A benchmark dataset for canopy crown detection and delineation in co-registered airborne RGB, LiDAR and hyperspectral imagery from the National Ecological Observation Network

Graves, Sarah J.; Marconi, Sergio; Singh, Aditya; Zare, Alina; Stewart, Dylan; Bohlman, Stephanie A.; White, Ethan P.

doi:10.1371/journal.pcbi.1009180

Cited by 27 publications

(20 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Further research could also explore the utility of this product to inform structural metrics of fire severity [48,84,85] and ecosystem vulnerability under a range of stressors [86]. With the recent publication of open-source, multi-sensor (e.g., LiDAR, high-resolution hyperspectral imagery) benchmark datasets that include numerous forest types and structures [87] there are even more opportunities to evaluate existing algorithms on a standard dataset.…”

Section: Discussionmentioning

confidence: 99%

Accuracy of a LiDAR-Based Individual Tree Detection and Attribute Measurement Algorithm Developed to Inform Forest Products Supply Chain and Resource Management

Sparks

Smith

2021

Forests

View full text Add to dashboard Cite

Individual Tree Detection (ITD) algorithms that use Airborne Laser Scanning (ALS) data can provide accurate tree locations and measurements of tree-level attributes that are required for stand-to-landscape scale forest inventory and supply chain management. While numerous ITD algorithms exist, few have been assessed for accuracy in stands with complex forest structure and composition, limiting their utility for operational application. In this study, we conduct a preliminary assessment of the ability of the ForestView® algorithm created by Northwest Management Incorporated to detect individual trees, classify tree species, live/dead status, canopy position, and estimate height and diameter at breast height (DBH) in a mixed coniferous forest with an average tree density of 543 (s.d. ±387) trees/hectare. ITD accuracy was high in stands with lower canopy cover (recall: 0.67, precision: 0.8) and lower in stands with higher canopy cover (recall: 0.36, precision: 0.67), mainly owing to omission of suppressed trees that were not detected under the dominant tree canopy. Tree species that were well-represented within the study area had high classification accuracies (producer’s/user’s accuracies > ~60%). The similarity between the ALS estimated and observed tree attributes was high, with no statistical difference in the ALS estimated height and DBH distributions and the field observed height and DBH distributions. RMSEs for tree-level height and DBH were 0.69 m and 7.2 cm, respectively. Overall, this algorithm appears comparable to other ITD and measurement algorithms, but quantitative analyses using benchmark datasets in other forest types and cross-comparisons with other ITD algorithms are needed.

show abstract

Section: Discussionmentioning

confidence: 99%

Accuracy of a LiDAR-Based Individual Tree Detection and Attribute Measurement Algorithm Developed to Inform Forest Products Supply Chain and Resource Management

Sparks

Smith

2021

Forests

View full text Add to dashboard Cite

show abstract

“…In terms of benchmarks, there had been timely limited comparisons of different methods on shared datasets in the past [20,22,23]. However, only recently, efforts were made for a curated, open, and long-term benchmark for individual tree detection in forests [107].…”

Section: Ground Truth Datamentioning

confidence: 99%

Individual Tree Detection in Urban ALS Point Clouds with 3D Convolutional Networks

2022

View full text Add to dashboard Cite

Since trees are a vital part of urban green infrastructure, automatic mapping of individual urban trees is becoming increasingly important for city management and planning. Although deep-learning-based object detection networks are the state-of-the-art in computer vision, their adaptation to individual tree detection in urban areas has scarcely been studied. Some existing works have employed 2D object detection networks for this purpose. However, these have used three-dimensional information only in the form of projected feature maps. In contrast, we exploited the full 3D potential of airborne laser scanning (ALS) point clouds by using a 3D neural network for individual tree detection. Specifically, a sparse convolutional network was used for 3D feature extraction, feeding both semantic segmentation and circular object detection outputs, which were combined for further increased accuracy. We demonstrate the capability of our approach on an urban topographic ALS point cloud with 10,864 hand-labeled ground truth trees. Our method achieved an average precision of 83% regarding the common 0.5 intersection over union criterion. 85% percent of the stems were found correctly with a precision of 88%, while tree area was covered by the individual tree detections with an F1 accuracy of 92%. Thereby, we outperformed traditional delineation baselines and recent detection networks.

show abstract

“…Many free, open‐source software tools exist for working with geospatial data products produced by UAS and NEON including QGIS (https://qgis.org/en/site/) for visualization and GUI‐based manipulation of raster and vector data types, CloudCompare (https://www.danielgm.net/cc/) for visualization and GUI‐based manipulation of point clouds, and a suite of packages (https://cran.r-project.org/web/views/Spatial.html) for the R programming language (R Core Team, 2021). Several packages have also been developed specifically for working with NEON data, including neonUtilities (Lunch et al, 2021), neonhs (Joseph & Wasser, 2021), geoNEON (National Ecological Observatory Network, 2020), and NeonTreeEvaluation (Weinstein et al, 2021). A recent review by Atkins et al (2022) describes the ecosystem of R packages available for working with forestry data, many of which are relevant for the types of geospatial data produced by UAS and NEON.…”

Section: Core Principles For Uas/neon Integrationmentioning

confidence: 99%

Democratizing macroecology: Integrating unoccupied aerial systems with the National Ecological Observatory Network

et al. 2022

View full text Add to dashboard Cite

Macroecology research seeks to understand ecological phenomena with causes and consequences that accumulate, interact, and emerge across scales spanning several orders of magnitude. Broad-extent, fine-grain information (i.e., high spatial resolution data over large areas) is needed to adequately capture these cross-scale phenomena, but these data have historically been costly to acquire and process.Unoccupied aerial systems (UAS or drones carrying a sensor payload) and the National Ecological Observatory Network (NEON) make the broad-extent, fine-grain observational domain more accessible to researchers by lowering costs and reducing the need for highly specialized equipment. Integration of these tools can further democratize macroecological research, as their strengths and weaknesses are complementary. However, using these tools for macroecology can be challenging because mental models are lacking, thus requiring large up-front investments in time, energy, and creativity to become proficient. This challenge inspired a working group of UAS-using academic ecologists, NEON professionals, imaging scientists, remote sensing specialists, and aeronautical engineers at the 2019 NEON Science Summit in Boulder, Colorado, to synthesize current knowledge on how to use UAS with NEON in a mental model for an intended audience of ecologists new to these tools. Specifically, we provide (1) a collection of core principles for collecting high-quality UAS data for NEON integration and (2) a case study illustrating a sample workflow for processing UAS data into meaningful ecological information and integrating it with NEON data collected on the groundwith the Terrestrial Observation System-and remotely-from the Airborne Observation Platform. With this mental model, we advance the democratization of macroecology by making a key observational domain-the broad-extent, fine-grain domain-more accessible via NEON/UAS integration.

show abstract

A benchmark dataset for canopy crown detection and delineation in co-registered airborne RGB, LiDAR and hyperspectral imagery from the National Ecological Observation Network

Cited by 27 publications

References 34 publications

Accuracy of a LiDAR-Based Individual Tree Detection and Attribute Measurement Algorithm Developed to Inform Forest Products Supply Chain and Resource Management

Accuracy of a LiDAR-Based Individual Tree Detection and Attribute Measurement Algorithm Developed to Inform Forest Products Supply Chain and Resource Management

Individual Tree Detection in Urban ALS Point Clouds with 3D Convolutional Networks

Democratizing macroecology: Integrating unoccupied aerial systems with the National Ecological Observatory Network

Contact Info

Product

Resources

About