Research And Application Of Inspection Of Solar Cell Surface Defects Based On Biologic Visual Model

With research and development of solar energy, solar panels research andmanufacturing were concerned which transform solar energy into electricity. Thedefects solar cell product will greatly affect the life and the photoelectric efficiency.Therefore, defect inspection had become a very important part of production. Inorder to develop the robotization and relaxhuman eye,software-based imageprocessing had a high requirements of image acquisition, but the recognition ratewas less than the requirements of the industrial inspection.According to the latest research progress and research results in computervision, the Serre standard model was used that can simulate the object recognitioncapabilities of the ventral visual pathway. In the pre-processing, denoising spot wascombined the LOG and adaptive Wiener filter, and got LGBPHS characteristic. Inthe next step, percolation-based image processing method to determine the crackpositioning and the final outcome based on an improved standard model which theC2layer output MAX value whether classified the crack. Analysis of experimentalresults that the method overcame the low recognition rate of the traditional imageprocessing methods, but was still a lot of that should be improved, such as theintroduction of the Bayesian model for the model provides a feedback mechanismto enhance self-learning ability and improved layers of algorithms to improve theefficiency of feature calculation, they were the directions of future research.As we all know the human eye can identify what the machine cannot do whichusing conventional identification methods. With the human eye has a fast responseand learning and comparing right thing to identify, through the study of the humaneye recognition principle, developing them a chine to simulate the human eyerecognition and apply that into industrial production.