Data for: Manual Labelling Artificially Inflates Deep Learning-Based Segmentation Performance on RGB Images of Closed Canopy: Validation Using Terrestrial Laser Scanning

2025/03/18 by Matthew J. Allen, Harry J. F. Owen, Harry Owen +7 · 1 voice · 1 citation
Environmental Science · Engineering · #Remote Sensing and LiDAR Applications #Advanced Measurement and Detection Methods #Infrared Target Detection Methodologies

paper · pdf · doi:10.48550/arxiv.2503.14273

Abstract

All users should cite this paper, for which this dataset was created: Allen MJ, Owen HJF, Grieve SWD and Lines ER, (2025) Manual Labelling Artificially Inflates Deep Learning-Based Segmentation Performance on RGB Images of Closed Canopy: Validation Using Terrestrial Laser Scanning. arXiv. https://doi.org/10.48550/arXiv.2503.14273

Citations

• Responses of tree species to heat waves and extreme heat events

Cited by

• Benchmarking individual tree segmentation using multispectral airborne laser scanning data: the FGI-EMIT dataset

Discussions

• There's a lot of interest in #deeplearning for tree crown segmentation from imagery, but most methods are tested in manual annotations, checked visually. How well do these perform on rigourously gro [bsky, 24 points, 3 comments]

Related

• Low-Cost Tree Crown Dieback Estimation Using Deep Learning-Based Segmentation
• Bringing SAM to new heights: Leveraging elevation data for tree crown segmentation from drone imagery
• Benchmarking tree species classification from proximally-sensed laser scanning data: introducing the FOR-species20K dataset
• PointsToWood: A deep learning framework for complete canopy leaf-wood segmentation of TLS data across diverse European forests
• TreeSatAI Benchmark Archive: a multi-sensor, multi-label dataset for tree species classification in remote sensing