基于高光谱遥感技术的生态伪装监测应用研究  

Application Research of Ecological Camouflage Monitoring Based on Hyperspectral Remote Sensing Technology

作　　者：张清林 ZHANG Qinglin(Fujian Mapping Institute,Fuzhou 350001,China)

机构地区：[1]福建省制图院,福建福州350001

出　　处：《测绘与空间地理信息》2024年第11期52-54,58,共4页Geomatics & Spatial Information Technology

基　　金：2021年度福建省测绘地理信息发展中心科技创新项目(FJCH-2021XMK-07)资助。

摘　　要：对生态修复中存在挂防航拍伪装网的“虚假造绿”现象进行分析,采集当前市面上主要的迷彩伪装网光谱信息并建立伪装网光谱数据库。以高分五号高光谱影像数据为基础,绿色伪装网光谱数据为目标端元,采用LSU、CEM、ACE、OSP、MTMF、FCLS及SAM等方法进行混合像元分解与波谱识别检测实验。将检测定位结果与高空间分辨率影像进行叠加分析提取目标地物信息,改变以往基于高空间分辨率遥感影像,人工逐屏比对分析目视判读提取地物信息的方式,将大量的人工判读影像工作交给计算机去搜索完成,再对结果进行少量的人工干预,可极大地提高工作效率和准确性。This paper analyzes the phenomenon of "false greening" of laying anti-aerial photography camouflage nets in ecological restoration,gathers the spectral information of the main camouflage nets currently on the market,and establishes a spectral database of camouflage nets.Based on the hyperspectral image data of GF-5 and the spectral data of green camouflage net as the target endmember,the hybrid pixel decomposition and spectral identification detection experiments were carried out using methods including LSU,CEM,ACE,OSP,MTMF,FCLS and SAM,etc.The detection and positioning results with high spatial resolution images are superimposed and evaluated to extract the target feature information,which modified the existing methods based on high spatial resolution remote sensing images and manual screen-wise comparison and analysis of visual interpretation to extract feature information.Lots of workload of manual interpretations are handed over to the computer to complete the screening and then a small amount of manual intervention is performed on the results,which can greatly improve the work efficiency and accuracy.

关 键 词：生态修复 伪装网 虚假造绿 混合像元分解 目标检测 

分 类 号：P237[天文地球—摄影测量与遥感]