冬小麦高光谱信息提取方法的研究  

The Study of Winter Wheat Hyperspectral Information Extraction Method

作　　者：李志花[1] 冯美臣[1] 王超[1] 赵佳佳[1] 王慧琴[1] 刘婷婷[1] 杨武德[1] 

机构地区：[1]山西农业大学旱作农业工程研究所,山西太谷030801

出　　处：《山西农业大学学报（自然科学版）》2015年第5期467-473,共7页Journal of Shanxi Agricultural University(Natural Science Edition)

基　　金：国家自然科学基金项目(31371572;31201168);山西省科技攻关项目(20110311038);山西省青年基金项目(2012021023-5)

摘　　要：针对高光谱数据波段多、数据量大和冗余度大等特点，本文以 ASD 便携式高光谱仪为光谱数据获取手段，以不同生育时期冬小麦冠层高光谱为研究对象，采用主成分分析和波段自相关分析两种方法来进行数据降维，通过采用基于高光谱与叶面积指数估算（LAI）相关系数法进行验证，以确定冬小麦高光谱遥感信息提取的最佳波段。结果表明，主成分分析法（PCA）和波段自相关分析法选择的波段主要在可见光区域（350～450 nm 和600～700 nm）、近红外区域（1100～1200 nm）和短波红外区域（1500～1750 nm），包含了验证方法基于高光谱与 LAI 相关系数法所选择的主要波段范围：可见光区域和近红外区域。综合考虑，用主成分分析法和波段自相关分析法两种方法对冬小麦高光谱提取的信息是全面、可靠的，包含了针对某一生理指标如 LAI 的有关信息。In accordance with the high spectral data band ,the large amount of data and the redundancy of the high data , the method of principal component analysis （PCA） and band intercorrelation analysis were used in the paper. The ex‐periment was conducted to obtain the hyperspectral data under different growth stages of winter wheat. Moreover ,the extracted hyperspectral information was validated by the sensitive bands of leaf area index （LAI） selected with the method of correlative analysis. The results showed that the hyperspectral information selected with methods of principal component analysis （PCA） and intercorrelation analysis are mainly centered in the area of 350 ～ 450 nm ,600 ～ 700 nm , 1 100 ～ 1 200 nm and 1 500 ～ 1 750 nm which covered the visible ,near‐infrared and shortwave infrared bands. To vali‐date the selected hyperspectral information ,the sensitive bands of LAI extracted with the correlative coefficient analysis were contained in the hyperspectral information of winter wheat. The paper indicated that the method of PCA and cor‐relation analysis was available and reliable in reducing the hyperspectral redundancy and extracting hyperspectral infor‐mation of winter wheat.

关 键 词：高光谱遥感 信息提取 主成分分析 自相关分析 

分 类 号：S12[农业科学—农业基础科学] S126