Multi-mode detection techniques of video visibility based on improved dual differential luminance algorithm

The traditional dual differential luminance algorithm was influenced by the surrounding environment, and was not applicable to accurate detection at night. In view of the interference factors of reality, this paper improved the dual differential luminance algorithm by using the method of amending coefficient. Meanwhile, combined with the traffic emergency control measures under different visibility, ten detection modes were subdivided. The next, an experimental system was built in the Meteorological Observatory of Beijing, and the detection software was developed. At last, the experimental data was compared with the correlation values of scatterometer. The results show that the detection accuracies are raised, and the optimized algorithm and system can be used for highway network visibility detection. © 2016 SERSC.