This file was generated on 2025-07-15 by Chong, Yue Linn. A GENERAL INFORMATION 1. Title of the dataset: MuST-C Dataset: The Multi-Sensor and Multi-Temporal Data Set of Multiple Crops for In-Field Phenotyping and Monitoring 2. Brief description of the research project and its aims: Phenotyping is crucial for understanding crop trait variation and advancing research, but is currently limited by expensive, labor-intensive monitoring. New phenotypic trait monitoring methods are being proposed to reduce this so-called phenotyping bottleneck via automation. These methods are often data-driven, requiring a dataset recorded with a specific sensor and corresponding reference values for developing novel methods. To this end, we present the MuST-C (Multi-Sensor, multi-Temporal, multiple Crops) dataset, which contains field data from various sensors collected over a growing season, covering six crop species. All data was georeferenced for alignment across sensors and dates. To collect our dataset, we deployed aerial and ground robotic platforms equipped with RGB cameras, LiDARs, and multispectral cameras, aiming to capture a wide variety of modalities and observations from different viewpoints. In addition to sensor data, we also provide manually collected leaf area index and biomass reference measurements. Our dataset enables the development of novel automatic phenotypic trait estimation methods, allows comparisons across different sensors, and generalizability across crop species. 3. Author Information A. Investigator Contact Information Name: Chong, Yue Linn Institution: University of Bonn Address: Nussallee 15, 53115 Bonn, Germany. Email: linn.chong@uni-bonn.de B. Project Supervisor (Principal Investigator) Contact Information Name: Lasse Klingbeil Institution: University of Bonn Address: Nussallee 17, 53115 Bonn, Germany. Email: klingbeil@igg.uni-bonn.de C. In case of questions related to this dataset, please contact: Name: Lasse Klingbeil Institution: University of Bonn Address: Nussallee 17, 53115 Bonn, Germany. Email: klingbeil@igg.uni-bonn.de 4. Date of data collection: 17-05-2023 to 08-09-2023 5. Information about funding sources that supported the collection of the data: This work has been partially funded by the German Research Foundation under Germany’s Excellence Strategy, EXC-2070 - 390732324 – PhenoRob. 6. Language of the dataset: English 7. Geographic location of data collection : Campus Klein-Altendorf research facility, University of Bonn, Germany (50° 37' North, 6° 59' East) B DATA & FILE OVERVIEW 1. File List: See folder_structure_v2.pdf for the overview of the directory structure. To download a specific file, look for the zip file with the path in snake case. For example, the images of UAV2-RGB can be found in the zip file images_UAV2-RGB_YYMMDD.zip, where YYMMDD represents the date of data collection. The data is first sorted based on its modality. Specifically, all raw images are located in the images directory, all point clouds are stored in the point_clouds directory, and all raster data (orthophotos) are placed in the raster_data directory. We place all reference measurements, field information, and further trial metadata in the LAI_biomass_and_metadata directory. We further organise each modality based on the sensor and platform used to collect the data. We collected data using multiple robotic platforms equipped with various sensors. The first part of each data package ID is the platform name (UAV1/UAV2/UAV3/UGV), and the second part is the sensor type, where RGB stands for RGB camera data, MS stands for multispectral data, Lidar, LMI, and Ouster stands for lidar data. 2. Are there multiple versions of the dataset? no 3. Relationship between files: Files in the plot-wise directory are duplicated and post-processed specifically for users who are interested in plot-level data. We organised the RGB images from UGV-RGB, and point clouds from UGV-LMI, and point clouds from UAV1-RGB into separate individual zip files, due to size concerns. All other data packages were compiled into a single zip per plot. Please download the LAI and metadata file separately, i.e., these reference data are not duplicated in the plot-wise directory. 4. Additional related data collected that was not included in the current data package: N/A 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: Y. L. Chong, J. Krämer, E. Chakhvashvili, E. Marks, F. Esser, A. Dreier, R. A. Rosu, K. Warstat, R. Pude, S. Behnke, O. Muller, U. Rascher, H. Kuhlmann, C. Stachniss, J. Behley, L. Klingbeil. The Multi-Sensor and Multi-Temporal Dataset of Multiple Crops for In-Field Phenotyping and Monitoring, (under review) (2025). 4. Links to other publicly accessible locations of the data: https://www.ipb.uni-bonn.de/data/MuST-C/ 5. Links/relationships to ancillary datasets: N/A D METHODOLOGICAL INFORMATION 1. Description of methods used for collection/generation of data: We collected the data with three separate UAVs (UAV1/UAV2/UAV3) and a ground vehicle (UGV). The data was collected at a field trial consisting of multiple crops. We also collected manual data for LAI and biomass measurements. Most sensor modalities are self-explanatory, except the UGV Lidar data. We mounted two types of Lidar on our UGV: LMI Laser Triangulation scanners and an Ouster multi-beam Lidar. As such, data from the LMI is stored in UGV-LMI, and data from the Ouster is stored in UGV-Ouster. 2. Methods for processing the data: We used the Agisoft Metashape Pro software package to perform structure-from-motion for all images. The raw point clouds from individual lidar scans were merged into a single point cloud for each sensor at each data collection date. The LAI reference measurements were taken non-destructively with a SunScan Plant canopy analyzer and repeated destructively with LAI measurements from the WinDIAS Leaf Image Analysis System. The above-ground fresh biomass was weighed with destructive measurements. 3. Instrument- and/or software-specific information needed to interpret the data: All data formats used in this dataset are open, and we provide a developer's kit for users to process our data: https://github.com/PRBonn/MuST-C easily. The full list of the required libraries is listed on our GitHub repository: https://github.com/PRBonn/MuST-C/blob/main/requirements.txt. Alternatively, for users who prefer GUI tools, we recommend QGIS for our raster data and shapefiles (.shp and corresponding .cpg, .dbf, .prj, .qmd, .shx files) and CloudCompare for point cloud files. 4. People involved in sample collection, processing, analysis and/or submission: Yue Linn Chong, Julie Krämer, Erekle Chakhvashvili, Elias Marks, Felix Esser, Ansgar Dreier, Radu Alexandru Rosu, Kevin Warstat, Ralf Pude, Sven Behnke, Onno Muller, Uwe Rascher, Heiner Kuhlmann, Cyrill Stachniss, Jens Behley, and Lasse Klingbeil. 5. Describe any quality-assurance procedures performed on the data: N/A 6. Standards and calibration information: Calibration files from structure-from-motion methods used to obtain the orthophotos are provided with the images. 7. Environmental/experimental conditions: N/A