Related Experiment Video
Updated: Jul 5, 2025
08:47
Author Spotlight: UAV Remote Sensing for Efficient Invasive Plant Biomass Estimation
Published on: February 9, 2024
1.4K
Maize plant detection using UAV-based RGB imaging and YOLOv5
Chenghao Lu1, Emmanuel Nnadozie1,2, Moritz Paul Camenzind1
1Precision Agriculture Lab, School of Life Sciences, Technical University of Munich, Freising, Germany.
Frontiers in Plant Science
|January 19, 2024
Summary
This study uses You Only Look Once version 5 (YOLOv5) and Unmanned Aerial Vehicle (UAV) images to accurately detect and count maize plants. The developed method shows high performance in realistic field conditions, even with weeds or occlusions.
Area of Science:
- Agricultural Science
- Computer Vision
- Machine Learning
Background:
- Computer vision (CV) and machine learning (ML) offer advanced image analysis for object detection.
- Unmanned Aerial Vehicle (UAV) high-resolution imagery enables CV and ML applications in plant science.
- Automated plant detection is crucial for crop monitoring and management.
Purpose of the Study:
- To develop a practical workflow for detecting and counting maize plants using YOLOv5 and UAV images.
- To evaluate the performance of YOLOv5 across different maize developmental stages.
- To reduce manual labeling efforts using the Segment Anything Model (SAM).
Main Methods:
- Utilized You Only Look Once version 5 (YOLOv5) for object detection.
- Employed Unmanned Aerial Vehicle (UAV) based high-resolution imagery.
- Integrated a semi-automated labeling method with the Segment Anything Model (SAM).
- Applied image-rotation augmentation and low-noise weight adjustments.
Main Results:
- Achieved high mean average precision (mAP@0.5): 0.828 (3-leaf stage) and 0.863 (7-leaf stage).
- YOLOv5 demonstrated robust performance in challenging field conditions (weeds, occlusion, blur).
- Image-rotation and low-noise weight enhancements improved accuracy by 0.024 and 0.016 mAP@0.5, respectively.
Conclusions:
- The YOLOv5 workflow provides a practical and accurate solution for automated maize plant detection using UAV imagery.
- The study highlights the effectiveness of YOLOv5 in real-world agricultural scenarios.
- This research offers a valuable reference for applying deep learning in plant growth characterization.
