Polarization imaging haze-penetrating method with adaptive non-uniform scattering characteristics
Year: 2026
Type: article
Abstract: Scattering removal for scenes in a haze environment is a key means to improve imaging quality. However, most existing polarization imaging defogging methods fail to establish an adaptive model for the interaction between targets and scattering in non-uniform haze environments, which easily leads to overexposure or incomplete defogging in dense haze regions. To address this defect, a polarization imaging haze-penetrating scheme with adaptive non-uniform scattering characteristics is proposed in this paper. Firstly, a polarization dual-channel variable-scale degradation imaging model is constructed. By analyzing the attenuation law of target light in the brightest and darkest polarization channels, the degradation process of target light in non-uniform scattering environments is accurately reconstructed, and the target polarization information is effectively preserved. Secondly, an adaptive scattering estimation method guided by polarization combination is designed, which fuses scene intensity information and degree of linear polarization (DoLP) constraints to alleviate the scattering estimation deviation caused by the extreme difference in non-uniform scattering intensity and improve the estimation accuracy of scattered light. Experimental results show that in scenes with high-concentration haze and non-uniform haze cluster interference, the imaging effect of the proposed method is superior to that of existing defogging methods based on the intensity domain or polarization domain. It can more clearly restore the texture details and hierarchical structure of the scene, and effectively suppress the overexposure phenomenon. The NIQE value is reduced by an average of 9.84%, the FADE value is reduced by an average of 50.12%, and the EME value is increased by an average of 14.51%.
Source: Optics Continuum
Authors: Chen Shen, Zhiguo Fan, Haihong Jin, Ke Zhao, Wenhong Gao
Language: English
Topic: Image Enhancement Techniques
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $1,372
Award: 大气偏振模式的罗盘特征解析与建模方法研究
OpenAlex ID: W7125401939