Semi-Automated Ground Elevation Correction from Waveform LiDAR Data
Large footprint waveform LiDAR (LFL) systems provide canopy and relative height metrics which are important to describe forest vertical structures.Yet the ability of canopy height metrics is limited by the accuracy of ground detection algorithm.Elevations from large footprint LiDAR (LFL) were compared with Digital Elevation Model (DEM) from discrete return LiDAR (DRL) and National Elevation Data (NED),referenced to the same horizontal and vertical datum (World Geodetic System,WGS84 Ellipsoid).Self-corrected ground elevations from LFL were validated by NED.Sensitivity of ground elevation errors to related factors,including topography (e.g.slope),vegetation (e.g.leaf area index and canopy coverage),and viewing geometry (e.g.scan angle) was analyzed.The proposed approach utilizes statistical information of ground elevation from neighboring footprints to correct the ground detection error.It was implemented to 2003 and 2009 LiDAR Vegetation and Imaging Sensor (LVIS) data and validated by NED and Airborne Laser Terrain Mappers (ALTM 3100,Optech) data collected at Howland Forest,Maine in northeast of United States.Results from this study will provide practical procedure for ground elevation corrections in LFL data.This will improve the accuracy of derived relative height metrics and thus the further applications in forestry such as estimates of forest aboveground biomass and habitat structure.
Laser Vegetation Imaging Sensor (LVIS) LiDAR Filtering Howland
Wenli Huang Guoqing Sun Wenjian Ni Zhiyu Zhang Dongliang Wang
University of Maryland, Department of Geographical Sciences, College Park, MD 20742, USA University of Maryland, Department of Geographical Sciences, College Park, MD 20742, USA;Chinese Aca Chinese Academy of Sciences, Institute of Remote Sensing Applications and Digital Earth, Beijing 100 Wuhan University, School of Remote Sensing and Information Engineering, Wuhan 430079, China
国际会议
北京
英文
226-238
2013-10-09（万方平台首次上网日期，不代表论文的发表时间）