🚙 Car Colour Detection along with People Count

Detects blue cars, counts vehicles, and counts people at traffic signals.

---

📌 Overview

This project uses YOLO object detection with a color classification model to detect cars at traffic signals, identify their colors, and count people.

It’s useful for traffic monitoring, enforcement, and analytics.

---

📂 Repository Structure

Car-colour-detection/
├── app.py
│ 
├── models/
│ └── blue_car_classifier.keras
│
├── yolo/
│ └── yolov8s.pt
│
├── style.css
│
├── notebooks/
│ └── Car_Colour_Detection.ipynb
│
├── data/
│ ├── input/
│ │ └── sample_inputs
│ └── output/
│   └── sample_outputs
│
├── results/
│ ├── evaluation_metrics.png
│ ├── confusion_matrix.png
│ └── classification_report.png
│
├── requirements.txt
└── README.md

---

🚀 Features

• 🚗 Colour-specific car detection
• 🔲 Bounding Box for cars
• 📊 Display of car colour & count + people count
• ⚡ Works with images and video files
• 🎯 Achieved ~96% Accuracy

---

📂 Dataset

• Source: Car Colours Dataset
• Format Used: Images sorted into color-labeled folders

---

📜 How It Works

1. Detection → YOLO detects cars & people in each frame
2. Classification → Car colours are identified
3. Bounding Boxes →
   • Blue cars → Red boxes
   • Other cars → Blue boxes
4. Counting → Number of cars & people displayed on top

---

📈 Example Output

---

▶️ How to Run

In GitHub Codespaces

# Update the package list and then install the libgl1 graphics library
sudo apt-get update && sudo apt-get install libgl1

# Install dependencies
pip install -r requirements.txt

# Run the Streamlit app
streamlit run app.py

On Local System

# Clone the repository
git clone https://github.com/ShravyaMalogi/Car-colour-detection.git
cd Car-colour-detection

# Install dependencies
pip install -r requirements.txt

# Run the Streamlit app
streamlit run app.py