Analysis and experimental validation of solar potential in urban roads using Google panorama images for solar electric vehicles

In this study, Google panoramic images were used to analyze reductions in direct and diffuse solar radiation caused by surrounding structures on urban roads and to predict road-scale solar irradiance for estimating the energy generation of solar electric vehicles (SEVs). Hemispherical images were generated from panoramic views, and a U-Net model classified sky and non-sky regions. By overlaying the annual solar trajectory, reduction ratios for annual direct and diffuse solar radiation were calculated. Annual solar potentials were estimated using typical meteorological year (TMY) data. A case study in Nam-gu, Busan, showed average irradiance of 0.93 MW/m2, with the highest at 1.47 MW/m2 on elevated roads and the lowest at 0.01 MW/m2 in tree-covered areas. Roads near residential buildings exhibited higher shading losses, with up to 17 % greater loss in October and up to 1.7 km/kWp shorter daily driving range in September compared to non-residential roads. Driving experiments validated the irradiance predictions by measuring actual reductions caused by buildings, trees, and elevated structures. The findings of this study can guide the development of navigation systems for SEVs.