Rain Removal Algorithm Based on Retinex in Low Light Automatic Driving Scenario

Abstract

In the field of automatic driving, clear image is one of the important prerequisites for the decision-making system relying on computer vision to make correct decisions. Rainy days and low illumination environment are two common scenes that seriously affect image quality. Researchers have proposed many effective methods for these two research fields. However, few applications have been studied in the field of automatic driving comprehensively. To solve the problem, a low illumination image enhancement method in rainy days is proposed, which combines the two scenes and can be applied to the field of automatic driving. This algorithm takes good advantage of the fact that raindrops cannot appear in the same frame at different times to eliminate the noise of raindrops in the image .And its main process is to fuse continuous image frames by maximum method through multi-scale Gaussian filtering. Thus, the fused image is used as the illumination map and the original image to be denoised is used as the input of Multi-Scale Retinex(MSR) method.Finally, the reflection image which can be directly applied in automatic driving is obtained.