Review of polarimetric image denoising
Hedong Liu, Xiaobo Li, Zihan Wang, Yizhao Huang, Jingsheng Zhai, Haofeng Hu
Abstract
Polarimetric imaging, leveraging measurements of polarimetric parameters that encode distinct physical properties, finds wide applications across diverse domains.However, some critical polarization information is highly sensitive to noise, and denoising polarimetric images while preserving polarization information remains a challenge.The development of denoising techniques for polarized images can be roughly divided into three stages: The first stage involves the direct application of traditional image denoising algorithms, such as spatial/transform domain filtering.The second stage involves specially designed methods for polarized images, using image prior models for noise removal, such as principal component analysis and K-singular value decomposition.In the third stage, benefiting from advances in deep learning, denoising methods tend to integrate polarization characteristics with deep learning models for noise suppression.The residual dense network, U-Net, and other effective models are appropriately modified and supervised/self-supervised trained to handle the denoising problem of regular/extensive polarimetric images.In this paper, we perform a comparative study of polarimetric image denoising methods.These methods are first classified as learning-based and traditional methods.Then, the motivations and principles of different types of denoising methods are analyzed.Finally, some potential challenges and directions for future research are pointed out.