Framework For Simulating And Rem,/Oving Rain In Stereo-Image Videos

We present an implementation of a rain simulation and rain removal framework, which is based on video streams captured with a stereo camera system. This framework can be used to study the effect of rain on vision-based systems in the automotive domain, including obstacle detection, vehicles tracking and traffic sign recognition. We also present testing techniques, to verify the performance of rain simulation/ removal systems, and quantify the closeness of de-rained image to the original rain-free image.