Single Image Dehazing Jointly Utilizing Dark Channel Prior and Guided Filtering in Dual-Tree Complex Wavelet Domain
Yue Dongqiao, Huang Yuanyuan, Haozhe Tang, Xi Huang, Zhi Qin · 2020
In this paper, a new image haze removal algorithm is proposed by jointly utilizing dark channel prior and guided filtering in dual-tree complex wavelet transform domain. The algorithm firstly uses guided filtering to smooth the hazing image, then adopts dual-tree complex wavelet transform to decompose the hazing image into low-frequency and high-frequency sub-band components, and then uses the dark channel prior model to process the low-frequency sub-band of the image, finally adopts inverse dual-tree complex wavelet transform to get the dehazed image. The dark channel prior comes from the statistical law of the outdoor fog-free image datasets. Since dual-tree complex wavelet transform has approximate translation advantages such as degeneration, higher positioning accuracy, computational efficiency, and guided filtering can increase the scope of application of haze removal, which can greatly reduce the optimization time of initial transmittance and the complexity of the entire algorithm. Therefore, the new proposed algorithm outperforms traditional dark channel prior based defogging algorithm. The experimental results show that the proposed algorithm can effectively improve the contrast and clarity of the image.