eroniki / build_daq

build_daq contains the code and tools needed for multi-cam video acquisition and analysis of human poses

BUILD_DAQ: A data acquisition system for multi-room, multi camera setup

Maintainer: Murat Ambarkutuk

Email: [e-mail]

VAST Lab, Mechanical Engineering Department,

Virginia Tech, Blacksburg, VA

Summary:

This repository contains the necessary tools and packages needed for Steelcase funded project called BUILD. This project involves with human pose detection in 3D with multiple cameras placed in different room located in Cheatham Hall, Virginia Tech. In this repository, one can find a various python modules and classes to create a RESTful API to interact with the Image Acquisition system as well as processing the collected images.

Installation

Dependencies:

OpenPose is needed to extract poses from the collected images. Please make sure OpenPose is installed correctly (with Python bindings). You can follow this link to obtain a comprehensive guide to install Openpose.

Obtaining the Repository

1. Open a new terminal window with CTRL+ALT+T
2. Run the commands sequentially: Clone the repository:

git clone https://github.com/eroniki/build_daq.git

Enter to the repository folder:

cd build_daq

Install the dependencies:

sudo -H pip install -r requirements.txt

Reboot the system:

sudo reboot now

Usage:

Upon completion of the installation, the system can be started directly with

python main.py

This script creates the REST API and starts serving a WEB-interface with which the system can be manipulated.

One can find the served with interface (locally):

http://0.0.0.0:5000

or if you'd like to access the system remotely, you can use a link similar to the one in below where THE_IP_ADRESS_OF_THE_LOCAL_MACHINE is the global IP address of the computer serving the system.

http://THE_IP_ADRESS_OF_THE_LOCAL_MACHINE:5000

Documentation

Main Code

The main controller of the code lay in main.py file. This python script creates and initializes all the classes needed in the project.

Flask App

The basic User Interface is created by using Flask, a web development tool for Python.

All the necessary tools and classes were implemeted in this class, which reside in app.py module. The documentation of the computer_vision class is automatically generated from the docstrings of the classes and the functions in it. The documentation is located here.

Utility Class

The documentation of the utility class is automatically generated from the docstrings of the classes and the functions in it. The documentation is located here.

Experiment Class

The documentation of the experiment class is automatically generated from the docstrings of the classes and the functions in it. The documentation is located here.

Computer Vision Class

The documentation of the computer_vision class is automatically generated from the docstrings of the classes and the functions in it. The documentation is located here.

Pose Detection Class

Camera Class

The documentation of the camera module and its contents are automatically generated from the docstrings of the classes and the functions in it. The documentation is located here.

Calibration Process

This calibration process is provided here

Contribution

If you would like to contribute to this project, please make sure you read and understand the contribution workflow.