遥感影像融合评价定量指标选取问题  

Quantitative Index Selection for Effect Evaluation of Remote Sensing Image Fusion

作　　者：王恩鲁 汪小钦[2] WANG Enlu;WANG Xiaoqin(National Engineering Research Centre of Ceo spatial Information Technology,Fuzhou University,Fuzhou 350116,China;Key Laboratory of Spatial Data Mining & Information Sharing,Ministry of Education,Fuzhou 350116,China)

机构地区：[1]福州大学地理空间信息技术国家地方联合工程研究中心,福州350116 [2]空间数据挖掘与信息共享教育部重点实验室,福州350116

出　　处：《遥感信息》2017年第6期14-21,共8页Remote Sensing Information

基　　金：国家科技支撑计划课题(2013BAC08B01);福建省科技重点项目(2015N0025);中央引导地方发展专项(2017L3012)

摘　　要：鉴于目前遥感影像融合评价在定量指标的归类、选取和使用等方面尚未有较为全面、科学的标准和体系,从而定量融合评价工作中存在一些指标选用不明确、评价结果不一致的情况,该文将众多定量指标的融合评价结果进行分析,考察并遴选出具有代表性的定量指标作为参考。以WorldView-2影像、ZY-3影像和ALOS影像为数据源,选取常用的几种融合方法进行融合实验,将客观评价结果利用均值-标准差图的可视化方式进行分析,并与主观评判相结合,探讨遥感影像融合质量评价的最佳方案。实验结果表明,在评价光谱保持度方面,SAM与Q4/Q8评价结果之间的一致性最好,且与主观评判相符度最高;在评价影像清晰度方面,推荐使用角二阶矩和点锐度相结合的方式。The selection of the representative indexes is significant for the effective evaluation of image fusion.However,there is no comprehensive and scientific standard or system in this field currently.This paper analyzed a lot of quantitative evaluation outcomes to find the suitable indexes for fusion evaluation in the future.Several fusion experiments based on the common methods were carried out by using WorldView2,ZY3and ALOS images.For studying the selection of the optimal assessment method,the objective evaluation indexes with“mean std”charts combined with the subjective judgement by visual interpretation have been analyzed and evaluated in detail.The results which agreed well with those from subjective judgement indicated that both SAM and Q4/Q8are good for the spectral preservation evaluation,and angular second moment and point acutance arithmetic are recommended for image sharpness evaluation.

关 键 词：遥感影像融合 融合质量评价 定量指标 光谱保持度 影像清晰度 

分 类 号：P237[天文地球—摄影测量与遥感]