Cloud identification model for sky images based on Otsu

Irradiance changes caused by clouds movement are the main factors, which lead to mutation of photovoltaic power. In order to increase the accuracy of the ultra-short-term photovoltaic power prediction, based on of clouds images, the cloud identification for sky images is one of the primary aspects for improving photovoltaic power accuracy. A cloud identification model for sky images is established based on Otsu. Firstly, a series of sky images are collected by Total Sky Imager (TSI) with the regions of clouds and sky are distinguished. Secondly, the color features (based on RGB values) of pixels in sample images and the spectral and texture features of the whole image are extracted; Finally, we established a cloud identification model for sky images based on Otsu according to the clouds´ and sky areas´ division and the general characteristics featured by pixels and images. The simulated results show that the model in this paper can identify the clouds in a sky images effectively.