Method On Image Clearness Technique In Fog Weather

The image we obtained in bad weather conditions is often significantly degraded by fog and mist. With the development of the video surveillance system techniques, the demands of system's reliability and stability have raised. In order to make video surveillance systems operate normally in fog weather, image clearness technique is proposed in the paper which aims to enhance weak information by software methods.The paper studies on the image clearness technique in fog. Firstly, the paper presents two classes of ways: One based on image enhancement technique and the other based on image restoration technique. Then due to the difference degree of information lost in haze, fog and dense fog different methods are adopted. Method base on Retinex theory is used to enhancement image taken in haze. As the result, higher contrast of image and the better color is obtained. In fog weather a novel method is presented. It is based on atmosphere scattering model, but not need to know depth information previously, by reconstructing parameter of model to estimate depth of scene. According to the depth, foggy image can be segmented into establishing shot and close shot. Then two different enhance method are used to two parts respectively. According to the characteristics that limited dynamic range of images taken in dense fog, dynamic range stretch as a global enhancement method is used. In order to reduce the effect of noise it brings, sparse coding is adopted.In this paper, computer techniques and images process techniques are used to enhance the clearness of images taken in fog automatically, which has a great significance in improving the robustness of outdoor surveillance systems and also be helpful to the process of images taken in other bad weather.