Warehouse-Object-detection

Warehouse Package and Label Detection

This project uses YOLOv8 for detecting packages and labels in warehouse images and videos. The dataset is sourced from Roboflow and annotated in YOLO format.

Dataset

• Classes: label, package
• Source: Roboflow Universe
• Format: YOLOv8 (see data.yaml)
• Train images: train/images/
• Validation images: valid/images/
• Test images: test/images/

Getting Started

1. Clone the repository

git clone <repo-url>
cd <repo-directory>

2. Install dependencies

pip install -r requirements.txt

3. Training

Train a YOLOv8 model on your dataset:

python main.py

• This uses yolov8n.pt as the base model and trains for 100 epochs on GPU (CUDA).

4. Testing on Images

Run detection on test images and save results to the results/ directory:

python test.py

• Input: images from test/images/
• Output: annotated images in results/

5. Video Inference

Run detection on a video and save the output video:

python video_test.py

• Input: video from videos/1stvideo.mp4 (edit path as needed)
• Output: videos/output_video_with_counts.mp4

Model Weights

• Trained weights are saved in runs/detect/train5/weights/best.pt by default.
• You can change the weights path in test.py and video_test.py as needed.

Requirements

See requirements.txt for all dependencies.

Notes

• The project expects a CUDA-capable GPU for training and inference.
• The dataset and code are provided under the CC BY 4.0 License.

Acknowledgements

• Ultralytics YOLOv8
• Roboflow