A method of video defogging in computer

Abstract

The invention discloses a video defogging method in a computer, the steps of which include: step 1, extracting video key frames based on edge amplitude kernel density estimation information entropy of adjacent frame differences; step 2, extracting video key frames based on edge information amplitude information entropy Representative pixel points; step 3, register video key frames pixel by pixel based on cross-correlation matching algorithm and generalized alignment algorithm; step 4, estimate atmospheric illumination intensity based on color mean and optimize the propagation rate of video key frames based on two-dimensional kernel regression; step 4 5. Estimate the propagation rate of non-key frames of the video based on the Catmull-Rom cubic spline; Step 6. According to the atmospheric scattering model, use the estimated model parameters to output the restored haze video image sequence.

Description

technical field [0001] The invention belongs to the technical field of computer image processing, in particular to a video defogging method in a computer. Background technique [0002] Fields such as security systems, urban transportation, military technology, criminal investigation, navigation, meteorology, and astronomy often need to collect smog video image sequences from outdoor monitoring. Severe weather such as smog makes the images captured by video surveillance dim in color, low in contrast, and seriously degrades image quality, which directly affects the visual effect of images and seriously affects their applications. It has important research and practical significance to remove the influence of haze weather from the video and restore the color and contrast of the image. [0003] Now there are a variety of defogging methods for video: Wang Xiaotong of the Dalian Naval Academy, based on the depth of field distribution model, uses the dark primary color idea to obt...

AI Technical Summary

Problems solved by technology

[0005] Most of the existing video defogging methods mainly focus on processing each frame of the video separately, and do not use the redundant information between the frames of the video to reduce the amount of calculation, and there is a problem of relatively slow speed.

In addition, in the process of video defogging, the spatiotemporal consistency of the video is not considered, and the continuity and smoothness of the video frame and between frames cannot be maintained, resulting in flickering problems

Images