SolaCam: A Deep Learning Model for Solar Radiation Estimation Using Consumer Cameras

Abstract

This study proposes a deep learning approach called SolaCam to accurately estimate solar radiation from the images captured by cameras. The proposed SolaCam performs deep learning by utilizing both image features and theoretical maximum solar radiation that vary with time and location. The trained model is capable of accurately estimating solar radiation on the ground surface from sky images captured by smartphones, fixed-point cameras, and other devices. The developed SolaCam can use a remote sensing function, which estimates solar radiation, on inexpensive camera-equipped devices.