Skip to content

Latest commit

 

History

History
44 lines (39 loc) · 2.5 KB

README.md

File metadata and controls

44 lines (39 loc) · 2.5 KB

behaviourdetection

This repository contains the source code for the ROS workspace including the NTNU Cyborg's Computer Vision Module, which should be installed on the Jetson TX1 Development Kit. The system uses the first generation ZED Stereo Camera to detect human behaviour.

Node name: jetsontx1_cvmodule Language: Python

Dependencies

Following packages and ther dependendencies should be installed, before proceeding to the "Installation" section:

  1. OpenCV (Allready installed if TX1 is flashed with JetPack)
  2. CUDA (Allready installed if TX1 is flashed with JetPack)
  3. Python 2.7 (Allready installed if TX1 is flashed with JetPack)
  4. ZED SDK
  5. ZED Python API
  6. scikit-learn
  7. scikit-image
  8. FilterPy

Installation

  1. Install and configure a catkin ROS environment by following the ROS tutorials
  2. Clone this repository and copy the folder "jetsontx1_cvmodule" into the src folder in your ROS workspace where the ROS packages should be located.
  3. Build the ROS package by using the command "catkin_make" in the catkin workspace terminal window.
  4. Locate the pyyolo folder in the terminal and install using the commands as following:
    1. make
    2. rm -rf build/
    3. python setup_gpu.py build
    4. sudo python setup_gpu.py install

Run the system

To launch the main code publishing info about detected human behaviour:

rosrun jetsontx1_cvmodule behaviourdetection.py

To launch the example of a code subscribing to the Topics published by behaviourdetection.py:

rosrun jetsontx1_cvmodule subscribertest.py

Instead of running behaviourdetection.py, you can play the rosbag file testbagfile.bag:

rosbag play testbagfile.bag

If subscribertest.py is allready running it should visualize the a sample of the published topic ”videostream”. Also, it will print information about detected people ID and smile/face indication, in the terminal.

Note: This installation is tested on a Jetson TX1 board running a version of Ubuntu 18.04 with ROS Melodic.

Note: Read the report on how to integrate the CV-module on the Jetson board with the Cyborg.