基于高光谱反射特征的荒漠类型可分性评价  

Class separability evaluation of desert types based on the hyperspectral reflectance characteristics

作　　者：刘志飞 杨雪梅 王景瑞[3,4] 黄轲盼 徐浩杰 LIU Zhifei;YANG Xuemei;WANG Jingrui;HUANG Kepan;XU Haojie(State Key Laboratory of Herbage Improvement and Grassland Agro-ecosystems,Key Laboratory of Grassland Livestock Industry Innovation,Ministry of Agriculture and Rural Affairs,Engineering Research Center of Grassland Industry,Ministry of Education,School of Pastoral Agriculture Science and Technology,Lanzhou University,Lanzhou 730020,Gansu,China;Tourism School,Lanzhou University of Arts and Science,Lanzhou 730010,Gansu,China;Gansu Desert Control Research Institute,Lanzhou 730070,Gansu,China;School of Resources and Environment,Lanzhou University,Lanzhou 730000,Gansu,China;Center for Remote Sensing of Ecological Environments in Cold and Arid Regions,Lanzhou University,Lanzhou 730000,Gansu,China)

机构地区：[1]兰州大学草种创新与草地农业生态系统全国重点实验室,农业农村部草牧业创新重点实验室,草地农业教育部工程研究中心,草地农业科技学院,甘肃兰州730020 [2]兰州文理学院旅游学院,甘肃兰州730010 [3]甘肃省治沙研究所,甘肃兰州730070 [4]兰州大学资源环境学院,甘肃兰州730000 [5]兰州大学寒旱区生态环境遥感研究中心,甘肃兰州730000

出　　处：《干旱区研究》2025年第1期141-153,共13页Arid Zone Research

基　　金：国家重点研发计划(2020YFA0608401);国家自然科学基金(32060373);甘肃省自然科学基金项目(22JR5RA766,23JRRA1048);中央高校基本科研业务费专项资金(lzujbky-2022-27)。

摘　　要：本研究采用裸土高光谱反射曲线细节所提取的特征变量评价不同荒漠类型可分性,选择石羊河下游盐漠、砾漠、泥漠和沙漠为研究对象,运用累积差值、一阶微分、包络线去除、植被指数计算、主成分分析法等,辨别不同荒漠类型的高光谱反射特征,提取关键分类变量,量化不同荒漠类型区分度。结果表明:(1)各荒漠类型在446~600 nm和2150~2285 nm处存在差异明显的吸收谷。(2)Carter指数1、绿度指数(GI)、绿色归一化植被指数(GNDVIh2)等在不同荒漠类型间存在显著差异。(3)改进叶绿素吸收指数(MCARI)、土壤调整植被指数(SAVI)、2265 nm与1790~1810 nm反射率在主成分指标构建中的权重值较大。(4)各荒漠类型区分度:沙漠&盐漠>沙漠&泥漠>泥漠&盐漠>砾漠&盐漠>沙漠&砾漠>泥漠&砾漠。研究结果可为西北干旱区荒漠遥感监测提供地面验证和数据支持。Few studies have used the characteristic variables extracted from the details of the hyperspectral reflectance curves of bare soil to evaluate the separability of various desert types.In this study,salt desert,gravel desert,mud desert,and desert in the lower reaches of the Shiyang River were used as the research objects,and cumulative difference,first-order differentiation,continuum removal,vegetation index calculation and principal component analysis were used to identify the hyperspectral reflectance features of various desert types,extract the key categorical variables,and quantify the degree of differentiation of various desert types.The results showed that(1)the absorption valleys at 446-600 nm and 2150-2285 nm differed significantly among the desert types.(2)the Carter index 1,Greenness Index,and Green NDVI hyper 2 differed significantly among the desert types.(3)The Modified Chlorophyll Absorption Ratio Index,Soil Adjusted Vegetation Index,and 2265 nm and 1790-1810 nm reflectance had larger weight values in constructing the principal component indexes;and(4)the differentiation index of each desert type:desert&salty desert>desert&muddy desert>muddy&salty desert>gravelly&salty desert>desert&gravelly desert>mud&gravelly desert.These findings provide ground verification and data support for the remote sensing monitoring of deserts in the northwest Arid Zone.

关 键 词：荒漠土壤 高光谱特征波段 特征提取 主成分分析 类别可分性 

分 类 号：P237[天文地球—摄影测量与遥感] X87[天文地球—测绘科学与技术]