Image defogging method based on a generative adversarial network

Abstract

The invention discloses an image defogging method based on a generative adversarial network. The method comprises the following steps: 1) obtaining sample data; 2) taking a real foggy image in the sample data as input data of a first generator, wherein the first generator generates a primary fogless image; wherein a real fog-free image in the sample data is used as input data of a second generator, and the second generator generates a primary foggy image; the first discriminator feeds back an error between the primary foggy image and the real foggy image to the second generator, the second discriminator feeds back an error between the primary foggy image and the real fogless image to the first generator, the second generator and the first generator reduce the error, and the truth of the generated image is improved; the generator and the discriminator perform repeated confrontation training to obtain an optimal defogging network model; and (3) performing image defogging. A generative adversarial network structure and a loss function are adopted, network training does not need a foggy-fog-free matching image of the same scene, and meanwhile it is guaranteed that the color of the image is not distorted before and after defogging.

Description

technical field [0001] The invention relates to an image defogging method, in particular to an image defogging method based on a generative confrontation network, and belongs to the technical field of computer graphics and image processing. Background technique [0002] With the advancement of science and technology, a large number of outdoor digital images are collected and analyzed for various scientific research and production practices, such as target detection, terrain classification, outdoor photography, etc. However, due to the existence of water vapor or suspended particles in the air in the outdoor environment, the images collected outdoors are often accompanied by fog or haze, which causes a series of degradation phenomena such as reduced image contrast, missing parts of the scene, and color shifts. The effective information of the image is a great obstacle. Therefore, it is very important and necessary to find an effective method for digital image defogging, and ...

AI Technical Summary

Problems solved by technology

However, the algorithm still has problems such as color degradation and insufficient transmittance estimation. More research based on DCP aims to solve these problems.

Although the image defogging algorithm develops rapidly, due to the under-constrained characteristics of the image defogging problem itself, the prior information is not sufficient, and various prior assumptions are often accompanied by new problems when solving a certain type of problem. It is very difficult to artificially analyze and find an accurate prior model

Images