-
Notifications
You must be signed in to change notification settings - Fork 64
Quick Start Example
We have found it very helpful to keep a document on one of the tracking computers, listing, in the appropriate order, all the commands needed to calibrate and run tracking for the specific OpenPTrack system.
Below is an example of one of our quick start guides, which has three (3) Kinect v2s, named in the network_camera.yaml
as kinectv2_left
, kinectv2_right
, and kinectv2_center
on three computers with IP addresses of 192.168.100.101, 192.168.100.102, and 192.168.100.103 respectively:
First, on the master computer (192.168.100.101), run:
roscore
roslaunch opt_calibration calibration_initializer.launch
Now, on all computers in the OpenPTrack system including .101 (ssh -XC 192.168.100.102 and ssh -XC 192.168.100.103), run:
roslaunch opt_calibration listener.launch
Then, on each computer connected to a Kinect v2, run:
sudo ~/workspace/ros/catkin/devel/lib/kinect2_bridge/./kinect2_bridge
Ignore the error!
Now, on the computer with the IP address of 192.168.100.101, run:
roslaunch opt_calibration opt_calibration_master.launch
roslaunch opt_calibration sensor_kinectv2_left.launch
Again, on the computer with the IP address of 192.168.100.102 (ssh -XC 192.168.100.102), run:
roslaunch opt_calibration sensor_kinectv2_right.launch
Lastly, on the computer with the IP address of 192.168.100.103 (ssh -XC 192.168.100.103), run:
roslaunch opt_calibration sensor_kinectv2_center.launch
Calibrate per the standard procedure. After the completion of calibration, save the calibration file:
rostopic pub /opt_calibration/action std_msgs/String "save" -1
Next, with the listeners
still running:
roslaunch opt_calibration detection_initializer.launch
To finish the calibration process, kill all the terminal commands that are still running. Now you can move on to the detection and tracking start-up process.
N.B.: The sudo ./kinect2_bridge
command only needs to be run again if the computer has been restarted.
To start the detection and tracking process, on the master machine (192.168.100.101), run:
roscore
roslaunch tracking tracking_node.launch
roslaunch detection detection_node_kinectv2_left.launch
Next, on the computer with the IP address of 192.168.100.102, run:
roslaunch detection detection_node_kinectv2_right.launch
Last, on the computer with the IP address of 192.168.100.103, run:
roslaunch detection detection_node__kinectv2_center.launch
Now person detection and tracking will be running, and OpenPTrack will track people in your active tracking area!
N.B.: The instructions above for detection and tracking activate the person tracking, pose annotation and object detection modules at the same time. It is possible to activate or deactivate modules by running the commands with flags as shown below and setting the flags to true or false. In the examples shown below for the master machine(192.168.100.101), only the person tracking module is activated.
roslaunch tracking tracking_node.launch enable_pose:=false enable_object:=false enable_people_tracking:=true
roslaunch detection detection_node_kinectv2_left.launch enable_pose:=false enable_object:=false enable_people_tracking:=true
- System Requirements
- Supported Hardware
- Initial Network Configuration
- Example Hardware List for UCLA Setup
- Making the Checkerboard
- Time Synchronization
- Pre-Tracking Configuration
- Camera Network Configuration
- Single Camera
- Setting Parameters
- Multi-Sensor Person Tracking
- HOG vs YOLO Detectors
- World Coordinate Settings
- Single Camera
- Pose Initialization
- Multi Sensor Pose Annotation
- Pose Best Practices
- Setting Parameters
- Single Camera
- Setting Parameters
- Multi Sensor Object Tracking
- YOLO Custom Training & Testing
- Yolo Trainer
- Single Camera
- Setting Parameters
- Multi Sensor Face Detection and Recognition
- Face Detection and Recognition Data Format
How to receive tracking data in: