Research On Real-time Single Image Dehazing

Currently,due to the development of computer vision theory and computer hardware technology,computer vision system has been widely used in all aspects of daily life?However,the environmental pollution is still severe,foggy day is frequently appeared in recent years.Under foggy conditions,the visibility of outdoor scenes declines,images and videos captured by outdoor computer vision system have many problems,such as low contrast,color distortion,landscape details blur and other issues,having adverse influences on the subsequent object recognition and feature extraction,reducing the computer vision system robustness severely.Consequently,image and video dehazing become a research hotspot in computer vision areas.This article focuses on real-time single image dehazing.We solve the dehazing problems of hazy images with lower complexity in order to apply it in video dehazing.Firstly,we summarize the research status of domestic and foreign scholars in dehazing field and deeply research on the state-of-the-art methods.Then introduced the atmospheric scattering theory and the hazy image degradation model,and futher explained the key technologies of the Dark Channel Prior based dehazing algorithm.Based on the above,we propose some novel dehazing algorithms which can achieve excellent results as well as satisfy the real-time requirements..In order to evaluate the performance,the new algorithm is compared and analyzed from a qualitative and quantitative point of view.This article's innovation mainly includes:propose a dehazing method based on domain transformation-based edge-preserving filter.Domain transformation-based edge-preserving filter is adopted to dehazing algorithms for the first time.First,we compute the minimal compent image and estimate the atmospheric light based on the weighed quadtree subdivision strategy.Then,we utilize the domain transformation-based edge-preserving filter to refine the transmission map and finally obtain the recovered image.Experimental results demonstrate that the new method not only can completely recover the details of the scene,but also has a linear complexity,which can satisfy the real-time requirements of computer vision system.a novel approach using image fusion and contrast limit adaptive histogram equalization(CLAHE)for single image dehazing is proposed.Firstly,we split the sky region to compute the atmospheric light.Because of the original recovered image suffers from color distortion and darker brightness,we employ a color correction scheme based on the image fusion and CLAHE in L??color space and finally we obtain the haze-free image.Experimental results illustrate that the algorithm can effectively improve the clarity of the image and avoid color distortion.