SAHI & YOLOv8 Integration for Enhanced Object Detection

This repository provides a Python project that integrates SAHI (Slicing Aided Hyper Inference) with YOLOv8 for enhanced object detection. The project supports detection on images, video files, and real-time webcam feeds, enabling more accurate results even in high-resolution and complex scenes.

Getting Started

Installation

1. Clone the repository:

   git clone https://github.com/SihabSahariar/Yolov8-SAHI-Inference/
   cd Yolov8-SAHI-Inference

2. Install the required dependencies:

   pip install -r requirements.txt

3. Run the main script:

   python main.py

Yolov8 vs Yolov8+SAHI

Features

• Supports Multiple Inputs: Works with images, video files, and webcam streams.
• Enhanced Detection: Integrates SAHI to better handle large images.
• Real-Time Processing: Provides real-time object tracking and detection.
• Extensible Code: Easy to modify and extend with additional features.

Use Cases

• Small Object Detection: SAHI helps in detecting smaller objects in large images by slicing the image into manageable sections, ensuring that small objects are not missed.
• Drone-Based Object Detection: Ideal for aerial imagery captured by drones where the field of view is wide, and objects may appear small due to altitude.
• Surveillance Systems: Effective in analyzing high-resolution security footage for identifying people, vehicles, or unusual activities.
• Wildlife Monitoring: Useful for detecting animals in high-resolution aerial or ground camera images where details matter.
• Industrial Inspections: Assists in detecting defects or items in large-scale images captured for quality control.

Project Structure

Yolov8-SAHI-Inference/
├── helper.py          # Custom Draw Box Function
├── sort.py            # Implementation of the SORT (Simple Online and Realtime Tracking) algorithm for object tracking
├── test.jpg           # Sample image for testing the object detection functionality
├── YoloSahi.py        # Script that integrates SAHI with YOLOv8, handles input sources (image, video, webcam), and processes detections
├── main.py            # Main Script to run inference on Webcam/RTSP/Video/Image
├── yolov8n.pt         # YOLOv8 pre-trained model file used for detection (YOLOv8 nano model)
└── README.md          # (Recommended) Project description and usage instructions

Contributing

Contributions, suggestions, and feature requests are welcome! Please open an issue or submit a pull request.

License

This project is licensed under the MIT License.

Contact

For any questions or support, please contact me at: sihabsahariarcse@gmail.com