Tutorial

Alex Ward & Adrian Josan

2024:3




OptiTrack

The OptiTrack system is a high-precision motion capture technology used to track the movement of objects and people in real time. It utilizes a network of infrared cameras and reflective markers placed on subjects or objects to capture their precise 3D positions and movements. The system processes this data to create accurate, real-time representations of the captured motion.

The OptiTrack system is part of the interaction lab, as the 14 IR cameras are fixed on the ceiling fixtures, therefore, the tracking space falls within their perimeter, occupying roughly half of the lab’s surface. To use the system, contact the admin team to ask for their permission and rent the necessary gear, which includes:

  • Calibration wand (fig.1)
  • Plastic container, containing the calibration square (fig.2), and reflective velcro markers (fig.3)
Calibration Wand (fig.1) Calibration Square (fig.2) Velcro Reflective Markers (fig.3)
  • Mocap suit bag (fig.4) (including vest (fig.5), pants (fig.6), and a hat (fig.7))
Mocap Suit Bag (fig.4) Mocap Vest (fig.5) Mocap Pants (fig.6) Mocap Hat (fig.7)
  • OptiTrack-dedicated PC w. Motive software
OpitTrack PC
  • OptiTrack security key (which should be already be plugged into the PC)

Additionally, you need to ensure that the lab space is empty for the duration of the recording as moving objects may negatively affect the quality of the motion tracking. To power the system, click the side button on the system’s power socket switch. The power switch status light should be green when on.

Motive is the software that pairs with the OptiTrack cameras to track and capture the motion of the markers. Motive is installed on the OptiTrack PC. To get it running:

  1. Ask one of the lab technicians for the password to the OptiTrack PC. Turn on the computer and use the password to unlock it.
  2. Launch the Motive application from the Desktop. The program should be automatically unlocked by the hardware key connected to the PC.

Now, you should be greeted by the main UI of Motive. If all cameras are correctly detected and running, there should be 14 camera perspectives in the Cameras View, each showing currently visible (reflective) objects.

Motive's Main UI

To effectively use the OptiTrack system, you need to have a general idea of how Motive 3.x works and looks like, including its main UI elements, which include:

  1. Left UI column:
    1. Calibration Pane - used for the camera calibration process
    2. Data Pane - where all session directories and corresponding motion recording files are found
    3. Builder Pane - to create skeletons and rigid bodies
  2. Viewports (center UI column):
    1. 3D Viewport (top) - where all 3D data is shown, including the cameras, ground plane and tracked objects
    2. Cameras Viewport (bottom) - where you can monitor the view, apply mask filters, examine markers, etc. for each camera
  3. Right UI column:
    1. Devices Pane (top) - lists connected cameras and their configuration
    2. Assets Pane (top) - includes the currently tracked rigid bodies and skeletons
  4. Control Deck
    1. Switches between:
      1. Live mode - showing the real-time 3D view from all cameras
      2. Edit mode - used to process a recording file.
    2. Starts/Stops recording

To learn more about Motive's UI panels, consult the official OptiTrack documentation.

The calibration of the cameras is necessary for Motive to reconstruct 2D camera data into 3D markers in real time. During calibration, it computes the position and orientation of each camera and masks reflective objects within the tracking space.

If there are any setup changes, the system must be recalibrated. Also, the calibration accuracy might decrease over time if the cameras change their position or new reflective objects are introduced in the tracking space.

The Continous Calibration feature is enabled (in the bottom portion of the Calibration Pane) which continuously monitors and refines the camera calibration. In theory, you can use the last calibration file (which is loaded automatically) and not have to worry about small displacements on the camera due to their mounts, etc. However, considering Interaction Lab’s dynamic nature it is best to check the quality of new recordings to ensure that the system is calibrated correctly. If you notice that the recording is off, perform a new full calibration.

To perform a full calibration:

  1. Ensure that the cameras are pointing to the center of the tracking space.
    1. In the Cameras Viewport, right-click a camera perspective.
    2. For the “Video Mode” option, select “Grayscale”.
    3. Using the grayscale video image, adjust the position of the camera such that the crosshair points to the center of the tracking space.
    4. Change the “Video Mode” option back to “Object” when done.
  2. Remove all reflective objects from the tracking space. These objects are represented as white dots in a camera's perspective. If an object can’t be removed, try covering it with a non-reflective material.
  3. In the Calibration Pane, click on “New Calibration”.
  4. Mask the reflective objects that can’t be removed physically. By doing so the system will ignore them. To perform “masking”: (Example video 00:50-02:03)
    1. Click on the “Mask” button in the Calibration Pane. This will automatically mask all reflective objects detected by the cameras.
    2. If the system managed to mask the objects you should see the “No visible objects detected” status in the Calibration Pane. Next, click “Continue”.
  5. Perform “wanding”. By waving the calibration wand repeatedly throughout the volume, Motive will compute the position and the orientation of the markers on the wand in 3D space. To perform the “wanding” process: (Example video 04:48-08:07)
    1. Extract the calibration wand from the box. All components are depicted below in Fig. 8.
    2. Assemble the want by screwing in the handle to the center of the marker holding piece.
    3. Make sure that the markers are screwed in the “A” (500mm) position on the calibration wand. This is how the calibration wand looks when assembled. [fig. reference]
    4. In the Calibration Pane, under the “Wand the Volume” section, select the “CW-500 (500mm)” wand type.
    5. Click “Start Wanding”.
    6. Wave the volume slowly with the calibration wand while walking around the volume, allowing each camera to collect enough samples to cover most of its field of view. Perform figure-eights repetitively with the wand at varying orientations.
  6. Check the status of the wanding process per camera by looking at the status LED of each camera. when the entirety of the LED ring is green, the camera has collected enough samples. To check the status in Motive, click on “Show list” and check that all camera-representing squares are bright green. If all squares/LED rings are green, the “Start Calculating” button should appear.
  7. Click “Start Calculating”, and wait for Motive to calculate the cameras relative to each other.
  8. Make sure that the “Overall Result” is “Exceptional”. If the result is anything less than “Excellent”, it is recommended to try the calibration process again.
  9. The final step is setting the ground plane and the origin of the coordinate system. For this purpose, we will use the Calibration Square. The calibration square can be found in the plastic container for the OpiTrack system. To set the ground plane and origin: (Example video 08:08-08:50)
    1. Unfold the Calibration Square such that the legs are at a right angle.
    2. Place it in the center of the volume (or where you want the origin to be located). The longer leg represents the Z axis, while the shorter leg represents the X axis, The Y axis is automatically directed upward. TIP: Ensure the calibration wand is out of the tracking space so that it doesn’t interfere with the system’s ability to correctly detect the Calibration Square.
    3. The Calibration Square should be automatically detected in Motive as “CS-400”.
    4. Click “Set Ground Plane”. After the ground plane is set, the calibration process is complete.
Unassembled Wand (fig.8) Assembled Wand

Read more about calibration

In Motive, Rigid Body assets are used to track solid, inflexible objects. Markers are attached to these objects, and their specific placement data is utilized to recognize the object and deliver 6 Degrees of Freedom (6DoF) information.

Applying the Reflective Markers

A Rigid Body requires at least three markers to be tracked by the OptiTrack system. If possible more than three markers are recommended for better tracking accuracy. However, not too many markers should be placed in close vicinity as this might cause label-swapping. It is best to place the markers asymmetrically around the body of the object as this helps the system in distinguishing orientations. If you add multiple rigid bodies, make sure that each one has a unique marker placement, which helps the system’s Rigid Body solver identify each rigid body. In short, ensure that each Rigid Body has a unique marker configuration in terms of arrangement, marker-to-marker distances, and, ideally, marker count.

To create a Rigid Body: (Example video 46:58-47:30)

  1. Apply at least three reflective markers to the object.
  2. Select the Builder Pane in the left UI column.
  3. For asset “Type”, select “Rigid Body”.
  4. In the 3D Viewport select all the markers attached to the object by dragging over them with your pointer, or by holding the “ctrl” key + clicking each marker.
  5. Give a name to this Rigid Body in the “Name” input box.
  6. Click “Create + Refine”.
  7. Move the rigid body around the volume until the progress bar is full.
  8. Once Rigid Body is created, the markers will be colored and interconnected. The Rigid Body will be listed under the Assets Pane.

To perform a recording: (Example video 43:47-44:00)

  1. In the Control Deck, switch to “LIVE” mode.
  2. To start the recording click the Record button near the center of the Control Deck. The red border around the Viewports will indicate that the recording is in progress.
  3. Click the Stop button to end and save the recording.

Exporting Recording Data

The recordings will be stored in .TAKE files in the current session directory, which can be found in the Data Pane in the left UI column. To calculate the positions and orientations of the Rigid Bodies, the 3D data needs to be “solved”. Also, Reconstruction and Auto-Labeling are required to export the recording:

  1. To solve all assets from a take, right-click a recording file (or take) from the Data Pane, and select “Solve All Assets”.
  2. Select the “Reconstruct and Auto-Label” option in the right-click menu of a recording file. By following these steps, the tracking data should be ready for export.
  3. From the Data Pane, right-click the recording file and select “Export Tracking Data”.
  4. In the export dialogue window, select a file format and configure the export settings.

Motive can reconstruct the 3D data into multiple file formats, including C3D, FBX, BVH, and TRC. Read more...

Motion usage recording