobile phone, open the acoustic control functional switch among Fig. 3 then to screening-mode.Then the sound collection means in the mobile phone is that embedded in mobile phone formula microphone begins to collect sound, and the voice data of collecting is sent to the identification of sound source locating module, and identification of sound source locating module location sound source sends to camera rotation driver module with the result.When the location sound source was the mobile phone frontal, camera rotation driver module drove camera and rotates to the mobile phone front.When the location sound source was mobile phone back side direction, camera rotation driver module drove camera and rotates to the mobile phone back side.Thereby realize the acoustic control cam rotation.In the time need not using the acoustic control cam spinfunction, close the acoustic control functional switch among Fig. 3.

3D technique for taking implementation structure principle is as shown in Figure 6; Dual camera is through take the two dimensional image that obtains two different visual angles from different perspectives; Many two dimensional images are integrated via the 3-D view processing module; Send images to the 3-D view display module through 3D multimedia messages process chip, the 3-D view display module is shown to 3D rendering on the mobile phone screen then, thereby realizes that 3D takes.

When the acoustic control dual camera rotates to the mobile phone front, open shoot function, can realize that just 3D autodynes, the while mobile phone screen also can be worked as mirror have been made.

Open video call function, just can realize the 3D video calling.

Like Fig. 7, Fig. 8, shown in Figure 9, the rotation principle of the camera 2 of present embodiment is following:

Outside camera 2, be wound with firstmagnetic coil 4, energising back camera 2 just is equivalent to a magnet.When firstmagnetic coil 4 passed to set directional current, according to right-hand screw rule, four referred to bend towards the sense of current of firstmagnetic coil 4, and that end of thumb indication is exactly the N extreme direction of first magnetic coil 4.In this example, the N utmost point of camera 2 is positioned at camera end.

Just be equivalent to another magnet after 5 energisings of second magnetic coil.When second magnetic coil 5 passes to the first direction electric current, utilize right-hand screw rule: the right hand is held second magnetic coil, and four refer to bend towards the sense of current of second magnetic coil 5, and that end of thumb indication is exactly the N extreme direction of second magnetic coil 5.In this example, the N utmost point of second magnetic coil 5 is positioned at mobile phone protecgulum 1-1 end.In like manner, if second magnetic coil 5 passes to the second direction electric current, then the N utmost point of second magnetic coil 5 is positioned at cell phone rear cover 1-2 end.

All can produce magnetic field after first magnetic coil and the energising of second magnetic coil, according to the principle that the same sex is repelled each other, there is a natural attraction between the sexes, the N utmost point of camera 2 can point to the S utmost point of second magnetic coil 5, and the S utmost point of camera 2 can point to the N utmost point of second magnetic coil.

The realization principle of present embodiment identification of sound source locating module is following:

This identification of sound source locating module adopts the auditory localization technology based on microphone array, and the time difference that arrives two microphones in the microphone array according to sound is carried out DOA estimation, thus the localization of sound source position.

The auditory localization algorithm of difference mainly comprises two steps based on the time of advent:

(1) advanced line delay is estimated, therefrom obtain in the microphone array two array elements between sound arrive time delay.

(2) utilize time-delay to estimate to carry out DOA estimation.

The structure flow chart of this algorithm is shown in figure 10:

(1) at first obtains speaker's voice signal by first microphone and second microphone; Again through A/D sampling (transferring analog signal to digital signal) and low pass filter; Obtain voice signal to be imported at last, can be designated as X1 (n) and X2 (n) respectively, n is the time.

(2) bandwidth of voice signal is 0.3～3.4kHz, thereby needs the noise outside band pass filter filtering speech signal bandwidth of design.Promptly keep linear phase for the signal phase of handling is not changed, need to adopt the FIR filter.Behind the FIR band pass filter, can obtain X1w (n) and X2w (n) with half overlapping Hamming window to X1 (n) and X2 (n) windowing, through the FFT Fourier transform, convert signal into frequency domain from time domain then, the crosspower spectrum that can try to achieve X1 (n) and X2 (n) then does

P₁₂(k)＝X_1w(k)×X_2w(k)

In order to weaken the influence of noise and reverberation, can carry out frequency domain weighting.Power spectrum does after the weighting

C₁₂(k)＝ψ₁₂(k)×P₁₂(k)

(3) be further outstanding peak value, behind frequency domain weighting, can be level and smooth to the crosspower spectrum between microphone signal, obtain

$a_{12}^m=\left\{\begin{array}{c}{\mathrm{<figure-callout\; label="C"\; filenames="HDA0000139602850000111.png,HDA0000139602850000121.png"\; state="\{\{state\}\}">C</figure-callout>}}_{12}^1,m=1\\ a_{12}^{m-1}<figure-callout\; label="+\; C"\; filenames="HDA0000139602850000111.png,HDA0000139602850000121.png"\; state="\{\{state\}\}">+</figure-callout>C_{}^{}{}_{12}^1,m>1\end{array}\right.$

Wherein, m representes the level and smooth frame number that adds up.

(4) inversefouriertransform is asked cross-correlation function, and its peak value is exactly the time delay between microphone 1 and 2

(5) according to the angular distance positioning mode, obtain sound source position.The horizontal angle of the relative initial point of sound source does

$\mathrm{\&theta;}={\cos }^{-1}\left(\frac{d}{2a}\right)$

The elevation angle of the relative initial point of sound source does

$\mathrm{\&phi;}={\cos }^{-1}\left(\frac{d}{a}\right)$

Wherein, a be microphone to spacing, d is that sound source arrives the right range difference of microphone.

The d=sound source is to the aerial propagation velocity of time delay * sound of microphone 1 and 2.

The realization principle of camera rotation driver module is following in the present embodiment:

1. according to the result of identification of sound source locating module, if sound source is positioned at mobile phone protecgulum one side, pass to second direction electric current among Fig. 9 then for second magnetic coil 5, repel each other according to the same sex, principle of opposite sex attraction, mobile phone cam rotate to the mobile phone protecgulum.

2. according to the result of identification of sound source locating module,, pass to first direction electric current among Fig. 8 then for second magnetic coil 5, repel each other according to the same sex if sound source is positioned at cell phone rear cover one side, principle of opposite sex attraction, mobile phone cam rotates to cell phone rear cover.

The realization principle of 3-D view processing module is following in the present embodiment:

This module adopts a kind of 3-D view synthetic method based on image re-projection.This method is taken Same Scene with two cameras in different points of view and is obtained two width of cloth reference pictures; At first utilize the part of the synthetic target image of a homography matrix and a secondary reference picture; Utilize the 3-D view converter technique then; Remainder according to another width of cloth reference picture generation target image obtains target image.

This method may further comprise the steps:

(1) two camera is taken Same Scene in different points of view and is obtained two width of cloth reference pictures;

(2) utilize the corresponding camera inner parameter matrix of first width of cloth reference picture and the spin matrix of target image to calculate homography matrix from first width of cloth reference picture to target image;

(3) utilize homography matrix that first width of cloth reference picture is projected to target image, thus the part of synthetic target image;

(4) order of definite traversal second width of cloth reference picture pixel;

(5) traversal order of confirming according to step (4) travels through the pixel of second width of cloth reference picture; Utilize 3-D view transformation equation group that these pixels are projected to target image; If this pixel has been projected to the target image zone partly that step (6) is synthesized, then projection image's vegetarian refreshments is left intact, otherwise directly copies this pixel point value in second width of cloth reference picture to the corresponding projection image of target image vegetarian refreshments; Obtain the remainder of target image, thus synthetic target image.

The practical implementation step is following:

Figure 11 is two width of cloth reference pictures of two cameras shootings and the projection relation sketch map of target image.Wherein I1 and I2 are two width of cloth reference pictures, and Is is a target image.

Figure 12 is the building-up process sketch map of first width of cloth reference picture to target image.

Figure 13 is the building-up process sketch map of second width of cloth reference picture to target image.

Figure 14 is the hole-filling sketch map of target image.

Two cameras are taken Same Scene in different points of view and are obtained two width of cloth reference picture I1 and I2, and photocentre is respectively C1 and C2; The inner parameter matrix of the camera that first width of cloth reference picture I1 is corresponding is K1, and spin matrix is R2; The inner parameter matrix of the camera that second width of cloth reference picture I2 is corresponding is K2, and spin matrix is R2.The target image Is that synthesizes is that first width of cloth reference picture I1 rotates resulting image around its photocentre C1, and establishing its angle of rotation is θ, and rotating shaft is the Y axle.U is in the scene a bit, and its standardization homogeneous coordinates are [X_wY_wZ_w1]^T, U corresponding pixel in second width of cloth reference picture is u, and pixel corresponding in target image is Us, and the depth value that pixel u is corresponding is Zu.

(1) the spin matrix R of calculating target image Is_θ

(2) in this example, that the pixel coordinate of first width of cloth reference picture I1 system adopts is x ' o ' y ', is xoy and the coordinate of the target image that obtains at last adopts, therefore needs to calculate from xoy to x ' the transformation matrix of coordinates T of o ' y '_{Oo '}, in this example

$T_{{\mathrm{oo}}^{\&prime;}}=\begin{array}{ccc}1& 0& 0\\ 0& -1& 767\\ 0& 0& 1\end{array}$

(3) utilize the inner parameter matrix K 1 of the corresponding camera of first width of cloth reference picture, the spin matrix R of target image_θAnd transformation matrix of coordinates T_{Oo '}, calculate the homography matrix M of first width of cloth reference picture I1 to target image Is.

$M=T_{{\mathrm{oo}}^{\&prime;}}^{-1}{K}_1{R}_0{K}_1^{-1}{T}_{{\mathrm{oo}}^{\&prime;}}$

(4) apex coordinate and the homography matrix M with first width of cloth reference picture I1 multiplies each other, and gets the set Verts of the projection in target image Is to the limit_s, this gathers Verts_sBe point set.Calculate the closed polygon zone Polygon that the projection of first width of cloth reference picture in target image Is constitutes according to the border of this point set and target image Is.

(5) by line scanning polygonal region Polygon; Utilize homography matrix M to calculate the corresponding points of each o'clock in first width of cloth reference picture I1; This corresponding points corresponding pixel value is copied to the position of corresponding points among the polygonal region Polygon; Thereby first width of cloth reference picture I1 has been projected among the target image Is, synthesized the part of target image.

(6) the projection e of photocentre C1 in second width of cloth reference picture I2 of calculating first width of cloth reference picture I1, e=[e_x, e_y, e_z]

(7) coordinate according to projection e is divided into several zones with second width of cloth reference picture I2.By e_zEach zone of positive and negative definite traversal in the order of pixel.If e_z＞0, then this limit is positive limit, should be according to the pixel of handling towards the direction of projection e among second width of cloth reference picture I2; Otherwise, should be according to the pixel of handling away from the direction of projection e among second width of cloth reference picture I2.

(8) traversal order of confirming according to a last step travels through each the pixel u in the individual zone that second width of cloth reference picture I2 marks off, and utilizes 3-D view transformation equation group, promptly

These pixels are projected among the target image Is, thereby obtain target projection point Us.Wherein Zs is the scale factor that the vegetarian refreshments Us of projection image is become the standardization homogeneous coordinates.

(9) adopt " jumping fence " vegetarian refreshments Us of algorithm computation projection image whether outside the polygonal region Polygon in target image Is.If, then the pixel value with the pixel u among second width of cloth reference picture I2 copies the vegetarian refreshments Us of projection image to; If do not exist, then do not do any operation.So just obtain the remainder of target image, synthesized target image Is.

(10) next be that hole-filling is carried out in the outer zone of the polygon Polygon among the target image Is, like Figure 14.

(11) for 1 P in the cavity, four neighbor pixel P1 up and down that detection range P is nearest, Pr, Pu, Pd adopts the pixel value of following formula calculation level P to get estimated value Ep

Ep＝(V(Pl)/dl+V(Pr)/dr+V(Pu)/du+V(Pd)/dd)/(1/dl+1/dr+1/du+dd)

Wherein, V (.) expression is got. pixel value

(12) come filling cavity point P with Ep

If also have empty point in the target image, then continue to fill in order to last method.

The foregoing description is a preferred embodiments of the present invention; But execution mode of the present invention is not limited thereto; Other are any not to deviate from change, the modification done under spirit of the present invention and the principle, substitute, combination, simplify; Be the substitute mode of equivalence, be included within protection scope of the present invention.

Claims (4)

1. rotatable multi-cam mobile phone comprises mobile phone body, it is characterized in that: also comprise the baseband chip of being located in this mobile phone body; At least two cameras, acoustic control identification positioning module, camera rotation driver module; 3-D view processing module and display module; Said baseband chip rotates the driver module control connection with acoustic control identification positioning module and camera respectively, and this acoustic control identification positioning module is used for the localization of sound source coordinate, and the sound source coordinate signal is sent to camera rotation driver module through baseband chip; Camera rotation driver module is used to drive camera and rotates to Sounnd source direction; And obtain the picture signal of Sounnd source direction, at least two picture signals that at least two cameras obtain generate the 3-D view signal of simulation via 3-D view processing module integration processing; And being sent to display module, this display module is used to show the 3-D view signal that receives.

2. rotatable multi-cam mobile phone according to claim 1; It is characterized in that: described camera rotation driver module further comprises first magnetic coil of being located on the camera and second magnetic coil that acts on this first magnetic coil; Described camera rotatably is located in the mobile phone body, and this first magnetic coil drives the camera rotation under the magnetic drive of second magnetic coil.

3. rotatable multi-cam mobile phone according to claim 1 is characterized in that: said acoustic control identification positioning module comprises sound collection means and sound source identification positioning unit, and said sound collection means is the built-in mobile phone microphone.

4. based on the three-dimensional camera shooting implementation method of identification of sound source location, it is characterized in that this method comprises the steps:

(1) two camera is taken Same Scene in different points of view and is obtained two width of cloth reference pictures;

(2) utilize the corresponding camera inner parameter matrix of first width of cloth reference picture and the spin matrix of target image to calculate homography matrix from first width of cloth reference picture to target image;

(3) utilize homography matrix that first width of cloth reference picture is projected to target image, thus the part of synthetic target image;

(4) order of definite traversal second width of cloth reference picture pixel;

(5) traversal order of confirming according to step (4) travels through the pixel of second width of cloth reference picture; Utilize 3-D view transformation equation group that these pixels are projected to target image; If this pixel has been projected to the target image zone partly that step (3) is synthesized, then projection image's vegetarian refreshments is left intact, otherwise directly copies this pixel point value in second width of cloth reference picture to the corresponding projection image of target image vegetarian refreshments; Obtain the remainder of target image, thus synthetic target image.

Images