Adaptive Framework for the Delineation of Homogeneous Forest Areas Based on LiDAR Points

Moritz Bruggisser,Norbert Pfeifer,Markus Hollaus,Di Wang

doi:10.3390/rs11020189

Abstract

We propose a flexible framework for automated forest patch delineations that exploits a set of canopy structure features computed from airborne laser scanning (ALS) point clouds. The approach is based on an iterative subdivision of the point cloud using k-means clustering followed by an iterative merging step to tackle oversegmentation. The framework can be adapted for different applications by selecting relevant input features that best measure the intended homogeneity. In our study, the performance of the segmentation framework was tested for the delineation of forest patches with a homogeneous canopy height structure on the one hand and with similar water cycle conditions on the other. For the latter delineation, canopy components that impact interception and evapotranspiration were used, and the delineation was mainly driven by leaf area, tree functional type, and foliage density. The framework was further tested on two scenes covering a variety of forest conditions and topographies. We demonstrate that the delineated patches capture well the spatial distributions of relevant canopy features that are used for defining the homogeneity. The consistencies range from R 2 = 0 . 84 to R 2 = 0 . 86 and from R 2 = 0 . 80 to R 2 = 0 . 91 for the most relevant features in the delineation of patches with similar height structure and water cycle conditions, respectively.

Highlights

Forests provide a variety of ecosystem goods and services, ranging from timber production to climate mitigation [1,2]
Contrary to urban applications for which the optimum neighborhood can be derived from the point cloud [43], the scale in forestry applications has to match the extent of the objects, i.e., the trees
Starting with the unsegmented point cloud, the maps in the top row depict the spatial distribution of the two resulting segments after each bipartitioning step for iterations 1–3

Summary

Introduction

Forests provide a variety of ecosystem goods and services, ranging from timber production to climate mitigation [1,2]. For the computation of features describing the local forest structure around a point, all points lying within a certain point’s neighborhood are used. Segmentation of forest stands requires the generalization of the fine-scale variabilities that are introduced by, for example, small gaps between trees. Chose a search radius of 10 m around each point as the maximum neighborhood size, which exceeds the scale of individual tree crowns and lowers the impact of small-scale gaps. Along with this large neighborhood, we computed the features for search radii of 2 and 5 m in order to retain a certain amount of variance at smaller scales. The computation of the described neighborhood search within the point cloud was accomplished using the kd-tree search implemented in scipy.spatial from the SciPy library [45]

Objectives

Methods

Results

Discussion

Conclusion