Low-illumination image enhancement method and device based on conditional generative adversarial network

Abstract

The invention discloses a low-illumination image enhancement method and device based on a conditional generative adversarial network. In daily image acquisition work, a camera is used; in order to solve the problem that a low-illumination image is easily generated due to poor scene illumination conditions or insufficient light supplementing capability of equipment, a convolutional neural network model (CNN) with encoding and decoding functions is adopted as a generation model, and meanwhile, a CNN with a binary classification function is added as a discrimination model to form a generative adversarial network. In the process of model training, a real bright image is taken as a condition, the generation model is supervised by the discrimination model, and the mutual game between the discrimination model and the generation model is combined, so that the network model has better low-illumination image enhancement capability. According to the method, the problems of color distortion and edge blurring of the processed image are solved, the quality of the enhanced image is greatly improved, the image processing speed is high, compared with other image enhancement methods, the processingtime is short enough, and the requirement for real-time enhancement of the low-illumination image can be met.

Description

technical field [0001] The invention relates to the field of computer image processing, in particular to a low-illuminance image enhancement method and device based on a conditional generation confrontation network. Background technique [0002] With the development of technology and the popularization of video surveillance, cameras, mobile phones and other tools, people can obtain a large number of images more conveniently in today's production and life. However, due to the poor lighting conditions of the shooting scene or the lack of functions of the image acquisition equipment and other factors, it is very easy to produce low-light images. The overall imaging of these low-light images is gray or even completely dark, with low signal-to-noise ratio and extremely low use value, which seriously affects people's use. Therefore, some methods must be adopted to deal with low-illuminance images and restore the content of low-illuminance images clearly. [0003] Over the past f...

AI Technical Summary

Problems solved by technology

[0006] Aiming at the above technical problems, the present invention proposes a low-illuminance image enhancement method and device based on conditional generative adversarial network, which performs enhancement processing on low-illuminance images to solve the problems of color distortion and edge blur of the processed image, and improve the enhancement process. image quality, while speeding up image processing to meet the real-time enhancement requirements for low-light images

Images