Defogging device based on space and channel attention residual network and defogging method thereof

Abstract

The invention provides a defogging device based on a space and channel attention residual network and a defogging method thereof. The defogging device comprises a main calculation module, a circulation module, a first calculation module, a second calculation module, a third calculation module and a fourth calculation module, the defogging device is used for defogging the input image and outputting a defogged image; the input image and the defogged image are color images composed of RGB three-dimensional channels; according to the method and device, any atmospheric scattering model parameter does not need to be estimated, and weight values are adaptively allocated to different space pixels and different feature channels by analyzing the mutual relation between the different space pixels and the different feature channels, so that the neural network more focuses on analyzing and processing the space pixels and the feature channels with larger value and information amount, and the defogging capability of the deep neural network is effectively improved; and the restored image is clear in detail and rich in color.

Description

technical field [0001] The invention relates to the field of image defogging, in particular to a defogging device and a defogging method based on a spatial and channel attention residual network. Background technique [0002] In recent years, hazy weather has become more and more frequent, seriously affecting our daily production and life. The appearance of smog is a natural phenomenon, mainly caused by particles in the atmosphere absorbing scattered light. Under such weather conditions, optical equipment cannot obtain effective scene information, and the imaging quality is poor, which severely limits subsequent applications in video surveillance, automatic driving, satellite remote sensing and other fields. Therefore, how to effectively remove the haze in the image and restore the color and contrast of the image as much as possible without losing details or introducing additional interference information has important research significance. [0003] Image dehazing is an e...

AI Technical Summary

Problems solved by technology

[0004] 1. In the real environment, the causes of fog are intricate, and it is unreasonable to describe it through a simple linear mathematical formula;

[0005] 2. Many methods extract features from foggy images that are too single to cover complex scenes;

[0006] 3. The intermediate parameters of different atmospheric scattering models affect each other during parameter optimization, it is difficult to achieve the global optimum of all parameters, and the inaccurate estimation of parameters further reduces the performance of dehazing

Images