TrafficPulse is a project focused on automating traffic data collection using Machine Learning techniques. It leverages computer vision and neural networks to accurately count and categorize different types of vehicles in urban settings.
Follow these steps to run a demo of TrafficPulse:
-
Install Requirements
You can set up the environment using either pip or conda. Run one of the following commands:
Using pip:
pip install -r requirements.txtUsing conda:
conda create --name TPenv python=3.10 --file requirements.txt -
Run the Demo
In the root folder, execute the demo script:
python3 demo.py -
Line Counter Setup
Once the demo starts, you must select two points on the screen. These points will define the line used for counting vehicles.
-
Demo Execution
If the installation is correct, the tracker will run on a 30-second video demonstrating the traffic counting capabilities of TrafficPulse.
To maintain code quality and consistency, we adhere to the following standards:
- Code Style: Follow the PEP8 standard.
- Documentation: Every function must include a docstring explaining its purpose and usage.
- Git Practices: Commit messages and code reviews should be concise and informative.
- Pre-commit Hook: We use pre-commit hooks to enforce PEP8 standards, primarily through the 'black' code formatter.
Ensure that only necessary packages are included before freezing the requirements. To manage the dependencies, use:
To generate the requirements file:
pip freeze > requirements.txt
To install dependencies from the requirements file:
pip install -r requirements.txt