Research On Alogrithm Of Haze Image Restoration

In the weather condition of fog, due to the effect of particles scattering in atmosphere, the scene image is captured by devices has been seriously degraded with a low contrast. Fog greatly affects the work of outdoor vision systems, such as satellite remote sensing systems, aerial systems, outdoor surveillance and target identification systems. Therefore,the haze image clearness technology has been one of the popular research topic in the filed of computer vision and image processing.Firstly, analyze the characteristics and mechanism of atmospheric scattering model of fog image degradation based on algorithm of image enhancement and image restoration. Next, to solve the issue of the high complexity and distorted bright areas, proposed a improved algorithm based on the dark channel prior, using improved median filtering algorithm instead of soft matting, a nd revising the transmisson map of bright areas. Then, after analysing the relation of the minimum map contains depth information and the luminance component, a dehaze algorithm based on the depth of the fog is proposed which is effective to estimate the depth map, and obtains the image without fog quickly. Finally, the experiment results show that this proposed algorithm not only can significantly improve the contrast and color of haze image with strong robustness,but also the quality and the efficiency is superior to the latest comparison algorithms.