Improved colour attenuation prior based dehazing by edge attenuation method

Abstract

Outdoor images taken in bad weather usually lose contrast and delity, resulting from the fact that light is absorbed and scattered by the turbid medium such as particles and water droplets in the atmosphere during the process of prop-agation. Moreover, most automatic systems, which strongly depend on the de nition of the input images, fail to work normally caused by the degraded images. Therefore, improving the technique of image haze removal will bene t many image understanding and computer vision applications. The goal of Haze Removal algorithms is to recover and enhance the scene from foggy image. Existing dehazing methods does not estimate haze thickness up to the point and thus can-not e ectively provide satisfactory haze removal results. In proposed system an improved colour attenuation prior based dehazing by edge attenuation is used to get e cient dehazing results. method is used which dynamically repair the transmission map and thereby, achieve satisfactory visibility restoration alone with dehazing. It involves haze density estimation, transmission map recovery and image visibility restoration. The proposed dehazing algorithm takes advantage of edge collapse to self-adjust the estimated transmission map in order to achieve a better haze removal e ect and is also ca-pable of dealing with images corrupted by thick haze particles.