图像作为人类感知世界最直观信息, 为我们日常生活带来极大便利。然而, 受雾霾、光照差等环境影响, 实际成像存在朦胧不清晰等问题。当前对图像去雾增强处理以事后分析居多, 难以满足跟踪等实施性要求较高的应用需求。针对此问题, 文章设计一款基于DSP+FPGA的实时去雾增强系统。该系统集合了DSP大型数据处理能力和FPGA高实时性的优点, 很好地将改进型“暗通道”原理算法移植其中。有效满足了实时去雾增强功能要求, 并已应用于实际工程项目。实验证明, 该系统工作稳定有效, 能有效解决实际工程中雾霾、光照不足给成像造成的影响。

As a kind of the most intuitive information, image has brought great convenience to our daily life. However, due to haze and poor light environment impacts, actual image usually have blurred clarity and other issues. Currently, image defog and enhance processing is mostly in hindsight, it is difficult to meet the requirements for tracking and other high real-time requirements applications. To solve this problem,a real-time image defogging and enhancement system is designed based on DSP+FPGA. This paper collects the DSPs large-scale data processing capability and FPGAs high real-time advantage, and skillfully transplant the improved “Dark channel” principle algorithm. This real-time image defogging and enhancement systems has applied in practical projects, and effectively meets the requirements of real-time processing. Experimental results show that this system works stably and effectively, solves the problem caused by haze and insufficient light perfectly.