🌲 TreeScope: An Agricultural Robotics Dataset for LiDAR-Based Mapping of Trees in Forests and Orchards

Overview

TreeScope is a robotics dataset for precision agriculture and forestry addressing the counting and mapping of trees in forestry and orchards. TreeScope provides LiDAR data from agricultural environments collected with robotics platforms, such as UAV and mobile robot platforms carried by vehicles and human operators. We provide ground-truth data for semantic segmentation and diameter estimation with over 1,800 manually annotated semantic labels for tree stems and field-measured tree diameters. We share benchmark scripts for these tasks that researchers may use to evaluate the accuracy of their algorithms.

TreeScope processed data, raw data, and code are available to download. For more information about our dataset, please visit https://treescope.org or watch our video.

For detailed instructions on how to use this repository, please refer to this step-by-step tutorial

Converting Labels

For converting H5 labels into 2D range images:

python3 semantic_segmentation/h5-to-labels.py --file <h5-file> --output <labels>

For converting labeled 2D range images into H5 labels:

python3 semantic_segmentation/full_data_preprocessor.py -D <path-to-data/>
python3 semantic_segmentation/labels-to-h5.py -D <path-to-data/>

Diameter Estimation Benchmarks

For calculating root-mean-square error of diameter estimation results compared to ground-truth:

python3 diameter_estimation/evaluate_dbh_rmse.py <dataset.json> <predictions.yaml> [output.csv]

Semantic Segmentation Benchmarks

For calculating IoU of inference point cloud (projected to 2D range image) compared to ground-truth:

python3 semantic_segmentation/evaluate_iou.py stacked_image.png <env>.yaml
python3 semantic_segmentation/evaluate_iou.py ground_truth.png pred.png <env>.yaml

Citation

You can access the paper from arXiv. To cite our work, please use:

@misc{cheng2023treescope,
  title={TreeScope: An Agricultural Robotics Dataset for LiDAR-Based Mapping of Trees in Forests and Orchards}, 
  author={Derek Cheng and Fernando Cladera Ojeda and Ankit Prabhu and Xu Liu and Alan Zhu and Patrick Corey Green and Reza Ehsani and Pratik Chaudhari and Vijay Kumar},
  year={2023},
  eprint={2310.02162},
  archivePrefix={arXiv},
  primaryClass={cs.RO}
}

Contributions or Questions?

Please fill-out an issue if you have any questions. Do not hesitate to send your pull request.