可见光遥感成像性能指标与图像特征参量相关性分析  

Correlation analysis between imaging performance index and image features of visible remote sensing

作　　者：徐宁珊 任国强[1] 黄永梅[1] Xu Ningshan;Ren Guoqiang;Huang Yongmei(Institute of Optics and Electronics,Chinese Academy of Sciences,Chengdu 6102Q9,China;The School of Electronic,Electrical and Communication Engineering,University of Chinese Academy of Sciences,Beijing 100049,China)

机构地区：[1]中国科学院光电技术研究所,成都610209 [2]中国科学院大学电子信息与通信工程学院,北京100049

出　　处：《国外电子测量技术》2021年第5期114-120,共7页Foreign Electronic Measurement Technology

基　　金：国家重点研发计划(2016YFB0500201)项目资助。

摘　　要：无参考图像质量评价是遥感领域对图像质量进行表征和评价的一个新型研究方向,然而,这类方法在实际遥感应用中能否取代传统的成像性能指标评价方法还有待验证。从遥感应用的角度出发,对图像自然场景统计(natural scene statistics,NSS)特征参量与MTF成像性能指标在图像清晰度表征中的一致性进行分析与讨论,设计并进行了相关性验证实验,在两个不同的图像测试集上分别取得了均值为0.8085、0.8568以及最大值为0.9143、1.0000的相关系数结果,从而验证了NSS图像特征参量与MTF指标具有强相关性。这一结论为无参考图像质量评价方法在遥感工程应用中的推广和实现提供了理论支撑和指导意义。The no-reference image quality assessment method is a new research direction in the field of remote sensing to characterize and evaluate image quality.However,whether such methods can take the same effect as the traditional imaging performance evaluation methods in actual remote sensing applications remains to be verified.From the perspective of remote sensing applications,this study analyzes and discusses the consistency between the natural scene statistics(NSS)feature parameters of the image and the MTF imaging performance indicators in the image definition characterization,and designs and conducts correlation verification experiments.The correlation coefficient results with mean values of 0.8085 and 0.8568 and maximum values of 0.9143 and 1.0000 were obtained on two different image test sets,thus verifying that the NSS image feature parameters have a strong correlation with the MTF index.This conclusion provides theoretical support and guiding significance for the promotion and realization of the no-reference image quality assessment method in the application of remote sensing engineering.

关 键 词：图像质量评价 MTF NSS特征 相关性分析 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]