针对现有图像去雾算法时间复杂度高, 难于实时处理的问题, 提出一款基于片上可编程系统(SoPC)的视频图像快速去雾系统。介绍了平台的软件流程和硬件结构, 对暗原色先验和导向滤波理论进行深入研究, 并将暗原色先验和导向滤波算法移植到所设计的 SoPC平台上进行测试。测试结果表明, 该系统能满足视频图像实时去雾的要求, 并且恢复的图像清晰, 对比度好, 颜色自然。

A System on a Programmable Chip(SoPC) based fast dehazing system for video image is proposed, aiming at the problems of high time-complexity and difficulties in real-time processing for existing algorithms. Firstly, the hardware structure and software process of the SoPC based platform are introduced. And then the algorithms of dark channel prior and guided filter are studied. At last, the algorithms are transplanted to the proposed platform. The experiment demonstrates that the system can meet the requirement of the real-time dehazing effectively, and restored images are pretty clear with fine contrast and natural color.