Research On Sub-pixel Edge Detection Technique In The Image Measurement Of Small Module Gear

The small modulus gear is widely used in the precision instrument industry.Its accuracy will directly affect the life of the instrument and the work performance.So the small modulus gear geometric measurement is important.The traditional contact measurement method is cumbersome,and it can't meet the needs of modern measurement.And using the machine vision technology to finish the small modulus gear measurement can achieve accurate nondestructive testing.And it has become an important development trend.In this paper,the edge detection technology of the small modulus gear and the evaluation method of edge detection effect were studied based on the construction of machine vision measurement system.The main research contents are as follows:(1)According to the design options to select the hardware equipment,we built a small modulus gear visual measurement system for small scale gear precision measurement.(2)In order to improve the edge localization accuracy by hardware condition,the subpixel edge detection algorithms are studied.The antinoise ability and detection efficiency of different algorithms are analyzed,and the sub-pixel edge detection algorithm based on Zernike moments is chosen and implemented.(3)In order to evaluate the edge detection effect and quantify the deviation of the edge detection results,an edge detection effect evaluation method based on feature image is proposed.Using the rich local image information can evaluate the effect contour extraction of small modulus gear in the visual measurement system.The subpixel edge detection algorithm based on Zernike moments is better than 0.6 pixel for different feature images.(4)Combined with the edge detection effect evaluation method based on feature image,the detection accuracy of different edge detection algorithms is analyzed.And it's proved that the detection accuracy of subpixel edge detection algorithm based on Zernike moment is better.And using the small module gear vision measurement system to detect the actual gear to analyze the repeatability and uncertainty The experimental results show that the measurement uncertainty of the system is 2.5?m.