Research On Lens Defect Detection And Classification Technology Based On Machine Vision

With the widespread application of optical lenses in security,virtual reality,consumer electronics,etc.,the quality of optical lenses is particularly important.Traditional artificial naked eye quality inspections have the disadvantages of inconsistent implementation standards and high labor costs.In addition,the current high-precision optical lens inspection instruments are expensive and cannot adapt to the complex environment of the factory.Therefore,the realization of automatic detection of optical lenses is of great significance to the development of related industries such as optical lenses in my country.First,according to the requirements of system detection accuracy and detection speed,an optical lens detection system is built,which includes the selection of industrial cameras,lenses,and light sources,and performs image correction and image preprocessing on the images collected by the camera.In the image preprocessing Compare the denoising capabilities of median filtering,mean filtering,and bilateral filtering,and select the appropriate filtering algorithm based on the mean square error and peak signal-to-noise ratio data.Secondly,the defect detection processing of the optical lens image is carried out.Aiming at the shortcomings of the Otsu algorithm in image segmentation and combining the PSO algorithm,an improved PSO+Otsu threshold segmentation method is proposed.This method redesigns the weight update formula of the PSO algorithm,which can be more Obtain the image segmentation threshold faster and more accurately,and finally extract the edge of the defect through the Zernike algorithm,so as to accurately achieve image segmentation.Then extract feature information such as contour,area,perimeter,gray level,etc.of the defects separated by the image.In the research of lens defect classification,the characteristic information of optical lens defects is reduced by principal component analysis(PCA),and Support vector machine(SVM)is used for defect classification,and for the penalty factor parameters in support vector machine(SVM)and the parameters in the Gaussian radial basis kernel function,there are inaccurate shortcomings based on manual experimental settings.An improved particle swarm is proposed.The algorithm's support vector machine(SVM)parameter optimization algorithm.Finally,in the lens defect classification experiment,the recognition accuracy of the support vector machine(SVM)parameter optimization algorithm based on the improved particle swarm algorithm proposed in this paper is higher than that of the support vector machine(SVM)classifier based on the artificial selection of parameters and the improved genetic algorithm.The classifier of the algorithm,so the algorithm in this paper has a certain meaning.