Image Defogging Algorithm Based on Semi-Supervised Learning Neural Network
Xuan Chen,Jian-qing Shi,Fei Xie,Jun Wang,Zhilin Gao,Yue Yin
DOI: https://doi.org/10.1109/ICRAIC61978.2023.00023
2023-11-24
Abstract:To improve the quality of images acquired by imaging devices in bad weather such as haze, defogging algorithms are widely used to acquire clear images. This paper proposes two distinct image-defogging algorithms, based on semi-supervised learning, to address the issue of the limited performance of deep learning dehazing algorithms when using synthetic hazy images as their current dehazing datasets. To this end, the poor effect of real foggy images applied to network training is addressed. Generative adversarial networks, employed as an unsupervised branch, are initially utilized to introduce real foggy images into the model training and to train the network in an alternating fashion with synthetic datasets. Secondly, in the unsupervised branch, the fog residual K of the fogged image is estimated by AOD-net, and the defogged image is restored to a fogged image by the improved atmospheric scattering model in AOD-net. Finally, this paper compares the SOTS test set in the RESIDE dataset with five classical algorithms. The comparison algorithms in both PSNR and SSIM indicators are outdone by the algorithms proposed in this paper, as demonstrated by the experimental results.
Computer Science,Engineering,Environmental Science