A Fast Video Haze Removal Algorithm Via Dark Channel Prior
Nanjun Ma, Jianbo Xu, Hongchao Li · Procedia Computer Science · 2018
The existed-algorithms has a poor real-time performance, and need high-configuration device to implement, in term of high-resolution video.For resolving above problems, this paper presents a fast video haze removal algorithm base on dark channel prior to improve the performance of algorithm. The dark-channel-prior-based haze removal algorithm, the imaging model in fog weather and the optimizing transmissivity are important technical support. Fast video haze removal algorithm firstly, uses guided filtering technology, which fix the problem of massive RAM storage, replace the soft matting approach. Then, it explores that the methods of down sampling and interpolation to measure the transmissivity, which can decline the time on images defogging. compared with the algorithm of MSRCR. Finally, the experimental results show that the improved algorithm our proposed can improve the speed of video defogging, and reduce the cost of equipment on image and video processing.