Optical Remote Sensing Image Defogging Algorithm based on Different Wavelength
Shuo Qu, Zhuowei Wang, Genping Zhao, Jiahui Wu, Fuhua Zhang · 2023
The shooting conditions of remote sensing images often cause them to be covered by clouds and fog, and remote sensing images handle more details more strictly, so it is very difficult to defog remote sensing images. In this paper, we propose an optical remote sensing image defogging algorithm based on different wavelengths. Firstly, the algorithm increases the transmittance attenuation value of the three channels to the initial energy value of the image based on different wavelength attenuation. Secondly, we use deep learning methods to segment the dense fog areas and determine the global atmosphere light candidate area to estimate the global atmosphere light. Finally, the proposed algorithm uses the atmospheric scattering model to defog the three channels respectively, and the final experimental results are obtained. Experimental results show that our method can achieve better defogging image effect and quality.