植被指数纹理特征信息估测稀疏植被生物量  

Estimation of Sparse Vegetation Biomass Based on Grey-level Co-occurrence Matrix of Vegetation Indices

作　　者：牧其尔[1] 高志海[2] 包玉海[1] 王琫瑜[2] 白黎娜[2] 

机构地区：[1]内蒙古师范大学地理科学学院,呼和浩特010022 [2]中国林业科学研究院资源信息研究所,北京100091

出　　处：《遥感信息》2016年第1期58-63,共6页Remote Sensing Information

基　　金：国防科工委重大专项(21-Y30B05-9001-13/15)

摘　　要：针对基于反射率及植被指数的统计模型在估测荒漠化地区稀疏植被生物量时往往无法满足要求的问题,该文以RapidEye多光谱影像为数据源,尝试通过植被指数与纹理信息相结合的方法对甘肃省民勤县绿洲边缘稀疏植被区的生物量进行估测。选取4种典型植被指数,按照不同参数分别提取植被指数的灰度共生矩阵纹理特征值。将植被指数与实测生物量数据进行线性回归,同时,将纹理特征与生物量构建多元逐步回归模型。比较了单一的光谱信息与纹理信息估测荒漠化地区生物量的能力。研究结果表明,利用植被指数的纹理特征估测生物量的能力较高于单一的植被指数。Estimating biomass of sparse vegetation in desertified areas based on the reflectance and vegetation index statistical models does not meet the requirements. Therefore, research on remote sensing image texture information combined with spectral information sparse vegetation biomass estimation method is important. The techniques for sparse vegetation biomass estimation using RapidEye multispectral satellite image were studied in oasis Minqin county, Gansu province. Spectral derivatives including typical vegetation indices and textural information of Grey-level Co-occurrence Matrix （GLCM） indices extracted from different parameters were calculated to develop models separately using stepwise multiple-linear regression. By comparing these models,it can be found the texture indices of vegetation indices can estimate the sparse vegetation biomass more accurately than single vegetation index.

关 键 词：RapidEye 植被指数 灰度共生矩阵 生物量 纹理 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]