MatchPlant: An Open-Source Pipeline for UAV-based Single-Plant Detection and Data Extraction

Description:

This dataset accompanies the research study “MatchPlant: An Open-Source Pipeline for UAV-Based Single-Plant Detection and Data Extraction.” It includes UAV (Unmanned Aerial Vehicle) images, annotated bounding boxes for plant detection, and a pre-trained Faster R-CNN model for detecting individual maize in UAV imagery. The dataset is designed to support research in high-throughput plant phenotyping, plant genetics research, and agricultural computer vision applications.

Dataset Content:

1. UAV images: Undistorted images created by OpenDroneMap software based on high-resolution RGB (Red-Green-Blue) images collected by UAV during the 2021 growing seasons. The images contain a maize field for plant genetics research.
2. Annotation file: Bounding box annotations in COCO (Common Objects in Context) format (.json files). The bounding boxes were drawn on the UAV images.
3. Pre-trained model: The Faster R-CNN (Faster Region-based Convolutional Neural Network) model was trained on the UAV images and the annotation file dataset to detect individual maize in plant genetics research.

Data Collection and Processing:

• UAV images were captured at nadir angles using a high-resolution RGB camera (20 megapixels) mounted on a UAV.
• Orthorectification and preprocessing were conducted using OpenDroneMap software.
• Bounding boxes were manually created using developed by Python software (Module "4_bbox_drawer": GitHub: https://github.com/JacobWashburn-USDA/MatchPlant) and verified for individual plant detection.
• The Faster R-CNN model was trained on undistorted images to improve detection accuracy.

Usage:

This dataset can be used to train, validate, and test deep learning models for single-maize detection in agricultural fields. It also supports research in high-throughput plant phenotyping and UAV-based plant monitoring.

Acknowledgments:

This research was supported in part by an appointment to the Agricultural Research Service (ARS) Research Participation Program administered by the Oak Ridge Institute for Science and Education (ORISE) through an interagency agreement between the U.S. Department of Energy (DOE) and the U.S. Department of Agriculture (USDA). ORISE is managed by ORAU under DOE contract number DE-SC0014664. Funding was provided by the United States Department of Agriculture, Agricultural Research Service, and SCINet Postdoctoral Fellows Program. All opinions expressed in this publication are the author’s and do not necessarily reflect the policies and views of USDA, DOE, or ORAU/ORISE.

Contact Information:
For further inquiries regarding this dataset, please contact Jacob D. Washburn at jacob.washburn@usda.gov.