PV-VISION: An Deep Learning based Package for Automated Solar Module Inspection

Solar photovoltaic (PV) modules are susceptible to manufacturing defects, mishandling problems or extreme weather events that can limit energy production or cause early device failure. Trained professionals use electroluminescence (EL) images to identify defects in modules, however, field surveys or inline image acquisition can generate millions of EL images, which are infeasible to analyze by rote inspection. We developed an open-source computer vision package PV-VISION to automatically process the EL images using deep learning models, covering automatic image preprocessing, cell defect detection and crack feature extraction. We demonstrated the functions of PV-VISION on two tasks: investigating fire impacts on solar farms by inspecting 2.4 million cells and quantifying crack growth in solar modules under mechanical aging tests. We anticipate that PV-VISION can offer a supportive platform for researchers in the solar field, facilitating a more efficient and data-driven approach to EL image analysis.