摘要
本研究基于获取时间相近的“珠海一号”高光谱数据(OHS)、高分一号(GF-1)多光谱数据和资源一号02D(ZY-102D)多光谱数据,采用支持向量机、随机森林和神经网络方法提取广西扶绥县桉树空间分布范围,评估其分类精度。结果表明:1)使用3种遥感数据进行分类均能取得较好的效果。2)3种分类方法中神经网络的分类效果最好。3)采用3种数据源和3种分类方法获得的扶绥县桉树空间分布格局一致,扶绥县桉树主要分布在北部、南部和东部的山区。本研究为扶绥县桉树经营管理提供科学的数据支持,对利用遥感技术快速准确获取桉树种植分布数据具有一定的指导作用。
Based on the data of OHS,GF-1 and ZY-102D obtained at similar times,this study used support vector machine,random forest and neural network methods to extract the spatial distribution range of eucalyptus in Fusui county,Guangxi,and evaluated its classification accuracy.The results showed that:1)Using three types of remote sensing data for classification achieved good results.2)Among the three classification methods,neural networks had the best classification performance.3)The same spatial distribution pattern of eucalyptus trees in Fusui county was obtained using three data sources and three classification methods.Eucalyptus trees in Fusui county were mainly distributed in the northern,southern,and eastern mountainous areas.This study provides scientific data support for the management and operation of eucalyptus trees in Fusui county,and has a certain guiding role in quickly and accurately obtaining eucalyptus planting distribution data using remote sensing technology.
作者
黄友菊
韩广萍
崔云蕾
HUANG Youju;HAN Guangping;CUI Yunlei(Natural Resources Remote Sensing Institute of Guangxi Zhuang Autonomous Region,Nanning 530023,China)
出处
《测绘与空间地理信息》
2023年第9期20-23,共4页
Geomatics & Spatial Information Technology
基金
广西重点研发计划项目(桂科AB22080080)
高分辨率对地观测系统重大专项政府综合治理应用于规模化产业化示范项目(84-Y50G25-9001-22/23)资助。
作者简介
黄友菊(1980-),女,广西横县人,高级工程师,硕士,2008年毕业于云南师范大学地图学与地理信息系统专业,主要从事遥感技术研发、应用及测绘地理信息系统研发等方面的工作。
