Non-Homogeneous Haze Removal Based on Deep Unfolding Network for Remote Sensing Images

Le résumé fourni par la source

Optical remote sensing imaging has wide applications in many fields, but scattering effects will degrade the quality of images. To solve the challenge of removing non-homogeneous dense haze from the remote sensing images, we propose a novel alternating iterative dehazing algorithm based on the atmospheric scattering model. The algorithm employs TV regularization, which has excellent edge-preserving capability and is easy to implement, while incorporating gradient descent to enhance model stability and convergence. Building upon this algorithm, we further designed the deep unfolding network architecture. By incorporating the attention mechanism, the multi-scale pyramid module, and the post-processing module, the dehazing performance of the network is significantly improved. Experimental results show that the proposed network exhibits robustness in handling non-homogeneous dehazing and can effectively preserve the target information in remote sensing images.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.