ABSTRACT
A smartphone sky camera and associated image analysis algorithm has been developed and validated to determine total cloud cover and the percentage of thick and thin clouds using supervised image classification for a range of cloud types, cloud cover from 4% to 98%, and solar zenith angles between 6° and 49°. Additionally, this system provides the percentage of total, thick, and thin clouds in proximity to the solar disc. The measurement errors were comparable to those associated with values obtained by commercial sky camera systems. The benefits of increasing the availability of cloud fraction measurements include the potential to develop improved local ultraviolet indices and weather forecasts and to provide a better understanding of cloud patterns required for the generation of solar energy.