Dark channel prior defogging algorithm of multi-scale convolutional neural network

Abstract

The invention discloses a dark channel prior defogging algorithm for a multi-scale convolutional neural network, and the algorithm comprises the steps of carrying out the minimum filtering processing of an original fog image through employing a dark channel prior theory, and taking the mean value of illumination intensity at the positions of pixel points with the brightness at the top 0.1% in a dark color image as a global atmospheric light value; then, estimating the fine transmissivity by using a multi-scale convolutional neural network, and the proposed algorithm being composed of a coarse scale network for carrying out global transmission graph estimation on the whole image and a fine scale network for locally refining the transmissivity; and finally, substituting the global atmospheric light value and the refined transmissivity into an atmospheric scattering model to recover a fogless image. The algorithm not only can effectively remove haze in the image and improve the visual effect of the image, but also can effectively reduce the complexity of the algorithm and shorten the calculation time, so that the defogged image is closer to a real image.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a dark channel prior defogging algorithm of a multi-scale convolutional neural network. Background technique [0002] In severe weather conditions such as fog and thunder, the suspended particles in the air will affect the outdoor shooting of optical imaging equipment. The absorption, scattering and refraction of reflected light by the turbid medium will cause the incident light of the camera to obtain the scene image to attenuate. As a result, the imaging quality is cloudy, which seriously affects the extraction of valuable information from the surrounding environment by the camera. [0003] The method of dehazing based on traditional image enhancement technology usually does not consider the problem of image degradation caused by the attenuation of incident light in foggy conditions, but directly uses image enhancement technology to properly adjust the contrast, bright...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a multi-scale convolutional neural network dark channel prior defogging algorithm, aiming to solve the technical problems of poor defogging effect and long defogging time of the defogging method in the prior art

Images