Project Description

This repository contains the source code and instruction for blueberry flower counting. Detailed methodology and results can be found from our paper.

Setup Instructions

Pip install all requirements in a Python>=3.10 environment with PyTorch>=2.7.1.

Usage

1. Clone this repository.
2. cd ./src
3. Run ./train.py to train the detection model.
4. Run ./test.py to evaluate the model.
5. Run ./inference_on_raw_imgs.py to get the predictions on raw images.

Data Information

The dataset will be made available after the paper is published.

Acknowledgement

• ultralytics
• SAHI
• Powerful-IoU