CN101270981A - Material position measuring method and device based on machine vision - Google Patents

Abstract

The invention relates to a high reliability method and device for continuously measuring material level of the little grainy material based on machine vision, which belongs to the technology field of material level measurement. The device comprises a special point light source, an image harvester, and a machine vision module, etc. High brightness optical circle image on the material surface formed by light beam emitted from point light source is harvested by the image harvester, and analyzed and treated by verification type optical circular detection algorithm of the machine vision module, to calculate the diameter or area size of the optical circle to obtain material level value. The material level value is converted to 4-20mA signal by D/A to be output, while the value is encoded with compressed field JPEG image to form serial signal for output. In addition, the original image signal harvested by the image harvester also can be output.

Description

A kind of material level measuring method and device based on machine vision

Technical field

The present invention relates to a kind of high confidence level method and device thereof of the contactless continuous coverage little grainy material material level based on machine vision, belong to the technical field of level gauging.

Background technology

The material that generally needs to measure material level all leaves in the container, and this container is tens meters high cylindrical body, cone or half half cone shape mixture often.Along with material is sent into container continuously, material is more and more nearer apart from the distance of container top, feeds in raw material as untimely stopping, and just might overflow.Measurement to material level at present has several different methods, as the radio frequency admittance level meter, uses the radio frequency capacitance technology, a radio frequency is applied on the probe, by the variation of continuous analysis electric capacity, definite influence that is caused by surrounding environment is a kind of contact level meter; Ultrasonic material level meter is used the mistiming of measuring the ultrasonic pulse transmission and receiving and is calculated object distance; Radar levelmeter is used light velocity electromagnetic wave propagation, according to Time Domain Reflectometry principle or Continuous Wave with frequency modulation principle, and the distance between measurement and the tested interface.Although principle, precision and application scenario that they are measured are not quite similar, but they have a public not enough point, the on-the-spot outdoor scene that when showing material level, can not show material exactly, therefore in case when being subjected to certain influence and causing material level information inaccurate, will mislead operating personnel, accidents such as material overflows take place, some accident even can cause serious consequence.The coal bunker of fuel-burning power plant for example, as Fig. 1, it is the container of upper semi-body tubular lower semi-body taper, speed with nearly ten tons of per minutes adds the storehouse, each position in storehouse can hold nearly hundred tons of coals, in case level-sensing device breaks down, operating personnel find one minute every night, just meaning has nearly ten tons of coals to overflow, and tends to cause major accident.

Chinese patent application number 01107328.4, open day 2002.11.6, a kind of digital video frequency level indicator is disclosed, by video input unit, image pick-up card, microprocessor and display screen etc. interconnect and form, video input unit is aimed at tested material, video input unit is connected to the video inputs of image pick-up card again, microprocessor control image pick-up card is gathered the material level image by the sampling period that the user sets, and this image handled, the material level coordinate figure that analyzes and the calibration gauge of demarcation are relatively calculated the material level actual value, simultaneously with the material level image, the variation tendency of material level actual value and this value is presented on the display screen.Its main means of measuring material level are to utilize the reading on the calibration gauge of material level coordinate points demarcation of living in to obtain the material level actual value, and the present invention utilizes the light circular diameter or the area size that constantly change with the height of level of filler material to carry out level gauging.

Summary of the invention

The objective of the invention is to propose a kind of method of the contactless continuous coverage little grainy material material level based on machine vision and the device that can show the on-the-spot real scene image of material level simultaneously truly.

Technical scheme of the present invention is: adopt a kind of special pointolite, be installed in the top of feed bin, as shown in Figure 2, the light beam that pointolite penetrates is incident upon the material surface, is the taper of isosceles as can be seen from the vertical section of feed bin.And can see that on the level cross-sectionn of differing heights they are light circles of different-diameter.If vertical height is L between A-A xsect and the pointolite₁, xsect glazing radius of a circle is R₁, area is S₁, the vertical height between B-B xsect and the pointolite is L₂, xsect glazing radius of a circle is R₂, area is S₂, vertical height is L between C-C xsect and the pointolite₃, xsect glazing radius of a circle is R₃, area is S₃, the pass between vertical height L and the light radius of a circle R is$\frac{L_1}{R_1}=\frac{L_2}{R_2}=\frac{L_3}{R_3},$Pass between the vertical height L and the light area of a circle is$\frac{L_1}{\sqrt{\frac{S_1}{\mathrm{\π}}}}=\frac{L_2}{\sqrt{\frac{S_2}{\mathrm{\π}}}}=\frac{L_3}{\sqrt{\frac{S_3}{\mathrm{\π}}}},$Promptly$\frac{L_1}{L_2}=\frac{\sqrt{\frac{S_1}{\mathrm{\π}}}}{\sqrt{\frac{S_2}{\mathrm{\π}}}}=\sqrt{\frac{S_1}{S_2}},\frac{L_1}{L_3}=\sqrt{\frac{S_1}{S_3}},$Suppose L₁=1, S₁Be given value A, so$L_3=\sqrt{\frac{S_3}{A}},$That is to say, as long as know the area S of C-C xsect glazing circle₃, just can obtain the vertical range L between C-C xsect and the pointolite₃

Add in the process in storehouse at feed bin, material is upwards piled up by the bottom at leisure, and pointolite forms the light circle of different area on the material surface, and as Fig. 3, along with the accumulation of material, the area of light circle diminishes gradually, as Fig. 4.Therefore record the area of the light circle that forms on the material surface, just can obtain the vertical height between material surface and the pointolite and then obtain level of filler material.

The image acquisition machine is integrated in the pointolite, the image that the light circle that this image acquisition machine collection is formed on the material surface by material and pointolite is formed, because significantly border is arranged between light circle and the material medium, by checking type light circle detection algorithm, be that adaptive threshold is cut apart and this image of region growing algorithm process, detect the area of bright dipping circle; Adopt rim detection and hough transform algorithm to detect light round edge circle simultaneously and obtain the light circular diameter, further verify the confidence level that the light area of a circle or diameter detect, if twice result unanimity, then the explanation detection is accurately credible, otherwise will be filtered; And, also can utilize the confidence level of the roll-off characteristic checking testing result of the light area of a circle or level of filler material because the material in the storehouse is slowly upwards to pile up.

Apparatus for measuring charge level based on machine vision is a kind of embedded system, and block diagram as shown in Figure 5.

In this device, after the image acquisition machine obtains to comprise the Digital Image Processing unit of picture signal in machine vision module of light circle, the analyzing and processing of leading up to is obtained light diameter of a circle or area value, this value obtains the digital quantity of material level through the material level computing unit, and the material level digital quantity converts the output of 4-20mA signal to through D/A.The second road signal of exporting from the Digital Image Processing unit is through the video d/a modular converter, be reduced into video signal output, Third Road passes through the signal encoding unit through the check code of JPEG compression unit compression back and on-the-spot material level digital quantity and generation, form the serial composite signal, with the composite signal of two wires form output material level amount and on-the-spot outdoor scene jpeg image.For example the serial signal transfer rate is 57600b/s, can transmit resolution in 1 second and is about 160 * 120 jpeg image and material level amount signal by calculating.

Description of drawings

Fig. 1 is the coal bunker synoptic diagram of fuel-burning power plant;

Fig. 2 is the principle schematic according to light circular diameter or area acquisition level of filler material;

Fig. 3, Fig. 4 are the coal bunker side views that the light circle changes with the level of filler material height;

Fig. 5 is the apparatus for measuring charge level system construction drawing.

Claims (2)

1. method and apparatus based on the level gauging of machine vision, its characteristic is:

(1) method of this device employing is that special pointolite is installed at the feed bin top, and luminous intensity is big, launches light beam to fixing direction after the line focus and is incident upon the material surface, forms high-luminance light circle on this surface.This light diameter of a circle or area size constantly change with the height of level of filler material.

(2) this device adopts image acquisition machine and the integrated collection of DSP image processor and analyzing and processing by the round image of forming of light that material and material surface form, and obtains light round area or diameter through graphical analysis, obtains the material level value thus.

2. according to 1 described device, in order to improve accuracy and the confidence level that material level detects, proposed checking type light circle detection algorithm, its characteristic is: utilize two or more independently light circle detection algorithms.Detect the light area of a circle and then obtain diameter as cutting apart by adaptive threshold with the region growing algorithm, adopt rim detection and hough transform algorithm to detect light round edge circle simultaneously and obtain the light circular diameter, whether the diameter that two kinds of detection algorithms are obtained compares consistent, and final roll-off characteristic in conjunction with the utilization light area of a circle or level of filler material is verified accuracy and the confidence level that material level detects jointly.

Images