Cloud cover forecast from a ground-based all sky infrared thermal camera
We describe the contribution of thermal infrared ground-based cameras in the short term Global Horizontal Irradiance (GHI) forecasting. This contribution is compared to the one of visible cameras, the most widespread technology currently used for this application. Accurate forecasts at short term horizons (5 to 30 minutes) under various and changing weather conditions represent an essential data for various applications such as optical availability and solar plant control procedures. The work presented in this paper first draw up an overview of the two cameras used in the following comparative study. The segmentation methods chosen for each of the camera and the protocol are subsequently described. Finally, the results of the study are presented and discussed. Thanks to the new opportunities it offers in terms of feature extraction and its capacities to overcome visible limitations, the thermal infrared camera shows a sizeable improvement in this comparative study.