A Comparison of LiDAR Filtering Algorithms in Vegetated Mountain Areas. Issue 4 (4th July 2018)

• Record Type: Journal Article
• Authors: Zhao, Xiaoqian; Su, Yanjun; Li, WenKai; Hu, Tianyu; Liu, Jin; Guo, Qinghua
• Is Part Of: Canadian journal of remote sensing. Volume 44:Issue 4(2018)
• Page Start: 287
• Page End: 298
• Publication Date: 2018-07-04
• Subjects: Remote sensing -- Periodicals; 621.367805
• DOI: 10.1080/07038992.2018.1481738
• Languages: English
• ISSNs: 0703-8992

Abstract

Abstract: Filtering of airborne light detection and ranging (LiDAR) data is a challenging task in vegetated mountain areas. Environmental features and LiDAR data characteristics have significant impacts on the performance of filtering algorithms. This study aims to determine the effects of topographic and environmental features such as slope, canopy cover, elevation variability, and LiDAR point density on five widely used filtering algorithms, including multi-scale curvature classification (MCC), interpolation-based filtering (IBF) algorithm, morphological filtering (MF) algorithm, progressive triangulated irregular network densification filtering (PTDF) algorithm, and slope-based filtering (SBF). The results show that the performances of these filtering algorithms are all significantly influenced by the chosen factors, but the dominant influential factor varies with algorithms. The MCC works well in steep and dense forests; IBF and MCC outperform the rest of filtering algorithms in areas with steep terrain but low vegetation coverage; and PTDF is more reliable for low-density LiDAR data. Our results can provide guidance for choosing the appropriate filtering algorithm based on the specific topographic and environmental features of a study area.