Drone-based solar panel inspection with 5G and AI Technologies

Abstract

It’s been considered an incomplete task for years to maintain large solar power plants for years. Presented here is an Artificial Intelligence (AI) based defects detection of Photovoltaic(PV) modules using Thermal Images (TI) darknet YOLOV4 object detection, which can be processed in two ways: (1) Creating a huge number of high-resolution TI samples using a huge number of TI generation methods; and (2) using the generated TI’s, to develop an efficient method of defects classification. Convolution Neural Network (CNN) technology and traditional image processing technology are combined to result in the TI object detection method. This method has a capability of training a large number of high-resolution TI samples to give a good AI model output. Then, CNN is used to extract the deep feature of TI to show the defected cells. In other hand using enhanced 5G technology it is used to operate the drone for long range and by help of AI, can send the defected cells location to the ground station. Compared to other solutions, using it can improve PV module inspection and health management solutions significantly. It has been demonstrated experimentally that the proposed AI-based solution is efficient and accurate at detecting defects using TI and drones automatically.