Research On Solar Cell Defect Detection System

In today’s increasingly serious energy crisis, Solar energy as a green, environmentalprotection, new energy inexhaustible that more and more attention has been paid, Therefore,the development of solar cell production technology has become a hot topic. In the studyand manufacture of the solar cell, defect detection technology will have a direct affect thequality of manufactured products and is very important for businesses, So for solar celldefect detection technology has a future use.The basic structure and optical properties of the solar cell was studied in this paper.The defects of the cells are classified by the mechanism and production process. Severalcell defects detecting methods are interpreted and their principle,characteristics,advantagesand disadvantages are compared， and an electroluminescence-based solar cell defectdetecting system is designed.By using Visual C++6.0a detecting software system on PC was programmed. Thissoftware enables to display the defective images,view the image histogram, also to zoom inor out on an image. The image processing steps of this software are divided intopre-processing, identification and preservation of test results. First, the image medianfiltering, Gaussian smoothing, Laplacian sharpening, histogram equalization, andpseudo-color and other pretreatment. Then defect detection and identification, softwaredefects into the bad plaques, voids and cracks three cases were detected. The histogramsegmentation algorithm and Otsu algorithm can be used for bad plaques, in addition, animproved image segmentation algorithm can achieve better image segmentation results,used opening operation to removal the gate lines and other effects for the divided images,finally,using the point statistical decision methods for determining, bad plaques can bedetected by this system; Voids can use Roberts operator, Sobel operator, Laplacian, Prewittoperator and Krisch operator for edge detection, through experimental comparison foundRoberts operator edge detection works best, after binarization, use homemade recognitionalgorithm to identify, voids can be detected by this system; the crack detecting is taken inthree directions: vertical and±45°on the enhancing process. In this paper, the defect detection method of the solar cell was analyzed. A solar celldefect detection systems and PC software was designed.The software is designed to achievethe solar battery defect detection and identification, achieve the desired objective.