基于三维融合策略的高光谱遥感影像全变分去噪  

Hyperspectral image total variation denoising based on 3D fusion strategy

作　　者：马文[1,2] 史潇 岳建平[1] MA Wen;SHI Xiao;YUE Jianping(School of Earth Sciences and Engineering, Hohai University, Nanjing, Jiangsu 210098, China;School of Remote Sensing, Nanjing University o{ In{ormation Science ~ Technology, Nanjing, Jiangsu 210044, China;Jiangsu Meteorological Service Center, Nanjing, Jiangsu 210008, China)

机构地区：[1]河海大学地球科学与工程学院,江苏南京210098 [2]南京信息工程大学地理与遥感学院,江苏南京210044 [3]江苏省气象服务中心,江苏南京210008

出　　处：《中国矿业大学学报》2018年第3期678-684,共7页Journal of China University of Mining & Technology

基　　金：国家自然科学基金项目(41471312)

摘　　要：高光谱遥感影像可以提供更细致的地物光谱,但受物理条件限制,也更易受到噪声干扰.基于高光谱遥感影像去噪的重要性,探讨了利用融合策略提升高光谱影像去噪效果的可行性.根据高光谱影像数据三维空间特点,提出了利用Q指标加权平均方法,将已有的全变分高光谱影像去噪方法在3个不同视角条件下的结果进行融合,从而弱化单一视角下去噪方法对原始信息的破坏程度,加大了高质量的影像在融合过程中所占的比重,提高了每个融合波段的质量,有效融合互补信息,实现对原始无噪声影像更好的恢复.结果表明:在各种图像和噪声类型情况下,提出的方法相较于单一视角的高光谱遥感影像全变分去噪方法有明显提高,采用的三维融合策略简单有效,具有很高的鲁棒性.Hyperspectral images（HSIs）can offer more detailed spectral information while suffering from more disturbance of noise because of physical reasons.Due to the importance of HSI denoising,this work discussed the feasibility of using fusion strategy to improve the HSI denoising result.In this paper,based on the three dimensional properties of HSIs,Q metric with weighted average method was adopted to combine the results produced by available total variation HSI denoising method in three different perspectives.By this way,the original information was better retained,and the complementary information is better combined.Furthermore,the method increased the proportion of high quality images in the process of fusion,and improves the image quality of each fusion band.Therefore better results can be expected.Experiments were conducted and the results show that the proposed method is better than the existing total variation method conducted in a single view,which indicates the effectiveness and robustness of the 3 Dfusion strategy.

关 键 词：高光谱影像 去噪 三维融合 Q指标 全变分 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]