FPGA Implementation of Low-light Enhancement and Defogging for Video Images

In order to realize the enhancement of low-light and foggy video images, a real-time enhancement defogging system based on FPGA is established. Low-light enhancement, image defogging and clock converter modules adopted in this design are optimized using parallel criteria. Firstly, we adopted a structure with shift operation to achieve low-light enhancement. Then we divide the algorithm into two parallel progresses of light value estimation and transmissivity calculation based on the deep analysis of dark channel prior defogging algorithm. Last, the image recover module calculates the recovered images based on the known parameters. Experiments indicate that, the quality of output image has been significantly improved, and the system can handle 65 frames video image with size 640×480 per second. It can satisfy the system requirements of enhancing image qualities, fast processing, and real time.