yolo-coffee-detection

Preparation

wget https://public-scicrop.s3.amazonaws.com/academy/CoffeebeanDetection.v1i.yolov7pytorch.zip
wget https://public-scicrop.s3.amazonaws.com/academy/Coffee-Fruit-Maturity-v5i.yolov7pytorch.zip
wget https://github.com/WongKinYiu/yolov7/releases/download/v0.1/yolov7_training.pt
pip install -r yolov7/requirements.txt

Choose the dataset

For green coffee normal/defect detection:

unzip CoffeebeanDetection.v1i.yolov7pytorch.zip

For ripe/unripe/semi-ripe/dry coffee fruit detection:

unzip Coffee-Fruit-Maturity-v5i.yolov7pytorch.zip

Clean environment

Before use another dataset, you need to clean your environment:

rm *.yaml
rm *.txt
rm -rf valid/
rm -rf test/
rm -rf train/

Training

python3 train.py --batch 8 --epochs 55 --data data.yaml --weights 'yolov7_training.pt' --device 0

Detection with Trained weights

If you trained the model by yourself using the provided datasets you need to find the files best.pt often creted at yolov7/runs/train subfolders. However, if you prefer to run with our pre-trained weights you can download them from here:

• Green Coffee: https://public-scicrop.s3.amazonaws.com/academy/coffee-best.pt
• Coffee Fruit: https://public-scicrop.s3.amazonaws.com/academy/coffee_fruit_best.pt

Running the detection

Execute the app.py script from root of this repository, with the following parameters:

1. detect: Set detection mode
2. weights: Change by the *.pt weight file that you want to use
3. source: Change by the image source (for detection) you want to use: a) An image file (jpg, png, etc); B) A video file (mp4 etc); C) Camera source. If you choose a Camera source, you need to use the index number of the camera connected to your computer. A webcam, often is defined as 0.

python3 app.py detect <weights> <source>

Results

https://github.com/Scicrop/yolo-coffee-detection/assets/692043/513ce071-0a32-4556-bce7-0072345e10f9 https://github.com/Scicrop/yolo-coffee-detection/assets/692043/7c90975b-f0ba-4a20-840e-9f22d6814489