Individual tree point clouds and tree measurements from multi-platform laser scanning in German forests

Weiser, Hannah; Schäfer, Jannika; Winiwarter, Lukas; Krašovec, Nina; Fassnacht, Fabian Ewald; Höfle, Bernhard

doi:10.5194/essd-2022-39

Cited by 3 publications

(6 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Finally, Table 8 represents the quantitative assessment of the FVLS-DL, VLS-DL, and real (geometric only) models on different datasets. In addition to the point clouds created from Weiser et al (2023) (results reported in Tables 4 and 7), the classification performance of the models (Real, VLS-DL and FVLS-DL) was assessed with point clouds from Wang et al (2021), for both the isolated and the near trees case, and for point clouds used in Xi et al…”

Section: Leaf-wood Resultsmentioning

confidence: 99%

“…For the leaf-wood case, synthetic meshes generated with the algorithm by Weber and Penn (1995), and the Wytham Woods 3D model (Calders et al, 2018;Liu et al, 2022) are used. Real labeled point clouds (Hessigheim for the urban case, point clouds from Weiser et al (2023) for the leaf-wood case) serve as training data for the deep learning (DL) models. The performance of the models trained on real data and those trained on virtual data are evaluated and compared quantitatively and qualitatively.…”

Section: Methodsmentioning

confidence: 99%

“…The main validation datasets for this study were generated from the remaining trees of the labeled tree point cloud dataset (Weiser et al (2023)), respectively. For the isolated trees, these were PicAbi BR08 01, PinSyl KA09 T048, QueRub KA11 09, for the near trees with ground, these were FagSyl BR05 P8T4, PicAbi BR08 01, QueRub KA09 T053 (Figure B.13b).…”

Section: Virtual Laser Scanning Configurationmentioning

confidence: 99%

See 2 more Smart Citations

Deep learning with simulated laser scanning data for 3D point cloud classification

Esmorís Pena,

Weiser,

Winiwarter

et al. 2024

Preprint

View full text Add to dashboard Cite

Laser scanning is an active remote sensing technique applied in many disciplines to acquire state-of-the-art spatial measurements. Semantic labeling is often necessary to extract information from the raw point cloud. Deep learning methods constitute a data-hungry solution for the semantic segmentation of point clouds. In this work, we investigate the use of simulated laser scanning for training deep learning models, which are applied to real data subsequently. We show that training a deep learning model purely on virtual laser scanning data can produce results comparable to models trained on real data when evaluated on real data. For leaf-wood segmentation of trees, using the KPConv model trained with virtual data achieves 93.7% overall accuracy, while the model trained on real data reaches 94.7% overall accuracy. In urban contexts, a KPConv model trained on virtual data achieves 74.1% overall accuracy on real validation data, while the model trained on real data achieves 82.4%. Our models outperform the state-of-the-art model FSCT in terms of generalization to unseen real data as well as a baseline model trained on points randomly sampled from the tree mesh surface. From our results, we conclude that the combination of laser scanning simulation and deep learning is a cost-effective alternative to real data acquisition and manual labeling in the domain of geospatial point cloud analysis. The strengths of this approach are that a) a large amount of diverse laser scanning training data can be generated quickly and without the need for expensive equipment, b) the simulation configurations can be adapted so that the virtual training data have similar characteristics to the targeted real data, and c) the whole workflow can be automated through procedural scene generation.

show abstract

Section: Leaf-wood Resultsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Virtual Laser Scanning Configurationmentioning

confidence: 99%

See 1 more Smart Citation

Deep learning with simulated laser scanning data for 3D point cloud classification

Esmorís Pena,

Weiser,

Winiwarter

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…VALIDATION BY LIDAR DATA Finally, we study the uncertainty of trained models on stand level. We utilize a dataset that get assembled from forests in Baden-Württemberg, Germany in the years 2019 and 2020 [27]. It embraces 12 separate plots, each covering a spatial area of about one hectare.…”

Section: Uncertainty Evaluationmentioning

confidence: 99%

“…The analysis of high-resolution aerial imagery in [26] revealed the accuracy of estimated crown diameter significantly varied with plots (0.63 and 0.85) when compared with height estimations. According to [27], the relative error of tree crown diameter derived from airborne/UAV-borne LiDAR data is significantly larger than that of derived tree height (19.22% and 20.7% for crown diameter, 11.70% and 10.97% for tree height estimation respectively). In fact, the crowns of individual trees in e.g.…”

mentioning

confidence: 99%

Biomass Estimation and Uncertainty Quantification From Tree Height

Song

Albrecht

Xiong

et al. 2023

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

We propose a tree-level biomass estimation model approximating allometric equations by LiDAR data. Since tree crown diameters estimation is challenging from spaceborne Li-DAR measurements, we develop a model to correlate tree height with biomass on the individual tree level employing a Gaussian process regressor. In order to validate the proposed model, a set of 8,342 samples on tree height, trunk diameter, and biomass has been assembled. It covers seven biomes globally present. We reference our model to four other models based on both, the Jucker data and our own dataset. Although our approach deviates from standard biomass-height-diameter models, we demonstrate the Gaussian process regression model as a viable alternative. In addition, we decompose the uncertainty of tree biomass estimates into the model-and fitting-based contributions. We verify the Gaussian process regressor has the capacity to reduce the fitting uncertainty down to below 5%. Exploiting airborne LiDAR measurements and a field inventory survey on the ground, a stand-level (or plot-level) study confirms a low relative error of below 1% for our model. The data used in this study are available at https://github.com/zhu-xlab/BiomassUQ.

show abstract

An Automatic Individual Tree 3D Change Detection Method for Allometric Parameters Estimation in Mixed Uneven-Aged Forest Stands from ALS Data

et al. 2022

View full text Add to dashboard Cite

Forests play a central role in the management of the Earth’s climate. Airborne laser scanning (ALS) technologies facilitate the monitoring of large and impassable areas and can be used to monitor the 3D structure of forests. While the ALS-based forest measures have been studied in depth, 3D change detection in forests is still a subject of little attention in the literature due to the challenges introduced by comparing point cloud pairs. In this study, we propose an innovative methodology to (i) automatically perform a 3D change detection of forests on an individual tree level; (ii) estimate tree parameters with allometric equations; and (iii) perform an assessment of the aboveground biomass (AGB) variation over time. The area in which the tests were carried out was hit by an ice storm that occurred in the time interval between the two LiDAR acquisitions; furthermore, field measurements were carried out and used to validate the results. The single-tree segmentation of the point clouds was automatically performed with a local maxima algorithm to detect the treetop, and a decision tree method to define the individual crowns around the local maxima. The multitemporal comparison of the point clouds was based on the identification of single trees, which were matched when there was a correlation between the position of the treetops. For each tree, the DBH (diameter at breast height) and the AGB were also estimated using allometric equations. The results are promising and allowed us to identify the uprooted trees and estimate that about 40% of the AGB of the area under examination had been destroyed, with an RMSE over the estimation ranging between 4% and 21% in four scenarios.

show abstract

Individual tree point clouds and tree measurements from multi-platform laser scanning in German forests

Cited by 3 publications

References 23 publications

Deep learning with simulated laser scanning data for 3D point cloud classification

Deep learning with simulated laser scanning data for 3D point cloud classification

Biomass Estimation and Uncertainty Quantification From Tree Height

An Automatic Individual Tree 3D Change Detection Method for Allometric Parameters Estimation in Mixed Uneven-Aged Forest Stands from ALS Data

Contact Info

Product

Resources

About