---
license: agpl-3.0
library_name: ultralytics
tags:
- object-detection
- yolov8
- beetle
- insect
- computer-vision
datasets:
- roboflow
metrics:
- map
model-index:
- name: beetle-detection-yolov8
  results:
  - task:
      type: object-detection
    dataset:
      type: beetle-detection
      name: Beetle Detection Dataset
    metrics:
    - type: map
      value: 0.9763
      name: mAP@0.5
    - type: map
      value: 0.8956
      name: mAP@0.5:0.95
---

# YOLOv8 Beetle Detection Model

## Model Description

This is a YOLOv8-based object detection model fine-tuned for beetle detection. The model was trained on a custom dataset of 500 beetle images from Roboflow and achieves excellent performance with mAP@0.5 of 97.63%.

## Model Details

- **Base Model**: YOLOv8n (nano) from Ultralytics
- **Task**: Object Detection
- **Classes**: 1 (beetle)
- **Input Size**: 640x640 pixels
- **Framework**: PyTorch
- **License**: AGPL-3.0 (inherited from YOLOv8)

## Performance Metrics

| Metric | Value |
|--------|-------|
| mAP@0.5 | 97.63% |
| mAP@0.5:0.95 | 89.56% |
| Precision | 95.2% |
| Recall | 94.8% |
| Processing Time (CPU) | ~100ms per image |

## Dataset

- **Source**: Roboflow Universe
- **License**: CC BY 4.0
- **Images**: 500 annotated beetle images
- **Split**: 80% train, 15% validation, 5% test
- **Augmentations**: Applied during training for robustness

## Usage

### Installation

```bash
pip install ultralytics
```

### Python Inference

```python
from ultralytics import YOLO
import cv2

# Load the model
model = YOLO('best.pt')

# Run inference
results = model('path/to/image.jpg')

# Process results
for result in results:
    boxes = result.boxes
    for box in boxes:
        # Get coordinates and confidence
        x1, y1, x2, y2 = box.xyxy[0].cpu().numpy()
        confidence = box.conf[0].cpu().numpy()
        
        print(f"Beetle detected with confidence: {confidence:.2f}")
        print(f"Bounding box: ({x1}, {y1}, {x2}, {y2})")
```

### Command Line

```bash
yolo predict model=best.pt source='path/to/image.jpg'
```

## Training Details

- **Epochs**: 100
- **Batch Size**: 16
- **Optimizer**: AdamW
- **Learning Rate**: 0.01 (initial)
- **Hardware**: Google Colab GPU
- **Training Time**: ~2 hours

## Applications

This model is designed for:
- Agricultural monitoring
- Entomological research
- Biodiversity studies
- Educational purposes
- IoT-based pest detection systems

## Limitations

- Trained specifically on beetle images
- Performance may vary with different lighting conditions
- Best results with clear, well-lit images
- Single class detection only

## Model Files

- `best.pt`: PyTorch model weights (recommended)
- `best.onnx`: ONNX format for cross-platform deployment

## Citation

If you use this model in your research, please cite:

```bibtex
@model{beetle-detection-yolov8,
  title={YOLOv8 Beetle Detection Model},
  author={Insect Detection Training Project},
  year={2025},
  url={https://huggingface.co/Murasan/beetle-detection-yolov8}
}
```

## License

This model is licensed under AGPL-3.0, inherited from the original YOLOv8 implementation by Ultralytics.

### Base Model Attribution

- **YOLOv8**: [Ultralytics YOLOv8](https://github.com/ultralytics/ultralytics)
- **Original License**: AGPL-3.0
- **Paper**: [YOLOv8: A Real-Time Object Detection Algorithm](https://arxiv.org/abs/2305.09972)

## Related Projects

- [Base Training Repository](https://github.com/Murasan201/insect-detection-training)
- [Hailo 8L Deployment Guide](https://github.com/Murasan201/insect-detection-training/blob/main/HAILO_DEPLOYMENT_GUIDE.md)

## Contact

For questions or issues, please open an issue in the [base repository](https://github.com/Murasan201/insect-detection-training/issues).