Aerial image defogging method based on improved generative adversarial network

Abstract

The invention discloses an aerial image defogging method based on an improved generative adversarial network. The method comprises the following steps: establishing a data set; inputting the foggy sample image into a generation network to carry out defogging processing; inputting the sample image after defogging processing and the corresponding fog-free sample image into an adversarial network, carrying out threshold and true and false discrimination, and calculating a loss function model parameter; feeding back the loss function model parameters to the generation network, and updating the generation network model by the generation network; repeating the steps to obtain a training model; and inputting the foggy picture into the training model to obtain a fogless picture. The invention aimsat the defects in the existing image defogging method. According to the aerial image defogging method based on the improved generative adversarial network, on the basis of an original generative adversarial network, a network structure and a calculation method are improved, it is not needed to assume prior conditions such as a model or a scene depth during defogging, a foggy image is directly input into the network, the speed is high, and the defogging effect is good.

Description

technical field [0001] The invention relates to an aerial image defogging method, in particular to an aerial image defogging method based on an improved generation confrontation network. Background technique [0002] With the rapid development of the Internet and information processing technology, people have higher and higher requirements for image clarity. However, due to the limitations of physical imaging conditions and the acquisition environment, images are often disturbed by varying degrees of fog when collecting images. In natural fog and haze weather, there will be a large number of small water droplets and small dust particles in the atmosphere. There are problems that will increase the scattering of light, which will reduce the contrast of the image captured by the outdoor image acquisition sensor, narrow the dynamic range, decrease the definition, not rich in color, cover up some details, and even distort the color, which adds a lot to the image. Large noise dir...

AI Technical Summary

Problems solved by technology

This method has many steps, complicated operation, and poor efficiency in practical application.

[0015] Researchers such as Tang Huanrong directly input the image into the generative adversarial network for processing, but its network structure is relatively complex, with two generators and two discriminators respectively, the calculation amount of the defogging operation is greatly increased, and the processing speed is slow

Images