The lab contains a motion capture system consisting of six Motion Analysis Hawk Digital Cameras and EVaRT motion capture software.

Motion Capture Guidelines

Getting started

  1. Turn on the cameras by flipping the power switch on the camera hub
  2. Open EVaRT and choose the Setup tab
  3. Choose "connect to cameras" on the bottom left. The software should detect six cameras are connected.
  4. Press the "Run" button (bottom right) to enable the cameras.
  5. You can select individual cameras by clicking on the buttons 1 through 6 on the bottom of the screen. When a camera is selected, you can verify that it has been assigned an IP address. To see what a particular camera is sensing, switch to that camera and open the 2D view from the View menu. To see what all cameras are viewing, click the "show all cameras" button.
  6. Save the project.

Adjusting the cameras to cover the volume

  1. Choose the Calibration tab.
  2. Click the "Details" button on the right pane and set options as appropriate (measurement units, etc.). Here, you can set the dimensions of the volume to be captured here; this will help you adjust the cameras to better cover the volume.
  3. Once the volume is set, open the 3D view from the View menu. Right-click on the main display and select "show volume" and "camera FOV".
  4. Adjust the poses of the cameras so that their FOV intersects the volume.

Masking spurious light sources

  1. Remove any markers or light sources from the capture volume.
  2. Press the "Run" button (bottom right) to enable the cameras)
  3. In EVaRT, open the 2D view for each camera. On the top right in parentheses is the number of points that the camera is currently detecting. Ideally, if the volume is clear, there should be no points detected. Try to identify the sources of any spurious points (e.g. light from the windows, reflections from other equipment, etc.). If the source cannot be covered or moved, you can define a mask by middle-clicking and dragging around the source in the 2D view. The software will ignore any motion in this area for the selected camera.
  4. Once the masks are defined, save the project.

Calibration

Calibration involves two steps; orienting the cameras using the L-frame, and then adjusting their positions by waving the wand to generate many sample points.

  1. Place the L-frame in the center of the volume, aligned with the volume axes.
  2. Press the "Run" button (bottom right)
  3. Check the box to allow overwriting the previous L-frame calibration
  4. Press "Collect and calibrate" for the L-frame calibration
  5. Remove the L-frame.
  6. Check the box to allow overwriting the previous wand calibration
  7. Press "Collect and calibrate" for the wand calibration
  8. While the software is recording, wave the wand to try and cover the volume as much as possible.
  9. When the recording stops, check "heavily weighted seed" and press "Run"
  10. Uncheck "heavily weighted seed" and run several more times, until the residual stabilizes
  11. Choosing "save as system calibration" will use same calibration for future sessions
  12. Save project

Creating a marker set

The first step in capturing motion is to define the marker set. This involves taking a short recording of the subject in a neutral pose so that the markers can be identified and labeled.

  1. Choose the "Motion Capture" tab
  2. On the right pane, choose desired output formats; avi/anl records the raw video, while trb/trc recordings the tracking data in binary or ascii format
  3. Type in a name for the capture
  4. Place the subject in a neutral pose and record a brief capture (e.g. 5 - 10 seconds)
  5. Choose the "Post Process" tab
  6. Load the track file (trb/trc) that was just recorded
  7. Create a marker set:
    1. Go to "Model Edit"
    2. On markers tab on right, enter names for the markers
    3. Choose "Quick ID"
    4. For each marker name, identify the corresponding point by clicking on it
    5. Click the "Linkages" button on the right pane. Create linkages between neighboring markers by left-clicking and drawing lines between them
    6. Click "Create template" to save the marker configuration

Motion Capture

  1. Select the Motion Capture tab
  2. Verify the appropriate output formats and type in a name for the capture
  3. Use the record button to capture the motion.

Post Processing

  1. Click the "Post Process" tab
  2. Load the track file (trb/trc)
  3. Select the initial frame and press "Template ID"; this attempts to label the markers based on the previously defined template (it is helpful if the subject starts in a neutral pose)
  4. Select all frames (buttons on bottom right)
  5. Click template rectify; EVaRT will attempt to match the template to all frames. You can scrub through the frames to check that the markers are labeled consistently.
  6. Bring up the XYZ view from the View menu. This shows the XYZ coordinates for a particular marker. You may notice several places where the marker disappears or spikes appear in the motion
  7. If marker labels are erroneously swapped, select the markers in the frame where the swap occurs and press "exchange"
  8. To interpolate gaps in the tracking
    1. Press 'N' to jump to the next gap
    2. Select an interval from a few frames before the gap to a few frames after by using the middle mouse button
    3. Click "Join Cubic" to perform cubic interpolation, or "Join Virtual" to interpolate based on the position of the neighboring markers (good for rigid structures)
  9. To remove spikes that appear, select the frame(s) and delete the motion with 'X'. Then, use interpolation as above to fill in the hole