Research On Weak Surface Defects Of Solar Cell Based On Decomposition And Fusion Method

The solar cell is the core part of the photovoltaic power generation system,and its surface quality directly affects the power generation efficiency and service life of the photovoltaic system.The surface defect detection of solar cell has become a challenging task due to the variety of surface defects of the battery sheet,poor contrast between the defects and the background,and uneven background texture.Therefore,this paper proposes two methods from the perspective of texture analysis and feature fusion,and designs and studies the surface defect detection algorithm of solar cells.The specific research contents are as follows:(1)In order to solve the problem of random texture interference in complex background,a method for detecting weak surface defects of solar cells based on structural texture decomposition is proposed in this paper.Firstly,the collected solar cell is pre-processed and converted from a three-channel color image to a single-channel black-and-white image;then,the improved Retinex algorithm is used to enhance the defect information while removing the solar cell surface gate line.Secondly,use theTV-L~2structural texture decomposition model of adaptive parameters processes the image and extracts the structural part of the image.Finally,adaptive threshold segmentation is performed on the structural part,and the segmentation result is optimized by morphological operation to obtain the final defect detection result.The experimental results show that the method can effectively enhance the defects and overcome the interference of complex background.However,when the contrast between the defect and the surrounding background is low,the problem of missed detection is prone to occur,and the detection performance needs to be further improved.(2)Aiming at the problem of low contrast and complex background interference,a new method for detecting weak surface defects based on multi-feature fusion is proposed in this paper.The fusion process consists of four phases:multiscale decomposition,building decision maps,decision graph optimization,and image fusion.Firstly,Laplace multi-scale decomposition of the image is performed,the source image is decomposed into five different scales and the sub-image is scaled to the same size as the source image using an interpolation algorithm;then,the structure is extracted from the solar cell image separately.Sexual features and detail sharpness features are used to construct the initial decision graph.After that,the gradient map is used to optimize the decision graph by considering the correlation between neighboring pixels.Finally,the optimized decision graph is merged.The method fuses multiple features and makes full use of structural saliency information and detail definition information in the image.In addition,this paper also complete the selection,construction and testing of the solar cell visual acquisition system,and develop a software system for solar cell visual defect detection.