摘要
自然资源调查监测工作中存在遥感影像辐射不统一、自然资源样本繁杂难以更新、变化检测精度低等问题。为解决遥感影像变化检测精度低与可靠性低的难题,形成主动智能的自然资源调查监测技术,本文以自然资源典型地类变化监测需求为引导,构建多尺度高分遥感地面基准网,以及多尺度、多形态、多时相、多源异构自适应更新的自然资源典型地类样本库,研究耕地、水体、林地等自然资源典型地类的主动变化检测模型,分类精度与变化检测准确率分别达到82%与97.7%。该项技术提高了土地利用变化检测的精度与工作效率,减少了人工目视解译工作量,可为自然资源典型地类主动调查监测提供有力支撑。
In the investigation and monitoring of natural resources,there are problems such as inconsistent radiation of remote sensing images,samples being too complex to update,and low accuracy of change detection.In order to solve the problems of low accuracy and reliability of remote sensing change detection,an autonomous and intelligent technique for natural resources investigation and monitoring was proposed.Guided by the monitoring requirements of typical land classes of natural resources,a multi-scale high-resolution remote sensing ground reference network was constructed,and a multi-scale,multi-morphic,multi-temporal,multi-source heterogeneous adaptive update sample database of typical natural resources was formed.Furthermore,the change detection model of typical natural resources such as cultivated land,water,and forest was studied;the classification and change detection accuracy reached 82%and 97.7%,respectively.This technique can improve the accuracy and efficiency of classification and change detection for land cover,reduce the workload of manual interpretation,and provide strong support for active investigation and monitoring of typical natural resources.
作者
闫利
张毅
杨见兵
李希
王剑
YAN Li;ZHANG Yi;YANG Jianbing;LI Xi;WANG Jian(School of Geodesy and Geomatics,Wuhan University,Wuhan 430079,China)
出处
《地理信息世界》
2022年第5期66-73,共8页
Geomatics World
基金
国家重点研发计划项目(2020YFD1100203)。
关键词
自然资源调查监测
遥感地面基准网
变化检测
智能主动
natural resources investigation and monitoring
remote sensing ground reference network
change detection
active intelligence
