Surface Defect Detection of Solar Cells Based on Multiscale Region Proposal Fusion Network

Abstract

Manufacturing process and human operational errors may cause small-sized defects, such as cracks, over-welding, and black edges, on solar cell surfaces. These surface defects are subtle and, therefore, difficult to observe and detect. Accurate detection and replacement of defective battery modules is necessary to ensure the energy conversion efficiency of solar cells. To improve the adaptability to the scale changes of various types of surface defects of solar cells, this study proposed a multi-feature region proposal fusion network (MF-RPN) structure to detect surface defects. In such a network, region proposals are extracted from different feature layers of convolutional neural networks. Additionally, considering that multiple aspect ratios and scale settings and the use of multiple RPNs, result in an overlap of candidate regions and lead to information redundancy, we designed a multiscale region proposal selection strategy (MRPSS) to reduce the number of region proposals and improve network accuracy. Owing to the complete learning of shallow-detail texture information and deep semantic information, our multiscale RPN fusion structure can effectively improve an object's multiscale feature extraction ability for various scales and types of surface defects of solar cells. Experimental results demonstrate that our method outperforms other methods by achieving a higher detection accuracy.

Surface Defect Recognition of Solar Panel Based on Percolation-Based Image Processing and Serre Standard Model

“…The multiscale RPN model proposed by them could fully explore the surface defects of solar panel of different sizes and types. Experiments have shown that compared to other methods, this algorithm had a significant improvement in detection accuracy [8].…”

Investigation of crack detection in solar cells by context integration and attention enhanced feature pyramid network

“…Currently, machine vision-based electroluminescence (EL) imaging detection technology has been widely employed for defect detection in solar cells. Deep learning-based object detection algorithms have shown the capability to automatically extract features and patterns from images, demonstrating strong robustness and generalization in practical engineering environments [1][2][3][4][5][6][7][8][9] .…”

“…Chen et al [6] introduced a Bidirectional Path Group Attention Detector (BPGA-Detector) based on the Faster R-CNN, suppressing the background interference and highlighting the defect locations. Zhang et al [7] proposed a Multi-Feature Region Proposal Fusion Network (MF-RPN) structure for detecting surface defects in solar cells, enhancing the algorithm's multiscale feature extraction capabilities. They also suggested an improvement in accuracy and recall by merging the results of the Faster R-CNN and Region-based Fully Convolutional Networks (R-FCN) [8] .…”

Research Progress on Deep Learning Based Defect Detection Technology for Solar Panels

“…In order to achieve the goal of carbon neutrality, solar power generation will become a strategic industry prioritized by the state, and photovoltaic companies will continue to expand the scale of production [1] . Solar panels may be improperly operated during the production process, resulting in defects such as broken grids, missing corners, color differences, dirt, cracks and other defects on their surfaces, which will not only reduce the service life of solar panels, but also affect their work efficiency [2] . Therefore, the defect detection of solar panels has become an important guarantee for the reliable operation of solar panels, and the study of defect detection methods for solar panels has important engineering practical significance.…”