Abstract

This paper presents a video adaptive enhancement method based on Retinex illumination theory. The bidirectional equalization algorithm is adopted to make the gray distribution more uniform and to suppress the noise to a certain extent. The self-adaptive enhancement function stretches the radiation component and the overall component and improves the global and local visual effects. In the splicing process of the reflection component, the self-adaptive weighted method is effective to solve the fuzzy block effect, and the introduction of the similarity between frames validly improves the processing efficiency of the algorithm. Experiments have demonstrated that the proposed method can effectively deal with low-contrast video, and the enhancement effect is obvious, fast and effective.

Authors and Affiliations