The Research Of The Thickness Detection For Flat Glass Based On Machine Vision Mechanism

The thickness measurement is one of technical indicators for plate glass detection. Tobe more precise and rapid measurement of the thickness for the float glass, non-contactmeasurement methods have become more advanced measuring instruments, while laser isone of the most popular tools in the field of the thickness detection. A precise and rapiddetection system for the thickness measurement of the glass is designed using machinevision mechanism in this paper.Firstly, there is a brief overview, which contains optical properties of the glass andmethods of the thickness detection of the glass by optics principles. Based on the abovemethods, a reasonable thickness detection system of the flat glass is designed.Secondly, the feasibility analysis, to the program of the glass’s thickness detectionwhich is based on the machine vision mechanism, is made in theory due to the opticalimaging principle. After that, the demands of the system for light source and imagingdevice are analyzed. A line laser is selected as the light source of the system as well as alinear CCD camera as the visual inspection tool. Subsequently, the performance indicatorsof testing components including laser, CCD camera, industrial lens and image collectioncard are analyzed.Then, the processing methods of grayscale image have been studied in accordancewith the requirements of the test. Compared with the reflected light, the image obtained bydiffuse reflectance light appears dim. It mainly includes filtration and grayscaleenhancement for the image processing and the center of light stripe can be found by usingthe method of Gray-gravity.Finally, the experiment is conducted to detect the thickness of the flat glass.According to the design of the thickness detection system, the experimental elements areinstalled, and the experimental platform is constructed on the optical table. A standard flatglass is selected as the experimental sample, with determining the pixel differencebetween the double light bands of the gray image by MATLAB. Combined with the realvalue of the flat glass’s thickness, the coefficient between the thickness and the pixel difference of the system is given. Glass samples with standard and different thicknessesare measured using this system. The design of glass thickness measurement system isfeasible by comparing the measured value and the true value.