openpose_ros

Example ROS catkin package that utilizes the OpenPose library from https://github.com/CMU-Perceptual-Computing-Lab/openpose.

System

Tested on:

• Ubuntu 18.04
• ROS melodic
• OpenCV 4.5
• CUDA 10.0
• cudnn 7.5.0
• Openpose GPU Version

Installation Steps

1. install Openpose. see: https://github.com/CMU-Perceptual-Computing-Lab/openpose/blob/master/doc/installation.md or see: https://medium.com/@erica.z.zheng/installing-openpose-on-ubuntu-18-04-cuda-10-ebb371cf3442

2. Clone this repository into your catkin_workspace/src directory.

   git clone https://github.com/MarkChangKM/openpose_ros.git

3. Modify the model_folder line in openpose_ros/src/openpose_flags.cpp to where openpose is installed (line 18).

   DEFINE_string(model_folder,             "/path/to/openpose/models/",      "Folder path (absolute or relative) where the models (pose, face, ...) are located.");

4. Modify the image_topic parameter in openpose_ros/launch/openpose_ros.launch to the image_topic you want to process.

   <param name="image_topic"     value="/camera/color/image_raw" />
   <param name="depth_topic"     value="/camera/aligned_depth_to_color/image_raw" />

5. Modify the other parameters in openpose_ros/src/openpose_flags.cpp and openpose_ros/launch/openpose_ros.launch to your liking such as enabling face and hands detection.

6. Run catkin_make from your catkin_workspace directory.

Running

1. setup enviroment

source catkin_workspace/devel/setup.bash

2. run with rosbag or realsence 3d:

roslaunch openpose_ros openpose_ros_3d.launch

realsence:

roslaunch openpose_ros openpose_ros_with_rs.launch

TODO

1. filter depth image(filling nan point)
2. filter blocked point(estimate keypoint that has been blocked by other body part)