The paper focuses on improving the reliability of remote sensing of rooftop PV installations using deep learning and the wavelet scale attribution method (WCAM). The WCAM is used to assess the scales on which a PV model operates and to derive methods to increase robustness to shifts in acquisition conditions. This increased reliability and robustness of deep learning models for rooftop PV mapping would facilitate the mapping and integration of rooftop PV into the electric grid, contributing to the decarbonization of energy systems.
Publication date: 21 Sep 2023
Project Page: https://arxiv.org/abs/2309.12214v1
Paper: https://arxiv.org/pdf/2309.12214
