Image Dehazing Based on Multi-scale Retinex and Guided Filtering
Gao Zhihui, Yishu Zhai · 2022 International Conference on Image Processing, Computer Vision and Machine Learning (ICICML) · 2022
Haze removing from single image is a challenging task in the field of computer vision. Aiming at the problems of low contrast, details loss and serious color distortion, this paper proposed an effective dehazing method. Firstly, the original image is decomposed into reflective image channel and illumination image channel by multi-scale retinex theory; Secondly, the traditional histogram equalization and guided filter are used in the reflective image channel to enhance contrast and compensate details from guided image; Then, 2-D gamma transform in the illumination image channel for light compensation; Finally, white balance on composite image to solve the problem of color distortion. Experimental results demonstrate that the proposed algorithm can efficiently enhance image contrast, as well as improving color visual effects and details of images.