Skip to content

lbristiel-psee/prophesee_ros_wrapper

 
 

Repository files navigation

Prophesee ROS Wrapper

Event-based vision by Prophesee

The aim of this metapackage is wrapping event-based data from Prophesee sensors using OpenEB software and publishing the event-based data to ROS.

This metapackage contains ROS wrapper (prophesee_ros_driver) and messages (prophesee_event_msgs) for Prophesee sensors.

prophesee_ros_driver is a ROS wrapper including the following nodes:

  • prophesee_ros_publisher - publishing data from Prophesee sensor to ROS topics
  • prophesee_ros_viewer - listening data from ROS topics and visualizing them on a screen

prophesee_event_msgs package contains ROS message types for Prophesee event-based data, including:

  • Event - an event from a Prophesee camera (uint16 x, uint16 y, ros::Time ts, bool polarity)
  • EventArray - a buffer of events (Event[] events)

Supported Prophesee Evaluation Kit Cameras :

  • EVKV2 Gen4.1
  • EVKV3 Gen3.1, Gen4.1
  • EVKV4 IMX636

Requirements

  • Ubuntu 20.04 or 22.04
  • ROS Noetic or ROS Melodic
  • OpenEB - starting from v2.2.0

Installation

First, retrieve and compile OpenEB.

Then, compile the packages:

  • Clone the source code to your catkin workspace (create a workspace, if needed)

        cd catkin_ws/src
        git clone https://github.com/prophesee-ai/prophesee_ros_wrapper.git
        cd ..
    
  • Compile

        catkin_make
    
  • Source the workspace

        source ~/catkin_ws/devel/setup.bash
    

Getting Started

Publishing data from a camera and listening to them

To publish data from Prophesee camera to ROS topics, run:

      roslaunch prophesee_ros_driver prophesee_publisher.launch

The following topics will be published:

  • /prophesee/camera/cd_events_buffer - buffer of CD (Change Detection) events
  • /prophesee/camera/camera_info - information about the camera

To listen data from ROS topics and visualize them:

      roslaunch prophesee_ros_driver prophesee_viewer.launch

Recording data from a live camera to rosbag

To record data from live camera to rosbag:

  • Start the publisher:
      roslaunch prophesee_ros_driver prophesee_publisher.launch
  • Start rosbag recording (choose the topics to record or record all available topics):
      rosbag record -a

Publishing data from a RAW file

To publish data from RAW file to ROS topics and view the data:

  • Update the prophesee_publisher.launch file to set the path to your RAW file (in raw_file_to_read parameter)
      rosed prophesee_ros_driver prophesee_publisher.launch
  • Start the ROS core
      roscore
  • Start the viewer, at first to be sure to not miss any data:
      roslaunch prophesee_ros_driver prophesee_viewer.launch
  • Start the publisher:
      roslaunch prophesee_ros_driver prophesee_publisher.launch

At the end of the RAW file, the publisher will stop on its own, but the viewer won't stop, so it's up to you to quit the viewer.

Recording data from a RAW file to rosbag

To record data from RAW file to rosbag:

  • Update the prophesee_publisher.launch file to set the path to your RAW file (in raw_file_to_read parameter)
      rosed prophesee_ros_driver prophesee_publisher.launch
  • Start the ROS core
      roscore
  • Start rosbag recording (choose the topics to record or record all available topics):
      rosbag record -a
  • Start the publisher:
      roslaunch prophesee_ros_driver prophesee_publisher.launch

Going Further

The ROS wrapper we propose in this repo is a minimal example to get you started. There are multiple ways to enhance it in terms of features and performance. The code is open to contributions, thus do not hesitate to propose pull requests.

We also recommend you to check the metavision_ros_driver repo from berndpfrommer that offers a ROS driver for event-based cameras with some speed and features enhancements.

Contact

In case of any issue, please raise an issue here on GitHub. For any other information contact us.

About

ROS driver for Prophesee event-based sensors

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • C++ 92.4%
  • CMake 7.6%