Image Defogging Based on Joint Contrast Enhancement and Multi-scale Fusion

Abstract

Although the single image dehazing techniques now in use work well in certain circumstances, real-world scenarios' complex and ever-changing nature makes it impossible for them to function well. This research offers a contrast enhancement and exposure fusion-based efficient paradigm for picture dehazing. Three basic preprocessing techniques are utilized to create intermediate pictures for contrast enhancement. The first processing method is gamma correction (GC), employed to improve an image's local quality. The second technique is the newly developed color-preserving histogram equalization, which combines each channel to enhance the overall contrast. The third kind of processing is called ACE, which uses adaptability to deal with a complex and shifting environment. This paper fuses the input from GC, CPAHE, and ACE during the fusion stage using the adaptive kernel size fusion technique based on rapid structural patch decomposition. According to the experimental findings, it outperforms other cutting-edge visual and quantitative analysis techniques. The average FADE scores for the daylight and night-time scenes have increased by at least 0.09 and 0.08, respectively. This paper also describes how this strategy may be used in various contexts.