Removing rain and snow in a single image using saturation and visibility features

摘要

Rain and snow are two of the major obstacles in processing the photograph captured in the outdoor bad weather conditions. Rain and snow will cause the performance of vision algorithm become worse. There are a lot of methods have been proposed to reduce the raindrops and snowflakes in video. However, how to remove rain and snow and keep the detail of the background in a single image is still quite challenging since it is hard to detect the pixels of rain and snow. In this paper, we proposed a new method based on features on saturation and visibility. The results show that it can achieve better performance than previous methods.