Two-channel Image Dehazing Algorithm Based on the Improved Guided Filter for Intelligent Automobile

Abstract

This work proposes a two-channel image dehazing strategy based on threshold segmentation. The algorithm first uses the Otsu threshold method to segment the sky and the background, and then estimates the dark channel and bright channel atmospheric light values of the sky and background respectively, and uses a weighting function to fuse them, and finally obtains the accurate mixed atmospheric light of each area. In order to improve the detail effect of the restored image, an improved guided filtering technique is proposed to further process the obtained transmission image. Experiment tests show the performance of the proposed method.