Automatic Detection of Single Trees in Airborne Laser Scanning Data through Gradient Orientation Clustering

Abstract

Currently, existing methods for single-tree detection based on airborne laser scanning (ALS) data usually require some thresholds and parameters to be set manually. Manually setting threshold or parameters is laborious and time-consuming, and for dense forests, the high commission and omission rate make most existing single-tree detection techniques inefficient. As a solution to these problems, this paper proposed an automatic single-tree detection method in ALS data through gradient orientation clustering (GOC). In this method, the rasterized Canopy Height Model (CHM) was derived from ALS data using surface interpolation. Then, potential trees were assumed as approximate conical shapes and extracted based on the GOC. Finally, trees were identified from the potential trees based on the compactness of the crown shape. This method used the gradient orientation information of rasterized CHM, thus increasing the generalization of single-tree detection method. In order to verify the validity and practicability of the proposed method, twelve 1256 m 2 circular study plots of different forest types were selected from the benchmark dataset (NEWFOR), and the results from nine different methods were presented and compared for these study plots. Among nine methods, the proposed method had the highest root mean square matching score (RMS_M = 43). Moreover, the proposed method had excellent detection (M > 47) in both single-layer coniferous and single-layered mixed stands.

An Individual Tree Segmentation Method Based on Watershed Algorithm and Three-Dimensional Spatial Distribution Analysis From Airborne LiDAR Point Clouds

“…To handle the symmetry of the citrus tree, Ok and Ozdarici-Ok [22] presented an original approach to detect and delineate citrus trees using unmanned aerial vehicle (UAV) based photogrammetric DSMs, where an orientation-based radial symmetry transform was performed and then the individual citrus trees were delineated using active contours. Similarly, Dong et al [23] also developed an automated single tree detection framework based on gradient orientation clustering from rasterized airborne LiDAR point clouds. Pirotti [24] adopted a template matching method to extract a correlation map from LiDAR-derived digital CHM for obtaining stem density, position, and height values.…”

LiDAR Applications to Estimate Forest Biomass at Individual Tree Scale: Opportunities, Challenges and Future Perspectives

“…In early studies, the accuracy for ALS-based individual tree detection and crown segmentation ranged from 29.7 to 48.3% [72]; the development of the Canopy Height Model (CHM) significantly improved the accuracy for tree detection at the single tree level [79]. CHM is associated with the canopy maxima algorithm to detect tree tops from the rasterized canopy heights from ALS point clouds [80]. High commission or omission rates might bring basis into individual tree detection outputs from CHM due to forest structure complexity and different tree species [81].…”

“…This significantly reduced commission errors produced by local maxima filtering and increased overall accuracy by 10%. To improve its accuracy, manually set thresholds and parameters were normally required [80]. To improve the processing accuracy and provide comparable results among locations, automatic individual tree detection was attempted.…”

“…Heurich [83] delineated individual tree crowns using automatic recognition for ALS-derived tree height and crown diameter that achieved an overall accuracy of 79.6% for individual tree detection from the upper canopy. Dong et al [80] developed an automatic single tree detection method using gradient orientation clustering (GOC) based on the rasterized CHM from ALS point clouds, significantly improving accuracy. Other efforts were made to detect individual trees using ALS fusion and optical imagery [73], alternative methods (e.g., nearest-neighbor estimation [84]), ALS-derived shape metrics and structure parameters as inputs for classification algorithms (e.g., k-Most Similar Neighbor (K-MSN) and the Random Forest (RF) [75]).…”

“…However, height estimation still varies in complex canopy structures and topography conditions due to low vertical point density [73]. More importantly, most of the current techniques for single tree detection are inefficient for dense forest ecosystems due to high commission and omission rates [80]. Therefore, researchers take advantage of both ALS (i.e., extracting tree heights, number of trees and tree crown width) and TLS (i.e., DBH, stand volume and tree height) for AGB estimation [65,79,86].…”