Research On Differential Case Recognition Technology Based On Machine Vision

In recent years,machine vision technology develops rapidly and plays an extremely important role in automatic industrial production.Machine vision recognition technology is to get the image that can be used for recognition through the digital processing of the image.After different recognition methods to recognize the image,instead of completing the recognition task manually.In view of the problem of model diversity in the mixed production system of differential case,this paper proposes the research of differential case recognition technology based on machine vision to identify different types of differential case.First of all,according to the specific requirements,the overall scheme of the differential case visual recognition system is designed,the main hardware of the visual system is designed and selected,and the hardware experiment platform of the differential case visual recognition is completed.Secondly,the image processing technology is studied,and the gray-scale processing,contrast enhancement,image filtering,binarization and edge detection processing are carried out for the color image of the differential case.The basic principles and practical effects of various methods are analyzed,and the specific implementation methods of image filter and edge detection are determined.Thirdly,the image features are analyzed,two image shape feature extraction algorithms,Hu moment and Zernike moment,are studied,and the Zernike moment is improved to reduce the calculation,shorten the CPU running time,and improve the feature extraction speed of Zernike moment.Finally,the neural network is studied.Aiming at the problems in the practical application of BP algorithm,this paper proposes a combined optimization improved BP algorithm.The structure model of neural network is designed and tested.According to the comparison of the results,the recognition rate of Zernike moment combined with BP neural network is higher.A set of differential case identification software system is developed by computer programming,which can identify differential case quickly and accurately.