Towards general deep-learning-based tree instance segmentation models

TL;DR

This paper investigates the generalization of deep-learning models for tree segmentation across diverse forest point cloud datasets, highlighting the potential and challenges of domain transfer in different data resolutions and forest types.

Contribution

The study trains a single model on seven diverse datasets to assess its generalization capabilities across different forest types and data resolutions.

Findings

Generalization from coniferous to deciduous forests is feasible.

High-resolution to low-resolution generalization is challenging.

Diverse data is essential for robust model development.

Abstract

The segmentation of individual trees from forest point clouds is a crucial task for downstream analyses such as carbon sequestration estimation. Recently, deep-learning-based methods have been proposed which show the potential of learning to segment trees. Since these methods are trained in a supervised way, the question arises how general models can be obtained that are applicable across a wide range of settings. So far, training has been mainly conducted with data from one specific laser scanning type and for specific types of forests. In this work, we train one segmentation model under various conditions, using seven diverse datasets found in literature, to gain insights into the generalization capabilities under domain-shift. Our results suggest that a generalization from coniferous dominated sparse point clouds to deciduous dominated high-resolution point clouds is possible. Conversely, qualitative evidence suggests that generalization from high-resolution to low-resolution point clouds is challenging. This emphasizes the need for forest point clouds with diverse data characteristics for model development. To enrich the available data basis, labeled trees from two previous works were propagated to the complete forest point cloud and are made publicly available at https://doi.org/10.25625/QUTUWU.