摘要
针对机载激光雷达与航空光学影像的互补特性,提出了一种基于多源遥感数据的高精度地物信息提取和分类方法。首先从激光雷达的全波形数据获得数字高程模型(DEM)、地物的正规化数字表面模型(nDSM)和激光雷达回波相对强度信息,从航空数码相机影像获得植被指数信息;然后利用决策树方法进行地物识别。选取"黑河综合遥感联合试验"中的3种典型区域(城市、农田和水体)进行分类,结果表明:该方法能够有效地分离建筑物、高大植被、低矮植被、裸土地以及水泥地等基本地物。
According to the complementarities of airborne laser scanning(ALS) data and aerial images,an accurate method of classification based on multi-source remote sensing data is presented.Firstly,DEM,nDSM,the relative intensity of return laser,and vegetation index can be extracted from the ALS data and aerial images,respectively.And then decision tree is adopt to recognize various ground objects.Finally,three typical areas of WATER(Watershed Airborne Telemetry Experimental Research) including city,cropland and water bodies are used to validate this approach.The result shows that the method can divide experimental area into building,high vegetation,low vegetation,cement and bare soil efficiently and reliably.
出处
《遥感技术与应用》
CSCD
北大核心
2010年第6期821-827,共7页
Remote Sensing Technology and Application
基金
中国科学院西部行动计划项目(KZCX2-XB2-09)
国家自然科学基金重点项目(40730525)
国家重点基础研究发展规划项目(2007CB714400)
关键词
全波形机载激光雷达
航空影像
决策树
地物分类
Full-waveform Airborne Laser Scanning
Aerial images
Decision tree
Classification