Detecting Durface Defects Dn Solar Cells Based On Frequency Domain Analysis

In recent years,with the continuous growth of energy demand,non-renewable energy has gradually used up,and solar energy as a clean and environment-friendly renewable resource has become the key planning direction of the country.Among them,one of the main solutions to solve such problems is solar photovoltaic power generation technology.In the process of making solar cells,the quality inspection requirements are very particular,such as whether there are physical damages,surface scratches,whether there are broken grids,whether there are microcracks and so on.In traditional factory production,the detection of the above defects requires professional inspectors to carry out visual inspection,which often leads to low detection efficiency,subjective assumption and fatigue,which may lead to detection errors.In recent years,computers have not only facilitated our daily life,but also developed rapidly in digital image processing technology,making it possible to use machine vision to detect defects in solar cells.Then this paper proposes a method for detecting surface defects in solar cells based on frequency domain analysis.The main contents are as follows.In this paper,first of all,some necessary preprocessing is carried out on the complete surface image of the collected solar cell:removing the black edge area in the collected image;The image with the black edge removed is divided into 60 small blocks,and the 60 small blocks are processed one by one to remove the black edge,and then are subjected to median filtering and Gaussian filtering noise reduction.Secondly,the processed image is de-rasterized:a method combining Fourier reconstruction and edge detection is proposed to remove the grid retention defects.Fourier reconstruction is adopted to remove the foreground to obtain the grid line;Edge detection is carried out on the reconstructed image and the original image,the position of the grid line is obtained by reconstructing the edge map of the image,and the found grid line,that is,the initial detection result of the defect,is removed from the edge map of the original image.Then,the defects in the initial detection result are screened:the real defects are found by setting a simple threshold.Finally,the detected defects are accurately located:an improved scanning line method is proposed to accurately locate the defects.In addition,a solar cell defect detection system is designed based on the proposed defect detection method.The system has a simple and humanized GUI interface and is also very convenient to use.In this paper,the test database collection of two types of solar cell surface defect images was collected from the solar cell production plant,so as to realize the subjective and objective experimental evaluation of defect detection on the library.The method in this paper has relatively high defect detection accuracy and defect location capability,and plays a significant role in greatly improving the factory qualification rate of solar cell products.