Abstract: In the process of production defects can reduce the efficiency of solar panels,so it need to defection the defect.In the process of solar cell defects automatic detection,defect segmentation process is the difficult process and the premise and key for subsequent identification.This paper based on household small solar panels for defect detection,The characteristics of this kind of panels is a rectangular frame,electrode in the central and unequal electroluminescent grayscale.According to the above characteristics,An innovative method of defect detection is proposed.First acquire the image of a priori knowledge,Then defect segmentation by using foreground segmentation,calculate the gray level difference,calculation of defect location of beginning and ending,connected to the defect area,morphological open operation processing,remove the electrode position.The experiment results show that the proposed method can successfully segment the defect.