Research And Implement Of Lens Residual Reflectance Color Detecting System Based On Computer Vision

In recent years,as the demand for lenses has increased,the lens manufacturing industry has come up with new requirements for scale,quality and productivity.Especially in quality monitoring,the method of manual detection is becoming more and more difficult.The lens residual reflectance color quality problem is one of the important factors affecting the quality of the lens.This paper mainly describes a method based on computer vision to quickly and accurately,non-contact detection lens residual reflectance color quality.Lens residual reflectance color quality problem is divided into two aspects,one is the single-sided is not in the standard range;Second,the difference of double-sided is too big.The research has done the following tasks respectively: Firstly,a set of image collection equipment was designed to collect the picture of lens residual reflectance color;Secondly,based on image processing theory,paper designed the corresponding image filtering algorithm and image segmentation algorithm to extract color features from the original image.Then modeled the color space of the extracted color features;Finally by analyzing of the current mainstream of the pattern recognition method,and through comparing the advantages and disadvantages of several common classifier,this paper put forward a suitable method for the quality of lens residual reflectance color classification.At the end of this paper,the algorithms proposed in this paper are systematically implemented,and the system is tested in the field.According to the test results,the related methods are compared and analyzed.The accuracy of the comparison displayed that system is higher than the artificial quality control,and the feasibility of the method is verified.