Real-Time Color Image Dehazing

Since the development of computer vision application, surveillance system enjoys increasing popularity in everyday life. However, in complex scenes such as haze, the visibility of outdoor scenes declines and the quality of the captured images degenerated extraordinarily in CCD based system which leads to video surveillance and information acquirement problems. Therefore, image dehazing becomes a research hot point in academic fields. Nowadays, the dehazing algorithms can be mainly divided into two categories:image enhancement based dehazing algorithms; image restoration based modeling ones. There are advantages and disadvantages in both categories. In general, the enhancement methods enjoy lower complexity and are able to procedure real-time process. However, the sort of algorithms is non-model based which causes less effective dehazed images. On the other hand, the restoration methods obtain effective dehazed results with high processing complexity. Moreover, the gray scale image restoring and enhancing attract more attention than color image which is often used in video surveillance systems,This article focuses on real-time color image dehazing. We analyze the cause of hazy image distortion and solve the dehazing problems. Firstly, we briefly review all related work domestic and aboard and then deeply research.on three algorithms which are Retinex based dehazing, Dark Channel Prior (DCP) based dehazing and Median Filter Visibility Restoration (MFVR) dehazing. Based on the algorithms above, we propose some novel dehazing algorithms to achieve clear dehazed images. Experiments’results show the vivid color and the accuracy of the restored edges and gradients. A comparison study is made among all these algorithms.This article makes innovation which mainly includes two items:.propose a real-time image defogging algorithm based on digital total variation filter (DTVF). Based on adaptive characteristic of the digital TV filter and the minimal component image, the article combines the digital TV filter with hazy image. The results demonstrate that the new method is better than the others both in effectiveness and details preserving. Moreover, the method can be adopted in real-time processing because of fast convergence of digital TV filter..propose a real-time image and video dehazing algorithm based on color transfer. Color transfer model is adopted to dehazing algorithms for the first time. In order to obtain a wealthy and vivid color of the dehazed images, color transfer is applied to make the dehazed images have the color information of the clear images. We propose a new real-time image dehazing algorithm to combine the digital TV filter and color transfer together. The new algorithm restores the hazy images details and obtains rich color information simultaneously. Results on variety of images and video sequences demonstrate that the novel combined algorithm is effective on both natural and synthetic images.