YOLO Object Detection & DeepSort Tracking

This project uses YOLOv8 for object detection and DeepSort for object tracking. It supports video file input or real-time camera processing, and is tested on both PC and Raspberry Pi (including Pi 5).

---

Requirements

• Python 3.8+
• Virtual environment (venv)
• Required libraries (see below)
• For Raspberry Pi: Pi OS (Bookworm recommended), camera enabled, GStreamer installed

---

Installation

1. Clone the repository

   git clone https://github.com/NoGenryFord/Exaple-YOLO-model.git
   cd Exaple-YOLO-model

2. Create and activate a virtual environment

   python -m venv venv
   # For Windows:
   .\venv\Scripts\activate
   # For Linux/Mac:
   source venv/bin/activate

3. Install dependencies

   pip install -r requirements.txt

---

Running the Program

Option 1: PyTorch Version (Original)

1. Start the script

   python main.py

Option 2: TensorFlow Lite Version (Optimized for ARM/Raspberry Pi)

1. Convert model to TFLite (if not already done)

   python src/convert_to_tflite/onnx2tf_converter.py

2. Run TFLite version

   python main_tflite.py

3. Test TFLite model

   python test_tflite.py

For Raspberry Pi 5 Deployment

# Install system dependencies
sudo apt update
sudo apt install python3-pip python3-venv libgstreamer1.0-0 gstreamer1.0-plugins-base

# Create virtual environment
python3 -m venv .venv
source .venv/bin/activate

# Install TensorFlow Lite runtime (lighter than full TensorFlow)
pip install tflite-runtime

# Install other dependencies
pip install opencv-python deep-sort-realtime ultralytics numpy

# Run TFLite version
python3 main_tflite.py

Controls (Both Versions)

• ESC: Exit
• g: Toggle grayscale mode
• c: Switch to default camera
• 1: Switch to Raspberry Pi camera (GStreamer)
• v: Restart video
• r: Reset tracker
• r: Reset selection (if implemented)

3. Functionality
   • Object Detection: YOLOv8 model detects objects in video/camera stream.
   • Object Tracking: DeepSort tracks detected objects across frames.
   • Interactive: Switch video sources, toggle modes, and view real-time FPS and confidence.

---

Project Structure

Exaple-YOLO-model/
│
├── main.py                  # Main script (all logic here)
├── weights/
│   └── YOLO/
│       └── model_3_best.pt  # YOLOv8 model weights
├── data/
│   └── tank1.mp4            # Example input video
├── venv/                    # Virtual environment (not in Git)
├── requirements.txt         # Dependencies
└── README.md                # This file

---

Dependencies

Main libraries:

• ultralytics
• deep_sort_realtime
• opencv-python
• numpy

Install with:

pip install -r requirements.txt

Note: Ensure the model weights file (model_3_best.pt) is in the correct folder. For camera use, make sure the camera is connected and enabled on your device.

---

Running on Raspberry Pi 5

• Make sure your Pi OS is up to date and camera is enabled (libcamera-hello should work).
• Install GStreamer and plugins:

  sudo apt update
  sudo apt install -y gstreamer1.0-tools gstreamer1.0-plugins-base gstreamer1.0-plugins-good gstreamer1.0-plugins-bad gstreamer1.0-plugins-ugly gstreamer1.0-libav
  sudo apt install -y gstreamer1.0-libcamera

• For best performance, use the GStreamer pipeline for the Pi camera:
  • In the app, press 1 to switch to the Pi camera (v4l2src device=/dev/video0 ! videoconvert ! appsink).

---

Performance Tips & Compilation

• For best speed on Raspberry Pi, use:
  • Lower video resolution (e.g., 240x240)
  • Lower FPS (e.g., 15)
• For maximum performance, use hardware acceleration (e.g., OpenVINO, Coral, or NPU if available on your Pi).

---

License & Legal

COMMERCIAL SOFTWARE
Copyright © 2025. All rights reserved.

This software is provided under a limited commercial license.
See LICENSE for details.

For licensing, contact:

• Email: your.contact@example.com
• Phone: +XX XXX XXX XXXX