Image defogging method for haze concentration of adaptive neural network based on end-to-end

Abstract

The invention discloses an image defogging method for haze concentration of an adaptive neural network based on end-to-end. The method comprises the following steps: constructing an image defogging model; acquiring foggy image data; using a feature enhancement module in the image defogging model for cascading the feature map with images recovered by different paths, combining fuzzy images with different dense haze degrees to help the network sense image haze concentration in a self-adaptive mode; reconstructing the features after function enhancement into a clear fog-free image through a multi-scale feature attention module; calculating mean square errors and perception losses of the restored images and the corresponding clear images, and updating an image defogging model; wherein the meansquare errors guide the image defogging model to learn the content of the clear images, the perception loss is used for quantizing the visual difference between the recovered images and the corresponding clear images, and the two loss functions cooperatively optimize the defogging model. According to the technical scheme, effective defogging processing is carried out on the actually shot fog image, a high-quality image is recovered, and the practicability is good.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to an image defogging method based on an end-to-end haze concentration self-adaptive neural network. Background technique [0002] With the development of technology, some computer vision tasks such as object detection, object tracking, behavior analysis, face recognition, etc. have made great breakthroughs. However, advanced vision tasks such as detection and tracking rely on clear video and image data, and their performance is often greatly affected in actual scenes such as heavy fog and heavy rain. As a pre-task of some advanced vision tasks, image dehazing has attracted the attention of many researchers in recent years. [0003] Image dehazing is a typical image restoration problem, which can be traced back to 1924. McCartney et al. first proposed a classic atmospheric light scattering model. The model is as follows: [0004] I=tJ+A(1-t), [0005] t(x)=e βd(x) , ...

AI Technical Summary

Problems solved by technology

However, different priors depend on the estimation of a certain feature of the image. In real scenes, these prior information are often invalid and computationally complex.

Images