Research And Design Of Image Dehazing Method

With the rapid development of our country's economy and society, the degree of environmental pollution has increased, which makes haze weather appear frequently around the country in recent years. Haze not only hurt human health directly, but also lead a serious threat to traffic safety, and makes lower quality of image of kinds of surveillance systems,such as low contrast, fuzzy details, color distortion and so on, which disrupts the smooth progress of related work. Thus achieve clear with no fog image not only have highly application value and economic value, but also is a hot research direction in the field of computer vision and image studies.This paper works on the algorithm of single image dehazing, the study shows that traditional algorithm achieve a certain effect about the image in light fog weather, but not so well in heavy fog weather. According to the characteristics of the image in fog weather and the characteristics of the human eye, integrated histogram equalization, homomorphism filtering, dark channel prior three traditional methods dehazing, make a use of image segmentation and image fusion technology, this paper puts forward a kind of mixed algorithm of single image dehazing. this paper also compares it with the above three methods, the superiority of the design method is verified by the subjective and objective evolution standard. On the basis of this, the paper realizes the visual operation of the image dehazing. The main work of this paper is as follows:This paper introduces the research status of the algorithm of image dehazing in domestic and abroad, which contains two aspects from the image enhancement and image restoration, and summarizes the research background and significance of the algorithm of image dehazing.This paper analyzes the reason of image degradation in fog, and describes the subjective and objective criteria of image quality assessment.This paper focuses on histogram equalization which is based on airspace enhancement method, homomorphism filtering which is based on frequency domain enhancement method, dark channel prior which is based on image restoration method. This paper studies the principle and operation process of each method, and in a serious of experiments under various of complex weather conditions,the performance and the effect of these algorithms of image dehazing have been compared and analyzed and draw conclusions.This paper introduces the significance of image segmentation technology, and analyses the principles, performance and advantages and disadvantages of four categories of image segmentation techniques, which are threshold segmentation technology, zone segmentation technology, the edge segmentation technology and the region segmentation technology. This paper introduces the significance of image fusion technology, and analyses the principles, performance and advantages and disadvantages of five categories of image fusion techniques, which are selected weighted average fusion technology, HIS fusion technology, PCA fusion technology, Pyramid fusion technology, Wavelet transform fusion technology.On the basis of the above research of the algorithm of image dehazing, this paper puts forward a kind of mixed algorithm of single image dehazing, and proves the availability and reliability of the design method in dealing with the image in light fog weather and in heavy fog weather. At last, this paper designs a multi-method optional dehazing system software, which realizes the visual operation of the image dehazing.