Research Of Fog-degraded Image Enhancement Algorithm And DSP Realization

With the improvement of the human condition, computer intelligence increasingly demanding, intelligent equipment widely used in target identification, road monitoring, military reconnaissance and other fields. The degradation images of low contrast, color distortion acquire from outdoor due to the influencement of foggy environment, which seriously affecting people’s normal life and work. In order to improving the foggy image contrast and sharpness, eliminating color distortion, making images look more natural, currently researches of image defogging method mainly from two aspects: the foggy image enhancement and image restoration method. The main work is as follows: 1) A brief introduction and analysis of the research background and significance of the image dehazing, development trend and prospect at home and abroad, the basic theory of image enhancement, atmospheric scattering physical model and image quality evaluation criteria. 2) The theory of the global, local and dual histogram equalization algorithm are introduced and analyzed the advantages and disadvantages of the above algorithms. An daptive local contrast enhancement algorithm of minimum brightness error was proposed by brightness preserving bi-histogram equalization algorithm, this paper introduces the basic principle of the algorithm, implementation steps of the algorithm, through the experimental realization of the algorithm and the above algorithms, and a comparative analysis of the the advantages and disadvantages of the algorithm. 3) This chapter analyzes the principle and feature of single image haze removal using dark channel prior, an improved algorithm is proposed for solving the defects caused by this algorithm. Described in detail the basic principle of the improved algorithm, analyzed the advantages and disadvantages of the improved algorithm by the experiment of this improved algorithm compared with the previous algorithm. 4) In order to verify experiment result and real-time of the above two improved algorithms, this paper built a DSP embedded system, detailed description of the system components and integrated development environment, the experimental results of the two improved algorithms are compared and analyzed. Improved dark channel prior image restoration algorithms based on embedded image processing system, the image to improve more clear and nature.