https://github.com/ninenights195/smart-traffic-light
This is mini project that I made for DEPA Coding 2025 competitive
https://github.com/ninenights195/smart-traffic-light
edge-ai python-3
Last synced: 4 days ago
JSON representation
This is mini project that I made for DEPA Coding 2025 competitive
- Host: GitHub
- URL: https://github.com/ninenights195/smart-traffic-light
- Owner: NineNights195
- License: mit
- Created: 2025-08-16T10:59:46.000Z (2 months ago)
- Default Branch: main
- Last Pushed: 2025-08-27T09:14:32.000Z (about 2 months ago)
- Last Synced: 2025-09-06T18:23:41.134Z (about 1 month ago)
- Topics: edge-ai, python-3
- Language: Python
- Homepage:
- Size: 5.58 MB
- Stars: 4
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# 🚦 Smart Traffic Light
### 💻 The Project for Coding Thailand 2025 - Regional Competition
This project is built for the Coding Thailand 2025 - Regional Competition to simulate and use AI to solve traffic problems.
Using Python and a webcam, the system employs the advanced YOLOv8 (nano version) object detection model to:
- Count persons and vehicles in real time at an intersection.
- Change traffic lights based on what it detected (person is top priority).
The goals for this project are to qualify for the **National round** and to practice my Python skills at the same time. When I enter the national round, I will make the Raspberry Pi version. However, the AI may have issues detecting persons and vehicles and needs improvement in the future.
---
## 🛠️ Requirements
- Linux (Ubuntu/Debian recommended) or Windows with Python installed
- Python **3.12 or 3.13**
- Webcam or external camera
- Git
---
## ⚙️ Setup instruction
### 1. **Clone the Repository**
- Clone default branch (main):
```
git clone https://github.com/NineNights195/smart-traffic-light.git
```
- To clone specific branch:
```
git clone -b https://github.com/NineNights195/smart-traffic-light.git
```
### 2. **Create Virtual Environment**
- Go to your repo
```
cd
```
- Create virtual enviroment to avoids messing up system Python
```
python -m venv venv
```
### 3. **Activate Virtual Environment**
- Windows:
```bash
.\venv\Scripts\activate
```
- macOS/Linux:
```bash
source venv/bin/activate
```
### 4. Ensure pip is avaliable
- If pip is missing inside the venv, run:
```
python -m ensurepip --upgrade
```
### 5. Upgrade pip, setuptools, and wheel
```
pip install --upgrade pip setuptools wheel
```
### 6. **Install Dependencies**
```bash
pip install -r requirements.txt
```
- Requirements dependencies
```
opencv-python
numpy
torch
torchaudio
torchvision
ultralytics
ultralytics-thop
pandas
pillow
tqdm
```
### 7. Download YOLO model
- Create a folder called models inside your project (if it doesn’t exist yet):
```
mkdir -p models
```
- Download the YOLOv8 nano model (yolov8n.pt) into the models/ directory:
```
wget -O models/yolov8n.pt https://github.com/ultralytics/assets/releases/download/v0.0.0/yolov8n.pt
```
Alternatively, if wget is not available, use curl:
```
curl -L -o models/yolov8n.pt https://github.com/ultralytics/assets/releases/download/v0.0.0/yolov8n.pt
```
### 8. Test some scripts
- Now you can test some scripts in scripts folder
- Example the webcam-test.py
```
python webcam-test.py
```
Now it should work properly