面向三江保护区植被类型识别的最佳波段组合研究  

Optimal Band Combination for Vegetation Type Recognition in Sanjiang Nature Reserve

作　　者：姜怡 刘华[1] 赵峰[1] 马永康 刘浩栋 JIANG Yi;LIU Hua;ZHAO Feng;MA Yong-kang;LIU Hao-dong(Research Institute of Forest Resource Information Techniques,Chinese Academy of Forestry,Beijing 100091,China;Key Laboratory of Forestry Remote Sensing and Information System of National Forestry and Grassland Administration,Beijing 100091,China)

机构地区：[1]中国林业科学研究院资源信息研究所,国家林业和草原局林业遥感与信息技术重点实验室,北京100091

出　　处：《林业科学研究》2022年第1期115-122,共8页Forest Research

基　　金：高分辨率对地观测系统重大专项(21-Y30B02-9001-19/22-2);中国林业科学研究院资源信息研究所结余资金资助(2017JYZJ16)。

摘　　要：[目的]针对高分六号影像数据面向植被类型、特别是湿地植被信息识别的最佳波段组合进行研究,为高分六号数据在湿地资源监测中的进一步应用提供参考。[方法]以三江国家级自然保护区为对象,基于高分六号影像数据,结合当地植被特点,开展保护区植被类型识别中的最佳波段组合研究。从影像信息特征、最佳指数因子以及光谱特征曲线和地物可分性角度进行综合分析,获取基于GF6/WFV数据的三江保护区植被类型识别的最佳波段组合。同时,通过对比实验,利用支持向量机分类方法,对样区的植被类型进行信息提取,证实了结果的可靠性。[结果]波段组合4(R)-5(G)-1(B)识别植被类型精度最高。总体精度为89.15%,Kappa系数为0.8463。[结论]波段组合4(R)-5(G)-1(B)为三江保护区植被类型识别的最佳波段组合。本研究可为东北地区湿地保护区的退耕还湿工程监测提供参考。[Objective]To find the best band combination of GF6 image data for the identification of vegetation types,especially wetland vegetation information,so as to provide reference for the further application of GF6 data in wetland resources monitoring.[Method]Taking Sanjiang Nature Reserve as example and basd on the GF6 data,the research mentioned above was conducted combing with the features of local vegetation.From the perspective of image information characteristics,the best index factor,the spectral characteristic curve and the separability of ground objects,the best band combination for vegetation type identification in Sanjiang Nature Reserve based on GF6 data were obtained.At the same time,through comparative experiment,the support vector machine classification method was used to extract the vegetation type information of the sample area,and the reliability of the results was confirmed.[Result]The band combination of 1-4-5 showed the highest accuracy in vegetation classification with 89.15%overall accuracy and 0.8463 Kappa coefficient.[Conclusion]The band combination of 4(R)-5(G)-1(B)is the optimal band combination for vegetation classification in Sanjiang Nature Reserve.This study can provide a reference for monitoring the returning farmland to wetland project in northeast China.

关 键 词：最佳波段组合 植被类型识别 GF6/WFV数据 三江保护区 

分 类 号：X87[环境科学与工程—环境工程]