Generalizing projected LiDAR data with support vector machines for solar energy applications

This project aims to facilitate the application of Light Detection and Ranging (LiDAR) point-clouds for taking into account shading of photovoltaic (PV) systems in the urban environment. In the Netherlands, LiDAR data is available nationwide, with a 0.5 m resolution. We use machine learning to turn this raw point-cloud data into a valuable asset for shading calculations. Applications include PV and building design support, and providing input for making digital twins for fault detection of PV systems.