A Novel Approach for Rain Pixel Detection and Recovery by using Motion Segmentation

The visual effects of rain are complex. Rain removal is a very useful technique in applications such as security surveillance, video editing, vision based navigation, and video indexing/retrieval. Rain produces sharp intensity variation in images and videos, which put down the performance of outdoor vision systems. These intensity variations depend on different parameters, such as the camera properties, the parameter of rain, and the brightness and sharpness of the scene. Removal of rain stripes in video is a hard task due to the random spatial distribution and fast motion of rain. Photometric, chromatic, and probabilistic properties of the rain have been exploited to detect and remove the rainy effect. This paper is introducing the Rain Pixel Algorithm with better performance for rainy scenes with large motion than exiting algorithm.