摘要
崩岗是我国南方红壤区较严重的土壤侵蚀类型,也是威胁区域生态安全的重要因素。为研究GF-2号卫星影像提取崩岗分布的技术方法,以江西省赣县区白鹭乡金钩形小流域2017年的影像为数据源,对影像进行了光谱增强和纹理增强处理,采用面向对象的分类方法构建了崩岗信息提取的分类规则。精度验证结果表明,GF-2号卫星数据可满足崩岗调查信息提取的需求,提取精度可达91.7%;国产高分辨卫星影像在遥感监测崩岗中有较大的应用潜力。
Collapsed erosion is one of the severe land erosion types in the southern red soil region,as well as a major factor that threaten the regional ecological safety.In order to study the technical method of collapsed erosion distribution extraction based on GF-2 satellite images,taking the images of the collapsed erosion area in Ganxian District,Jiangxi Province as data source,we performed spectral enhancement and texture enhancement on the images,and used the object-oriented classification method to construct the classification rules for collapsed erosion information extraction.The accuracy verification results show that GF-2 satellite images can meet the information extraction demand of collapsed erosion investigation,and the extraction accuracy can reach 91.7%.Domestic high-resolution satellite image has great application potential in monitoring collapsed erosion by remote sensing.
出处
《地理空间信息》
2021年第11期85-87,107,I0007,共5页
Geospatial Information
基金
国家自然科学基金资助项目(41967011)
江西省自然科学基金资助项目(20181BAB203024)
江西省水利厅科技资助项目(KT201616、KT201724)
江西省重点研发计划资助项目(20171BBG70059)。
关键词
GF-2号
影像融合
崩岗提取
GF-2
image fusion
collapsed erosion extraction
