Video Haze Removal Approach Based On Dark Channel Prior

With the continuous improvement of computer technology, video surveillance systems used in people's daily lives in many aspects, such as monitoring traffic roads, residential, shopping malls monitoring. When property damage occurs, the content of the video surveillance video recorded can be used as supporting evidence to be used. However, in case of foggy weather conditions, the video image will become blurry, image quality, so you can not image feature extraction, target tracking and monitoring can not be performed. In order to remove the effects of fog video, video to this thesis, the fog method.Based on the previous studies on single video defogging first images were analyzed to fog. First study the foggy image of the physical model, and analyzed the formation of principle, but also describes the dark colors prior model, two models after math integration process obtained new physical model of a priori theory of dark colors on dark primary physical model parameter a priori theoretical studies included in the analysis, and then won the estimated transmittance, transmittance of the estimated optimized transmittance obtained after optimization, then get clear and effective defogging renderings.But also on the basis of video foggy dark colors based on the a priori theory defogging research, this analysis is the fixed background, outlook changed foggy video. Video background frame with dark colors prior to strike transmittance, the transmittance obtained a general view of this general transmittance FIG applied to each video frame to fog, haze results obtained to color coordination. But the dark colors prior knowledge limitations when dealing with sky and other bright areas, while de-mist in the sky area, transmittance of estimation error color "wave" shape, the sky area is distorted when defogging. To fix this problem, the paper added tolerance mechanism for weakening the sky area, without affecting the overall effect of the fog to improve the region's sky color distortion, enabling to have to deal with fog fog video.