68 lines
3.1 KiB
Markdown
68 lines
3.1 KiB
Markdown
|
Skeleton Tracking Based on Kinect Sensor V2
|
||
|
===========================================
|
||
|
|
||
|
This work is to realize communication between Windows and Ubuntu to interact skeleton tracking data based on [rosserial_windows](http://wiki.ros.org/rosserial_windows). Windows's advantage is that its kinect application is easily to install, and it has powerful Kinect SDK to get and process skeleton tracking data directly. Ubuntus's advantage is that it support ROS, and can make robotics system access these data. All of these inspired me to combine these two system, and provide a basis for roboticists to do further work on skeleton tracking.
|
||
|
|
||
|
For this repository, `Kinect_V2_Windows` is folder should only run on Windows, `kinect_v2_Ubuntu` is folder should only run on Ubuntu.
|
||
|
|
||
|
**note**: Before everything starting, make sure your Windows and Ubuntu both connect to the same local network to secure the communication between their ROS clients. And also make sure to plugin Kinect Sensor V2 to USB3.0 port on your windows computer.
|
||
|
|
||
|
### Windows
|
||
|
1. Download and install Kinect for Windows SDK 2.0 from [here](https://www.microsoft.com/en-us/download/details.aspx?id=44561)
|
||
|
|
||
|
2. Verify the successful installation in Device Manager.
|
||
|
|
||
|
3. Open `Kinect v2`, and then install `Kinect Configuration Verifier` to check the success of installation.
|
||
|
|
||
|
### Ubuntu
|
||
|
1. Clone the kinect v2 repository to your ROS workspace and do catkin_make
|
||
|
```bash
|
||
|
cd ~/catkin_ws/src
|
||
|
git clone git@github.com:msr-peng/kinect_v2_skeleton_tracking.git
|
||
|
cd kinect_v2_skeleton_tracking && sudo cp -r kinect_v2 ../ && cd ..
|
||
|
rm -r kinect_v2_skeleton_tracking
|
||
|
cd .. && catkin_make
|
||
|
|
||
|
2. Get your Ubuntu IP address in local network
|
||
|
```bash
|
||
|
ifconfig
|
||
|
```
|
||
|
Then you should get the IP address which behind "inet addr:"
|
||
|
|
||
|
3. Start ROS and rosserial_server
|
||
|
|
||
|
In order for rosserial_windows to communicate with the ROS master, a server socket is needed to connect. Now I just uses a TCP socket, but in theory it is also possible to sue a serial port. On the ROS master, start the ROS core:
|
||
|
```bash
|
||
|
roscore
|
||
|
```
|
||
|
In a separate terminal, start the rosserial server:
|
||
|
```bash
|
||
|
rosrun rosserial_server socket_node
|
||
|
```
|
||
|
Then go back to your windows machine and run the app.
|
||
|
|
||
|
### Windows
|
||
|
1. Clone the repository, and enter into Kinect_V2 folder
|
||
|
|
||
|
2. Set Ubuntu IP address in Windows app
|
||
|
Go to `Kinect_V2.cpp` source file, find line
|
||
|
char master_uri[] = "192.168.0.104";
|
||
|
Then replace the string with your Ubuntu IP address your just got.
|
||
|
|
||
|
3. Run Kinect_V2 solution in Visual Studio
|
||
|
Then windows should send skeleton tracking data to Ubuntu.
|
||
|
|
||
|
### Ubuntu
|
||
|
1. Process skeleton tracking data, publish them as `TF` forms
|
||
|
```bash
|
||
|
rosrun kinect_v2 show_joint
|
||
|
```
|
||
|
|
||
|
2. Visualize skeleton tracking data on `Rviz` (basic frame is `/kinect_v2`)
|
||
|
```bash
|
||
|
rosrun rviz rviz -f kinect_v2
|
||
|
```
|
||
|
Then in `Rviz`, add `TF` to display, you should see the frames of skeleton keypoints.
|
||
|
|
||
|
For more potential applications, please refer to [rosserial_windows](http://wiki.ros.org/rosserial_windows).
|