This file was generated on 2026-03-05 by Jonas Bömer A GENERAL INFORMATION 1. Title of the dataset: Sugar4D 2. Brief description of the research project and its aims: The Sugar4D project introduces a publicly available, high-quality 4D plant phenotyping dataset of sugar beet plants captured with terrestrial LiDAR. It provides densely sampled, temporally consistent point cloud data with detailed, point-wise organ-level annotations that enable tracking of individual leaves across growth stages. By combining annotated 3D data with extracted morphological traits and reference measurements, the dataset aims to reduce the bottleneck of data acquisition and manual annotation, and to support the development, benchmarking, and validation of methods for spatio-temporal plant analysis, including instance segmentation, temporal registration, organ tracking, and growth-related morphological studies. 3. Author Information A. Investigator Contact Information Name: Jonas Bömer Institution: Institute of Sugar Beet Research Address: Holtenser Landstr.77, 37079 Göttingen Email: boemer@ifz-goettingen.de Name: Elias Marks Institution: Center for Robotics, University of Bonn Address: Nussallee 15, 53115 Bonn Email: elias.marks@uni-bonn.de Name: Facundo Ramón Ispizua Yamati Institution: Institute of Sugar Beet Research Address: Holtenser Landstr.77, 37079 Göttingen Email: ispizua@ifz-goettingen.de Name: Prof. Dr. Cyrill Stachniss Institution: Center for Robotics, University of Bonn Address: Nussallee 15, 53115 Bonn Email: cyrill.stachniss@igg.uni-bonn.de Name: Dr. Stefan Paulus Institution: Institute of Sugar Beet Research Address: Holtenser Landstr.77, 37079 Göttingen Email: paulus@ifz-goettingen.de Name: Prof. Dr. Anne-Katrin Mahlein Institution: Institute of Sugar Beet Research Address: Holtenser Landstr.77, 37079 Göttingen Email: mahlein@ifz-goettingen.de B. Project Supervisor (Principal Investigator) Contact Information Name: Prof. Dr. Anne-Katrin Mahlein Institution: Institute of Sugar Beet Research Address: Holtenser Landstr.77, 37079 Göttingen Email: mahlein@ifz-goettingen.de C. In case of questions related to this dataset, please contact: Name: Jonas Bömer Institution: Institute of Sugar Beet Research Address: Holtenser Landstr.77, 37079 Göttingen Email: boemer@ifz-goettingen.de 4. Date of data collection: 2021-10-26 - 2021-12-21 5. Information about funding sources that supported the collection of the data: - Federal Ministry of Food and Agriculture (BMEL): 28DK108C20 - German Research Foundation (DFG): EXC 2070 – 390732324 - Federal Ministry of Research, Technology and Space (BMFTR): Robotics Institute Germany (RIG) 6. Language of the dataset: English 7. Geographic location of data collection: Göttingen, Lower-Saxony, Germany B DATA & FILE OVERVIEW 1. File List: - sugar4d_readme.txt -> readme file including metadata - sugar4d.zip -> Sugar4D dataset - point_clouds/ –> main directory containing annotated plant point clouds and data splits - 37_das/ to 93_das/ –> 16 subdirectories named by days after sowing, each containing 48 .ply files (XX_das_plant_001.ply … XX_das_plant_048.ply) - splits/test.txt, splits/train.txt, splits/val.txt –> predefined ML data partitions - measurements/ –> main directory containing morphological parameters and reference measurements - plant_related.csv –> 58 automatically extracted plant-level morphological parameters (for 768 plant point clouds) - leaf_related.csv –> 5 automatically extracted leaf-level morphological parameters (for 6778 leaf point clouds) - reference/ -> subdirectory containing reference measurements - leaf_related_manual.csv –> manual reference measurements of 450 individual leaves - ref_plant/ -> subdirectory containing all files related to the reference plant - plant_related_ref_plant.csv –> reference measurements for plant-level parameters using a 3D-printed reference model - point_clouds/ -> subdirectory containing 15 reference plant scans (ref_plant_001.ply … ref_plant_015.ply) and digital base model (ref_plant_reference.ply) - supplements/ –> main directory containing additional supplementary information - experiment_layout.png –> greenhouse trial layout including genotype information - growth_points.npy –> manually annotated plant growth points - sensor_calibration.pdf –> LiDAR sensor calibration certificate - visualize.py –> Python visualization script - label_position/ –> subdirectory containing per-plant annotation position diagrams (label_position_plant_001.png … label_position_plant_048.png) - label_presence/ –> subdirectory containing per-plant temporal label presence diagrams (label_presence_plant_001.png … label_presence_plant_048.png) 2. Are there multiple versions of the dataset? no 3. Relationship between files: - 4. Additional related data collected that was not included in the current data package: - C SHARING/ACCESS INFORMATION 1. Was data derived from another source? no 2. Licenses/restrictions placed on the data: CC BY 4.0 3. Links to publications that cite or use the data : - Bömer, J., Marks, E., Ispizua Yamati, F.R. et al. Spatio-temporal 4D phenotyping for automated morphological genotype differentiation of sugar beet. Precision Agric 27, 50 (2026). https://doi.org/10.1007/s11119-026-10337-6 4. Related Datasets: - Pheno4D - A large scale spatio-temporal dataset of point clouds of maize and tomato plants (https://www.ipb.uni-bonn.de/data/pheno4d/index.html) - LAST-Straw - Lincoln’s Annotated Spatio-Temporal Strawberry Dataset (https://lcas.github.io/LAST-Straw/) D METHODOLOGICAL INFORMATION 1. Description of methods used for collection/generation of data: - Terrestrial LiDAR (Faro Focus S70, 1550 nm, ±1 mm ranging error) - Sensor inverted on camera arm/tripod, positioned above plants - 10 scanning positions in circular pattern per session - 9 spherical reference targets for scan registration - 16 time points, biweekly intervals, 37–93 days after sowing - 48 plants, 12 genotypes, 4 repetitions each 2. Methods for processing the data: - Scan processing: stray point, low-reflectivity, edge artifact filtering (Faro Scene) - Registration into unified point cloud via reference targets - Coordinate transformation using floor-mounted photogrammetric targets - Ground plane removal via RANSAC; outlier removal via statistical filter - Manual pointwise annotation (3-annotator pipeline) - Temporally consistent instance labels; annotation integrity verified via centroid displacement tracking - Morphological parameters extracted automatically (58 plant-level, 5 leaf-level) 3. Instrument- and/or software-specific information needed to interpret the data: - LiDAR: Faro Focus S70 (calibration certificate included in supplements) - Processing: Faro Scene (scan filtering & registration) - File format: binary .ply (vertices with x,y,z, nx,ny,z, label_semantic, label_instance) - Annotation platform: segments.ai - Visualization: Python script (visualize.py) included, requires Open3D ≥ 0.16.0 or open source software like CloudCompare - Morphological parameters: Python extraction code described in detail and published as supplementary material in https://doi.org/10.1007/s11119-026-10337-6 4. People involved in sample collection, processing, analysis and/or submission: - Jonas Bömer - Kiara Burgsmüller - Jonathan Eggers - Saskia Flentje - Dirk Koops - Anne-Katrin Mahlein - Elias Marks - Stefan Paulus - Cyrill Stachniss - Facundo Ramon Ispizua Yamati 5. Describe any quality-assurance procedures performed on the data: - Spatial point cloud integrity check by target- and scan point-based statistics - Annotation integrity by three-step annotation procedure and temporal label checks for each plant - Plant-related morphological parameters validated against 3D-printed reference model (n=15) - Leaf-related parameters validated against invasive manual reference measurements (n=450) 6. Standards and calibration information: - Sensor calibration certificate in supplements 7. Environmental/experimental conditions: - Controlled greenhouse, Institute of Sugar Beet Research, Göttingen, Germany - September–December 2021 - Temperature: 20 °C (20.1 ± 4.9 °C), humidity: 41.2 ± 12.5 % - Artificial lighting, 16 h photoperiod (sodium-vapor lamps) - 2 L flowerpots, sand/topsoil substrate (1:1 by weight) - Uniform pest management, fertilization, and irrigation protocol - Data acquisition in early morning (maximal leaf turgor)