大气散射校正静态实时偏振图像去雾技术

发布时间：2018-01-25 16:44

本文关键词： 图像去雾大气散射大气溶胶偏振分析图像复原　出处：《北京邮电大学学报》2016年06期 　论文类型：期刊论文

【摘要】：为了在雾霾天气下实现静态实时图像去雾,研究了大气衰减模型,分析了大气溶胶对成像的影响和大气散的射校正方法,提出了偏振去雾方法的技术原理.利用斯托克斯参数和穆勒矩阵实现原始数据的静态获取,基于偏振测量结果实现天空区域的实时提取,通过提取结果定性和定量分析,找到了偏振角范围和大气散射光强度阈值的规律性,实现了静态实时偏振去雾.静态实时偏振去雾方法将原始数据由不确定的偏振图像改变为固定的偏振图像,克服了对手动操作和主观评价的依赖,简化了去雾流程,解决了复杂的全局参数估计问题,提高了图像复原的准确度.所提出的静态实时偏振去雾方法为雾霾天气下高分辨、高稳定性和高实时性成像提供了新的理论和技术手段.
[Abstract]:In order to realize the de-fogging of static real-time image in haze weather, the atmospheric attenuation model is studied, and the influence of large aerosol on the imaging and the emissive correction method of atmospheric dispersion are analyzed. In this paper, the principle of polarization defogging method is put forward. Stokes parameter and Muller matrix are used to obtain the original data statically, and the real-time extraction of sky region is realized based on the results of polarization measurement. Through the qualitative and quantitative analysis of the extracted results, the regularity of the polarization angle range and the threshold value of atmospheric scattering light intensity is found. The static real-time polarization de-fogging method changes the original data from the uncertain polarization image to the fixed polarization image, which overcomes the dependence on manual operation and subjective evaluation. It simplifies the defogging process, solves the complex global parameter estimation problem, and improves the accuracy of image restoration. The proposed static real-time polarization defog method is high resolution in haze weather. High stability and high real-time imaging provide new theoretical and technical means.
【作者单位】：中国科学院西安光学精密机械研究所中国科学院光谱成像技术重点实验室;中国科学院大学;
【基金】：国家高技术研究发展计划(863计划)项目(20137031071B) 国家自然科学基金项目(11327303)
【分类号】：TP391.41
【正文快照】： 目前雾霾问题十分严重[1],普通可见光成像难以获得清晰图像.去雾成像方法分为2大类:图像增强方法[2-3]和图像复原方法,如偏振成像[4-5].偏振去雾是基于大气衰减物理模型的图像复原,可恢复细节信息.Schechner等[6]提出图像偏振去雾,但该方法无法满足静态实时去雾需求.针对上述

【相似文献】