An Algorithm for Image Restoration Based on Underwater Video Series

摘要

The inspection for nuclear power facilities has been paid attention all the time. In underwater NP videos exit waggling water flow and much snowflake noise whose positions are undetermined. The former results from the cold-hot water with different temperatures, and the latter maybe are from high radiation. All these severely influence recognition for objective structures and flaws in the nuclear power facilities. In order to get clear images about the nuclear power facilities, in this paper we put forward an algorithm for image restoration based on underwater video series, in which, each frame is de-noised firstly, and then these images are registered and fused, finally the fused image is enhanced by histogram equalization or homomorphic filtering according to the requirement. The effectiveness of this algorithm is proved by an example in processing a series of underwater videos.