基于FPGA的近红外实时透雾成像方法

针对雾天图像的退化现象，提出采用近红外波段成像和视频图像处理技术相结合的透雾成像方法。根据视频图像相邻两场直方图相似性特点，提出了改进的直方图均衡化算法处理雾天降质图像。改进算法简化系统结构，降低逻辑设计复杂度，节约了高速存储器部分硬件资源。算法经过MATLAB仿真验证，并在FPGA视频图像处理平台上硬件实现。实验证明改进算法增强雾天图像的同时满足标准视频图像处理25帧/秒的实时性要求，结合近红外波段透雾能力使雾天能见距离提高1.5倍以上。

Near infrared real-time penetrating fog imaging based on FPGA

According to image degeneration on foggy days, a penetrating fog imaging method using near infrared imaging and video image processing is proposed. According to the characteristics of low contrast image on foggy days and high comparability of histogram of contiguous field in video image, we have proposed an improved histogram equalization algorithm.This algorithm simplifies the system architecture as well as reduces the complexity of logical design. Moreover it also reduces the hardware resources of high speed memory. It has been simulated and verified on MATLAB and implemented on the video image processing platform based on FPGA. The experiment shows that the near infrared wave band has a better penetrating ability on fog. The improved algorithm enhances the image quality on foggy days while meet the real-time requirement for currently standard video image of 25 frame/s. With these two methods, we increase the visibility distance over 1.5 times on foggy days.