🚗 Vehicle Detection, Classification, and Counting using OpenCV

📌 Jump straight to:

• Demo
• Repository Contents
• Usage
• Features
• Configuration Parameters

---

🎥 Demo

---

📁 Repository Contents

• 📜 main.py: Launches the vehicle counter.

• 🛠 utils.py: Houses utility functions and classes for vehicle operations.

• ⚙️ config.py: Manages parameters such as model paths, video input, and display settings.

• 📦 MODEL/: (yolov4.weights missing in this repo - download)

  • 🧠 yolov4.cfg: YOLO model config.
  • 🔖 coco.names: Recognizable classes by the model.

• 🎥 VIDEO/:

  • 📹 video2.mp4: A test sample.

---

🚀 Usage

1. 🔗 Clone this repository.

2. 📦 Install dependencies:

   pip install -r requirements.txt

3. ⚙️ Adjust paths and parameters in config.py.

4. 🏃‍♂️ Execute:

   python main.py

5. 🖱 In the new window, double-click to position the counting line. Watch as vehicle stats get tallied in real-time!

---

🌟 Features

• 🕐 Real-time detection via YOLOv4.
• 📏 Uses Euclidean distance for tracking.
• 🖱 Set counting line with a double-click.
• 📦 Non-Max Suppression (NMS) to declutter overlapping boxes.
• 📊 Classifies and displays counts: Car 🚗, Motorbike 🏍, Bus 🚌, Truck 🚛.

---

⚙️ Configurations Parameters

• VIDEO_PATH: Pathway to your footage.
• INPUT_SIZE: YOLO model's desired input dimensions.
• CONFIDENCE_THRESHOLD: Desired confidence level for detections.
• NMS_THRESHOLD: Non-max suppression's threshold.
• FONT_COLOR, FONT_SIZE, FONT_THICKNESS: Style the on-screen text.
• CLASSES_FILE: YOLO's class definitions file path.
• REQUIRED_CLASS_INDEX: Indices of classes you wish to monitor.
• MODEL_CONFIG: YOLO's config file path.
• MODEL_WEIGHTS: YOLO's weight file path (download separately).

🛠 Tweak these in config.py to fit your scenario.