Study On Image Dehazing Based On Adaptive Wavelet Fusion

Images captured in fog or haze weather condition often degrade seriously because of attenuation and scattering of fog or haze. They often exhibit low contrast, color distortion and poor visibility. Heavy fogs are not only the traffic obstacle of human daily life, but also affect some computer vision applications, such as traffic monitoring, military remote sensing, target detection and tracking. Hence, the research of how to restore the haze-free image has important theoretical and practical significance.Through analyzing the degraded reasons and the characteristics of hazy images, a method based on the atmospheric scattering model and dark channel prior is proposed in this thesis for removing the influence of haze. At first, the improved quad-tree subdivision method is used to estimate atmospheric light. Then, two atmospheric veils with different scales are first derived from the hazy image and a local similarity-based adaptive wavelet method is proposed to fuse the two atmospheric veils adaptively, which can reflect the scene depth information well. At last, the scene and transmission map is restored by solving the atmospheric scattering model.Both subjective and objective evaluation methods are used for testing the effectiveness of the proposed dehazing method. A lot of natural scene hazy images and the simulative hazy images are chosen in subjective evaluation for experiments. Experimental results demonstrate that the proposed method can yield comparative or even better results than several state-of-the-art methods. Focus on the issues of low contrast, color distortion, image contents missing, the dehazing images’ quality are evaluated by three objective evaluation indexes-average gradient, color consistency and structure similarity. The evaluation results show that our method exhibits better contrast, color and structure of the original scene.