Road traffic light-stained image enhancement method and device based on adversarial network

Abstract

The invention provides a light-stained image enhancement method and device, relates to the technical field of computer application and the traffic field, and aims to enhance low-light images and process image enhancement recovery of strong light source point pollution. The method comprises the following steps: acquiring a light fouling image, and generating a weight map of the light fouling image; inputting the light fouling image and the weight map into an adversarial neural network; and generating an enhanced image of the light fouling image through a generator of the adversarial neural network, wherein the adversarial neural network comprises the generator and a discriminator, the generator introduces a local attention mechanism to enhance the performance of the strong light source point in the generated image, and the discriminator is used for discriminating whether the image generated by the generator is a corresponding high-quality image or not when the adversarial neural network is trained, and performing feedback by means of a loss function and back propagation to realize adversarial learning optimization of the generator and the discriminator.

Description

technical field [0001] The present application relates to the field of computer application technology and the field of transportation, in particular to a method and device for enhancing road traffic light-defaced images based on an adversarial network. Background technique [0002] Images captured in low-light environments with insufficient lighting often suffer from low visibility, low contrast, and noise. This type of low-light image can be repaired by low-light image enhancement technology. Image restoration and enhancement technology is the task of the bottom layer and preprocessing stage of computer vision. Image restoration and enhancement technology can transform low-quality images such as blurred, underexposed, and overexposed images into clear, high-contrast high-quality images. [0003] However, the existing low-light image enhancement technology is usually only applicable to images in a low-light environment as a whole, and cannot be used for light pollution phe...

AI Technical Summary

Problems solved by technology

[0003] However, the existing low-light image enhancement technology is usually only applicable to images in a low-light environment as a whole, and cannot be used for light pollution phenomena such as low visibility and loss of objects caused by strong light sources to images under the condition of low light and strong light sources at the same time. Effectively deal with

Especially in image processing in the field of traffic, it often happens that, for example, images taken in the dark at night with oncoming headlights, the headlights in the image are overexposed, while other areas tend to be dark, which is not conducive to traffic Further advanced operations such as image processing and recognition in the field

Images