Image Dehazing in Wide Field of View Based on Dual-tree Complex Wavelet Transform
Xiya Lian, Zhenhong Jia, Gang Zhou, Jiajia Wang, Sensen Song, Xiaohui Huang, Fei Shi · 2024
With the development of science and technology, surveillance equipment that captures wide-field scenes has become common. In order to solve the problem of partial loss of details and distortion of color restoration after hazy images in wide field of view scenes are dehazed, we proposed a multi-scale image dehazing algorithm based on dual-tree complex wavelet transform. First, in view of the characteristics of wide field of view scenes that contain a large amount of near and distant view information, a method based on depth of field and gradient is proposed to optimize the rough transmittance map. Then, the dual-tree complex wavelet transform is used to decompose the image into low-frequency and high-frequency images in six directions. In the high-frequency part, a linear weighted scale factor is used to enhance the details of the high-frequency image. The edges of the transmittance map obtained by this algorithm are clearly visible, and more texture details can be retained. The foggy image after dehazing is clear and the colors are more natural. The algorithm was experimented with a real wide-field foggy data set. Comparative studies and quantitative evaluations demonstrated the superiority of this method in processing wide-field images.