NVIDIA Developer Blog

In this repository, we build ros2 wrapper for trt_pose for real-time pose estimation on NVIDIA Jetson.

• Pose message
• Pose detection image message
• Visualization markers
• Launch file for RViz

• ROS 2 Eloquent:
  
  • Docker Support
    
  • Install Instructions
    
• trt_pose
  • Dependencies for trt_pose
    
  • [Install trt_pose] (https://github.com/NVIDIA-AI-IOT/trt_pose#step-2---install-trt_pose)
    

• Clone repository under ros2 workspace
  

$ cd ros2_ws/src/
$ git clone https://github.com/NVIDIA-AI-IOT/ros2_trt_pose.git

• Install requirements using rosdep
  

$ rosdep install --from-paths src --ignore-src --rosdistro eloquent -y

• Build and Install ros2_trt_pose package
  

$ colcon build
$ source install/local_setup.sh

• Change Power Mode for Jetson sudo nvpmodel -m2 (for Jetson Xavier NX)
  sudo nvpmodel -m0 (for Jetson Xavier and Jetson Nano)
  

• Input Images are captured using image_tools package
  ros2 run image_tools cam2image

• Keep trt_pose related model files in base_dir, it should include:
  

  • Model files for resnet18 or densenet121 download link
  • Human Pose points json file

• For running you can use Launch file or manually run each node:
  

  • Run using Launch file
    

  $ ros2 launch ros2_trt_pose pose-estimation.launch.py
  

  Note: Update rviz file location in launch file in launch/pose_estimation.launch.py
  

  • Run ros2_trt_pose node
    

  $ ros2 run ros2_trt_pose pose-estimation --ros-args -p base_dir:='<absolute-path-to-base_dir>'
  

  • For following use separate window for each:
    
    • See Pose message
      

    $ source install/local_setup.sh
    $ ros2 run rqt_topic rqt_topic
    

    • Visualize markers
      

    $ ros2 run rviz2 rviz2 launch/pose_estimation.rviz
    

• ros2_torch_trt : ROS2 Real Time Classification and Detection
  
• ros2_deepstream : ROS2 nodes for DeepStream applications
  
• ros2_jetson_stats : ROS 2 package for monitoring and controlling NVIDIA Jetson Platform resources