Traffic-analysis
AI based Decision support system for Traffic Control
AI based Decision support system for Traffic Control is an helpful tool for traffic cops to analyze the traffic using ML and IoT.
Table of contents
- Motivation
- Problem Statement
- Setup
- Algorithm
- Deployment of OneAPI
- Output
- Web dashboard
- Overview of the project
01 Motivation
Knowing that traffic officers won’t have the view of the whole traffic, this project has a web-based dashboard to assist traffic officers in making decisions depending on the volume of traffic only by visualizing the count of vehicles from each lane. This would serve as a hybrid way of traffic support to the cop instead of fully automating traffic control.In this project we have also implemented the algorithm using Intel oneAPI Toolkit.
02 Problem Statement
Image Recognition: Develop a computer vision project using Scikit-Learn’s image processing capabilities. You could use a dataset of images to train a model to recognize specific objects or classify images into categories.
03 Setup
We have used IntelOneApi Toolkit to make the system run in any environment and it extends the code portability across all processor architectures. Hence, in our model, the usage of oneAPI served to be helpful in terms of optimization.
04 Overview of the project
Explanation & implementation of the project : Link
Source code: Link
05 Implementation - Algorithm
Deep Learning Algorithm - Yolov5 + Deepsort with PyTorch The detections generated by YOLOv5, a family of object detection architectures and models pretrained on the COCO dataset, are passed to a Deep Sort algorithm which tracks the objects. For backend, Flask and Jinga was used For database - Firebase (No-SQL) in cloud and Sqlite in edge was used.
06 Deployment of OneAPI
OneAPI is used enable the use of one platform for a range of different hardware, hence it eliminated the need for different languages, tools, and libraries when to code for CPUs and GPUs. Openvino was used in this project which helped in optimization of the computer vision packages that were used including OpenCV and other DL packages required for it.
Implementing OpenVINO link
Steps
Step 1: Create virtual environment
python -m venv openvino_env
Step 2: Activate virtual environment
openvino_env\Scripts\activate
Step 3: Upgrade pip to latest version
python -m pip install --upgrade pip
Step 4: Download and install the package
pip install openvino-dev==2022.3.0
How to run,
• Create a virtual environment and activate it • Download the packages using the command,
pip install -r requirements.txt
• In a terminal, run mainProgram.py
python mainProgram.py
• In another terminal, run the frontend using the command,
python -m flask run
• CTRL + click on the link to open the web dashboard.
07 Output - Count taken
Real time dataset
**Prototype live detection and count**
# 08 Web dashboard
Final view of the Traffic shown in the web dashboard.
