Camera enhanced compressive light beam induced current sensing for efficient defect detection in photovoltaic cells

摘要

Recently, use of the compressive sensing-based light beam induced current (CS-LBIC) method has led to significant improvements in the scanning speed of the standard LBIC method that made the method suitable for practical use in fast and detailed characterization of photovoltaic cells. However, the sparsity of the defect signal obtained using the compressive sensing model is greatly increased by presence of bus bars and finger structures in the testing area, meaning that more measurements are required for reasonable signal recovery. In this paper, a camera-enhanced CS-LBIC method is proposed that eliminates the effects of both bus bars and finger structures and thus further improves the scanning efficiency of the CS-LBIC method. An image of the testing area was captured using an optical camera and the areas without the presence of bus bars and fingers were extracted as photosensitive areas. Then, an effective sensing and reconstruction algorithm was proposed that allowed direct recovery of the defect signal from the photosensitive area. The experimental results indicate that in a 10,000 point scan, defects can be located with a sampling ratio of 2%-5% using the proposed method with a scanning duration on the scale of tens of seconds, which is approximately five times faster than the reported state-of-art results. The proposed method shows significant potential for improvement of the efficiency and speed of meticulous inspection and classification of photovoltaic cells.