Automated detection and tracking of photovoltaic modules from 3D remote sensing data

This study addresses the growing demand for increased performance and reliability of photovoltaic (PV) installations by developing innovative monitoring technologies. The strategy consists of flying an unmanned aerial vehicle (UAV) equipped with a dual camera (RGB and thermal) over the PV plant of interest, followed by the generation of photogrammetric 3D models derived from the overlapped aerial images. The resulting datasets involve orthoimages and point clouds by processing RGB and thermal imagery. The key contribution of this study is twofold: (1) the thermal image mapping on dense and high-resolution point clouds that represent the status and geometry of PV solar modules, and (2) the automatic identification of individual solar panels in 3D space and their thermal characterization along their oriented surface. Then, the vector layer of each PV panel is projected onto the 3D thermal point cloud to extract the thermal values associated with each panel. To evaluate the capability of the proposed method, it was replicated in different scenarios, considering rural and urban environments with different light conditions and PV structures. The results demonstrate the robustness of our method, which achieves a remarkably high detection rate, around 99.12% of true positives, and a low false positive rate, close to 0.88%. Consequently, this method means an advance over previous work by proposing a comprehensive and automated solution for individual and highly detailed monitoring of each solar panel from 3D remotely sensed data. This study opens up new frontier research related to real-time monitoring of photovoltaic modules, an inspection of solar photovoltaic cells, the simulation of solar resources and forecasting, the development of digital twins, solar radiation modelling, and analysis of modular floating solar farms under wave motion.