Efficient Diagnosis of Photovoltaic Cell Degradation Based on Deep Learning Using Drone Thermal Imagery

Solar energy has recently become the ideal solution for reducing the use of polluting energies. Due to the great importance of this natural and unlimited source of electrical energy, the latter requires a reliable fault diagnosis system with effective and rapid performance in detecting the degradation of PV cells, which are the main energy converters of the photovoltaic system, in order to avoid interruption of power generation and ensure its proper operation. It can be difficult to reach the location of these stations to Diagnose and Detect cell Failures (DDF), so the drone technology is useful in this case in parallel with deep learning techniques in image processing and analysis. In this work, depending on the photoelectric thermal image dataset, abnormal cells are detected only through the use of thermal data because they have a different temperature. We propose a model that detects the defective cell based on one of the deep learning tools called YOLO. The effective performance of the proposed model is verified in detecting degenerated cells with high accuracy and standard speed.