由于当前的计算机视觉系统对天气条件比较敏感，尤其是在能见度较低的雾天条件下，捕获的景物图像严重退化，这极大地限制并影响了户外系统效用的发挥。因此，为了保证计算机视觉系统能全天候正常工作，提高系统在恶劣天气下工作的可靠性，对雾天降质图像进行清晰化处理就显得尤为必要。本文首先对雾天降质图像的清晰化处理方法进行了分析研究，并以此为基础提出了一种新的基于联合双边滤波的单幅图像去雾算法，并基于FPGA平台实现了雾天环境下计算机视觉系统的降质图像清晰化处理过程。实验及硬件仿真结果可见：本文提出的去雾算法能有效完成有雾图像的复原，处理速度快，且恢复后的图像自然逼真、清晰度高；论文同时完成了所提算法在硬件平台的移植及实现。由于硬件平台资源有限，论文工作尚未能实现有雾视频的实时去雾处理，下一步的研究目标是优化算法的Verilog HDL代码，并实现实时视频去雾系统。

The current system of computer vision is more sensitive to weather conditions, especially in low visibility fog conditions, to capture the scene image degradation, which greatly restricted and influenced the effectiveness of outdoor surveillance system. Therefore, in order to ensure the computer vision system capable of all-weather work, improving the system reliability under adverse weather conditions , the fog-degraded image clearness processing is particularly necessary.In this paper, firstly, fog-degraded image clearness processing methods were analyzed, and on this basis a single image to fog algorithm based on the joint bilateral filtering was put forward, and then based on use the FPGA platform realizes the fog environment of computer vision system of the degraded image clearness processing process.Experiments and hardware simulation results of visible: this paper proposed a defogging algorithm can effectively accomplish the fog image recovery, fast processing speed, and the recovered image is natural and vivid, definition is high; at the same time, the proposed algorithm performed in hardware platform transplant and realize. Due to hardware platform of limited resources, research work is not yet to be achieved with fog real-time video processing to the fog, the next step of the research goal is to optimize the algorithm Verilog HDL code, and the realization of real-time video to defog system.