Research Of Fog CCTV Images Real-time Clearness Methods

Video surveillance products play an increasingly important role in our daily lives, which has been widely applied in various fields of security and defense, industrial and agricultural production, transportation, military science and technology. However, outdoor video surveillance system is sensitive to the weather conditions. Particularly by low-visibility fog weather, the scenery image degenerates seriously, it is difficult to distinguish valid information in time, which has limited outdoor video surveillance system. In order to enhance the effectiveness of outdoor video surveillance system, and make the outdoor video surveillance systems work around the clock, it is necessary to improve system’s robustness and real-time works in fog and other inclement weather. Therefore, there has very important practical value and scientific significance to research fog-degraded video image clearness methods.In the field of image processing, digital signal processor has a wide range of applications. It has special hardware and software structure and powerful computing capability. And, as to continuously improvement of its computational speed and data throughput capability, digital signal processor has been widely used in many video processing. In this paper, we have realized a video surveillance system based on DSP chip, which uses camera to capture video image and uses the DSP chip to make the fog video image clearness. Finally, display the processed image in the monitor equipment. The whole process fits the requirements of real-time process.First of all in this paper, we compared and analyzed three fog image clear methods:histogram equalization, the part over lapped histogram equalization (POSHE) and Retinex, and gives the MATLAB simulation results. And then, according to the relationship between neighbor frames of video image, we proposed a dynamic processing model based on moving target detection. In next part we gives the requirements of how to design and implement real-time fog-degraded video clearness system based on TMS320DM642. There are two parts of the system:hardware and software. The part of hardware achieves three modules of the design:the video decoder input, video cache and video encoder outputs. The part of software is the introductory chapters, which has discussed the design based on DSP/BIOS operating system and the various modules of standards-based software framework RF5. Finally, we give processing results based on this system, and analysis the processing the results.According to the experimental results, the proposed fog-degraded video image processing platform is stable, efficient, and has basically meet the requirements of the system design, which has broad application prospects.