Ishmam Tashdeed* · Md Taukir Azam Chowdhury*

VeloTrack is a vehicle speed analysis system designed to estimate vehicle speeds from CCTV footage. This repository contains the tools and code necessary to process video, detect vehicles, track their movement, and calculate their speed. The system also aims to document relevant details about the vehicles observed, providing a comprehensive analysis of traffic flow and potential speed violations.

Overview of VeloTrack Framework

The different features of VeloTrack – - Detecting Vehicles: Can detect multiple types of vehicles in a cluttered environment - Vehicle Tracking: Tracks each instance of a vehicle, assigning a unique identifier to them - Predicting Vehicle Speed: Finds vehicle speed from the captured frames and detection and tracking results - Producing Comprehensive Report: Logs each entry and creates a comprehensive report

Input and Output of the Framework

The system takes a CCTV video footage such as:

And produces two outputs: - Cropped frames containing a vehicle along with it's unique identifier - Comprehensive report containing the speed of each vehicle

Installation

This repository is tested on Python 3.8. Create a virtual environment to install all the dependencies.

1. To install torch visit the PyTorch website and follow the instructions.

2. Install requirements.txt file running the command:

   pip install -r requirements_cuda.txt

Video Folder

⚠️ Ensure that you have the video in the Data folder. ⚠️

Setting Road Parameters

The following parameters must be set in main.py from the input video before executing the program- - Add the ROI (Region of Interest) co-ordinates to further optimize the system - Add the co-ordinates for the entry, exit, and deletion areas from the CCTV footage - Add the distance between the entry and exit areas

Execution

To run the framework, execute the following command: python main.py

Results

The resulting .csv file contains each vehicle's identifier, timestamp, and speed.

Acknowledgement

The vehicle detection framework uses components from YOLOv5 and the vehicle tracking framework uses components from DeepSort

Citation

@misc{Tashdeed_VeloTrack_2023, author = {Tashdeed, Ishmam and Chowdhury, Md Taukir Azam}, title = {{VeloTrack: A Vehicle Speed Analysis System}}, version = {1.1}, year = {2023}, publisher = {GitHub}, journal = {GitHub repository}, url = {https://github.com/ishmamt/VeloTrack} howpublished = {\url{https://github.com/ishmamt/VeloTrack}} }