基于无人机遥感的黄河下游河南段河滩地植物群落多样性研究  

Diversity of Plant Community in Flood Land of Henan Section of the Lower Yellow River based on Unmanned Aerial Vehicle Remote Sensing

作　　者：朱红雷 黄艳伟[1] 李英臣 于飞[1,2] 涂田双[1] 王伟 张元培 李佳莉 罗懿哲 ZHU Honglei;HUANG Yanwei;LI Yingchen;YU Fei;TU Tianshuang;WANG Wei;ZHANG Yuanpei;LI Jiali;LUO Yizhe(College of Life Sciences,Henan Normal University,Xinxiang 453007,Henan,P.R.China;Research Center for Ecological Management and Protection of the Yellow Riuver Basin,Henan Normal University,Xinxiang 453007,Henan,P.R.China)

机构地区：[1]河南师范大学生命科学学院,河南新乡453007 [2]河南师范大学黄河流域生态治理与保护研究中心,河南新乡453007

出　　处：《湿地科学》2021年第1期17-26,共10页Wetland Science

基　　金：河南省重大科技专项项目(201300311700);河南省高等学校重点科研项目(17A170007);河南师范大学国家级项目培育基金项目(5101049170804)资助。

摘　　要：周期性的农业活动和水沙变化已经显著改变了黄河下游河滩地的植物群落结构,快速、准确地获取河滩地的植物群落多样性信息,可以为黄河流域生态保护和恢复提供参考依据。以位于河南省新乡市原阳县朱贵村南部的黄河下游河滩地的植物群落为研究对象,采用最大似然、人工神经网络、面向对象和随机森林分类方法,利用无人机多光谱遥感影像数据,对河滩地上的植物进行分类,计算出各种植物的相对盖度、相对频度、重要值和植物群落的Simpson多样性指数、Shannon-Wiener多样性指数。研究结果表明,利用无人机多光谱遥感影像数据,可以较为准确地获取黄河下游河滩地上无遮盖植物的物种信息,优选的遥感分类方法是人工神经网络分类方法,其分类结果的总体分类精度和Kappa系数分别为61.42%和0.52;其对河滩地上植物物种的分类结果与实地调查结果基本一致。无人机多光谱遥感方法是研究湿地中植物群落多样性的有效方法。Periodic agricultural activities and flow-sediment variation have significantly changed the plant community structure in the lower reaches of the Yellow River.The rapid and accurate acquisition information of plant community diversity can provide important reference and basis for ecological protection and management of the Yellow River Basin.In this paper,the typical flood plain area of Henan section of the lower reaches of the Yellow River in Yuanyang county,Henan province was taken as the research area.The multispectral images of the study area were obtained by remote sensing platform of the unmanned aerial vehicle(UAV).The classification accuracy of plant species in the flood land using 4 kinds of automatic methods was evaluated.Then,relative coverage,relative frequency,important value and the spatial distribution information of Simpson and Shannon-Wiener diversity indexes of the plant community in the study area were obtained.The results showed that multispectral remote sensing by UAV could accurately obtain the uncovered plant species distribution information in the lower reaches of the Yellow River.The optimal remote sensing classification and extraction method was neural network classification method,with the overall classification accuracy and Kappa coefficient of 61.42%and 0.52 respectively.The monitoring method of plant community diversity based on UAV remote sensing could obtain the spatial coverage of diversity indicators.The evalucated values of Simpson and Shannon-Wiener diversity indexes were basically consistent with those by the field survey.The investigation of plant community diversity based on UAV remote sensing is an effective way for investigation of the plants in the wetlands.

关 键 词：河滩地 植物群落多样性 无人机多光谱遥感 黄河下游河南段 

分 类 号：TP753[自动化与计算机技术—检测技术与自动化装置] Q948.15[自动化与计算机技术—控制科学与工程]