Motive 2.0 beta is Now Available

Check out the Release notes and the 2.0 Wiki

Rigid Body Tracking

From NaturalPoint Product Documentation Ver 1.10
Jump to: navigation, search
Back to the Main PageBack to Motive DocumentationBack to Assets


In Motive, rigid body assets are created from retroreflective markers that are attached to tracked objects. From these assets, position and orientation (6 degrees of freedom) can be obtained. This page details instructions on how to create the rigid bodies and other useful features associated with the assets.


Rigid Body Marker Placement[edit]



A RigidBody in Motive is three or more markers in 3D space that are interconnected to each other with an assumption that the objects are undeformable. More specifically, it is assumed that the spatial relationship among the attacted markers remain constant, and marker-to-marker distance does not deviate beyond the allowable deflection tolerance defined under the Rigid Body Properties. Otherwise, involved markers may become unlabeled. Cover any reflective surfaces on the rigid body with non-reflective materials, and attach the markers on the exterior of the rigid body where cameras can easily capture them. In 3D space, a minimum of three coordinates are necessary for defining a plane and calculating orientations using vector relationships. Likewise, at least three markers are required to create a rigid body in Motive.

Tip: If you wish to get more accurate 3D orientation data (pitch, roll, and yaw), it is beneficial to spread markers as far as you can within the rigid body. This way, any slight deviation in the orientation will be reflected from small changes in the position.


  • Retroreflective markers placed on a quadrocopter
  • The corresponding rigid body defined in Motive

Number of Markers

Additional markers can provide more 3D coordinates for computing orientation of the rigid body, and they can also be beneficial when accommodating marker occlusions. When any of rigid body markers are occluded, Motive can reference to other visible markers to solve for the missing data and compute position and orientation of the rigid body. Whenever possible it is best to use 4+ markers to create rigid bodies. This helps for a few reasons:

  • It adds at least one redundant marker which is helpful when markers get occluded because you do not lose tracking as easily.
  • The rigid body solve benefits from having a point not on the same plane as all the other markers, because it makes the overall shape more unique.
  • It adds increased stability. Having extra markers not in the same plane also helps the stability of the rigid body solve because it give you data in the missing axis.

However, placing too many markers is not recommended. When too many markers are placed in close vicinity, individual reflections may be indistinguishable from the camera view. Furthermore, it may increase the likelihood of label-swaps during capture. Securely place a sufficient number of markers (usually less than 10) just enough to cover the main frame of the rigid body.

Tip: The recommended number of markers per a rigid body is 4 ~ 12 markers. Rigid body cannot be created with more than 20 markers in Motive.



Asymmetrical Marker Placements

Within a rigid body, a set of markers should be placed asymmetrically, because it provides a clear distinction of orientation within the asset itself. Avoid symmetrical shapes such as squares, isosceles, or equilateral triangles. Symmetrical arrangements make asset identification difficult, and they may cause the rigid body assets to flip during capture.


Unique Marker Placements

Noncongruent, or unique, five markers placements in 2D

When tracking multiple objects, it is beneficial to create unique rigid body assets in Motive. Specifically, you need to place retroreflective markers in a distinctive arrangement for each object, and it will allow Motive to more clearly identify the markers on each rigid body throughout capture. In other words, their unique, non-congruent, arrangements work as distinctive identification flags among multiple assets in Motive. This not only reduces amount of required computation for the rigid body solver, but it also improves the stability of the tracking. Not having unique rigid bodies could lead to labeling errors especially when tracking several assets with similar size and shape.


What Makes Rigid Bodies Unique?

The key idea of creating unique rigid body is to avoid geometrical congruency within multiple rigid bodies in Motive.

  • Unique Marker Arrangement. Each rigid body must have a unique, non-congruent, marker placement creating a unique shape when the markers are interconnected.
  • Unique Marker-to-Marker Distances. When tracking several objects, introducing unique shapes could be difficult. Another solution is to vary Marker-to-marker distances. This will create similar shapes with varying sizes, and make them distinctive from the others.
  • Unique Marker Counts Adding extra markers is another method of introducing the uniqueness. Extra markers will not only make the rigid bodies more distinctive, but they will also provide more options for varying the arrangements to avoid the congruency.


What Happens When Rigid Bodies Are Not Unique?

The Unique setting set to true for a rigid body under its properties from the Project pane.

Having multiple non-unique rigid bodies may lead to mislabeling errors. However, in Motive, non-unique rigid bodies can also be tracked fairly well as long as the non-unique rigid bodies are continuously tracked throughout capture. Motive can refer to the trajectory history to identify and associate corresponding rigid bodies within different frames. In order to track non-unique rigid bodies, you must make sure the Properties → General Settings → Unique setting in Rigid Body Properties of the assets are set to False.


Even though it is possible to track non-unique rigid bodies, it is strongly recommended to make each asset unique. Tracking of multiple congruent rigid bodies could be lost during capture either by occlusion or by stepping outside of the capture volume. Also, when two non-unique rigid bodies are positioned in vicinity and overlap in the scene, their definitions may be swapped. In these cases, additional efforts will be required for correcting the labels in post-processing of the data.


Multiple Rigid Bodies Tracking

Depending on the object, there could be limitations on marker placements and number of variations of unique placements that could be achieved. The following list provides sample methods for varying unique arrangements when tracking multiple rigid bodies.

1. Create Distinctive 2D Arrangements. Create distinctive, non-congruent, marker arrangements as the starting point for producing multiple variations, as shown in the examples above.

2. Vary heights. Use marker bases or posts, with different heights to introduce variations in elevation to create additional unique arrangments.

3. Vary Maximum Marker to Marker Distance. Increase or decrease the overall size of the marker arrangements.

4. Add Two (or more) Markers Lastly, if an additional variation is needed, add extra markers to introduce the uniqueness. We recommended adding at least two extra markers in case any of them is occluded.


Creating Rigid Body[edit]



Creating a rigid body from Project Pane. Properties for the selected asset is described at the bottom.

A set of markers attached to a rigid object can be grouped and auto-labeled as a rigid body. This rigid body definition can be utilized in multiple takes to continuously auto-label the same rigid body for every Take. Motive recognizes the unique spatial relationship in the marker arrangement, and auto-labels each marker to recreate the rigid body. At least three coordinates are required to define a plane in 3D space, and therefore, a minimum of three markers are essential for creating a rigid body.


Steps to Create Rigid Bodies:

  1. Select all associated rigid body markers in the 3D view pane
  2. Right-click on the perspective view to access the context menu.
  3. Go to Create From Selected Markers under Rigid Body. This can also be done by using the Create Rigid Body Assets RBIcon.png button in Project Pane, or by using the Hotkey (Default: Ctrl +T) while the markers are selected.
  4. When a rigid body is created, the markers will be colored and interconnected to each other. A newly created rigid body will be displayed under Assets in the Project pane.


If the rigid bodies, or skeletons, are created in the Edit mode, the corresponding Take needs to be auto-labeled again for newly applied assets to be effective. Only then, new annotations to the markers will be applied to the 3D data.


Asset RBRightClick.png Asset RBbat.png
Creating a rigid body from the perspective view
Adding an extra marker to a existing rigid body asset.

Rigid Body Properties

When a rigid body is selected, rigid body properties will show up at the bottom of project pane. Here, you can access different settings, such as properties, display, transformation, and real-time info, for the selected rigid body. Default settings for the newly created rigid body are set from the Rigid Bodies tab under Application Settings.


Adding or Removing Markers

An existing rigid body can be modified by adding or removing markers.

  1. First select a rigid body from the Project Pane or by selecting the pivot point in the Perspective View.
  2. Ctrl + left-click the markers that you wish to add/remove.
  3. Left-click on the Perspective View pane to open the rigid body context menu.
  4. Under Rigid Body, choose Add/Remove selected markers to/from rigid body.
  5. If needed, right-click on the rigid body and select Reset Pivot to relocate the pivot point to the new center.



Tracking Rigid Body[edit]



Rigid body real-time information. The position is represented in respect to the global origin, and the orientation is represented in respect to the initial orientation when the rigid body was created.
Displaying rigid body pivot points and orientations

The pivot point of a rigid body is used to define both position and orientation. When a rigid body is created, its pivot point is be placed at the center, and its orientation axis will be aligned along the global coordinate axis, which was defined during the Calibration process. To view the pivot point and the orientation in the 3D viewport, set the Pivot and Orientation to true in the display tab of the selected rigid body in the Project Pane.

Real-time Information

Position and orientation of a tracked rigid body can be monitored in real-time. First, select a rigid body from the Project pane, and in the properties section at the bottom, access the real-time information tab where the tracking information of selected rigid body is displayed.


Orientation

As mentioned previously, the orientation axis for a rigid body is aligned along the global coordinate axis when a rigid body is created. To reset the orientation, first physically orient the tracked object in a desired angle in respect to the global system, select the corresponding rigid body in Motive, and press Reset To Current Orientation under the transformation tab in the Project Pane. Then the coordinate axis of the rigid body will be zeroed to the orientation. Precise adjustments on the pitch, roll, and yaw can also be applied from this section.


Adjusting Pivot Point

There are situations where the desired pivot point location is not at the center of a rigid body. The location of a pivot point can be adjusted by assigning it to a marker or by translating along the rigid body axis (x,y,z). For most accurate pivot point location, attach a marker on the desired pivot location, set the pivot point to the marker, and apply the translation for precise adjustments. If you are adjusting the pivot point after the capture, in the Edit mode, the Take will need to be auto-labeled again to apply the changes.


Setting Pivot Point to a Marker

To assign the pivot point to a marker, first select the pivot point in the Perspective View pane, and CTRL select the marker that you wish to assign to. Then right-click to open the context menu, and in the rigid body section, click Set Pivot Point to Selected Marker.

Pivot Assign1.png Pivot Assign2.png


Translating Pivot Point

To translate the pivot point, access the Transformation tab in the Project pane while the rigid body is selected. Then, on the Rigid Body Pivot-Point Translation section, insert the amount of translation (in mm) that you wish to apply. Note that the translation will be applied along the x/y/z of the rigid body orientation axis.

Pivot Translate1.png Pivot Translate2.png


Resetting the Pivot Point

If you wish to reset the pivot point, simply open the rigid body context menu in the Perspective pane and click Reset Pivot. The location of the pivot point will be reset back to the center of the rigid body again.

Pivot Reset.png


Pivot for Spherical Placement

This feature is useful when tracking a spherical object (e.g. ball). The Update Pivot for Spherical Placement feature will assume that all the rigid body markers are placed on the surface of a spherical object, and the pivot point will be calculated and re-positioned accordingly. To do this, select a rigid body, right-click to access the context menu, and click Update Pivot for Spherical Placement.



Export Assets Definition[edit]



Assets in a Take can be exported into Motive compatible rigid body (TRA) and skeleton (SKL) definition files. These files record calibrated spatial relationship of markers in each asset, and they are used to import the asset into different takes without creating a new one in Motive. Note that these files contain the spatial relationship and only the identical marker arrangements will be recognized and defined with the imported asset.

Rigid body definitions are exported as TRA files from the Export Rigid Body option under the File tab. When there are multiple rigid bodies in a Take, individual definitions can be exported by selecting a rigid body and clicking the Export Rigid Body from the context menu in the Project Pane. Likewise, skeleton definitions are exported as SKL files from the Export Skeleton option under the File tab, and individual definitions are exported through the Export Skeleton option located in the context menu from the Project Pane.

TIP: Rigid body and Skeleton definitions can also be quickly copied into different takes by dragging and dropping selected assets into a Take from the Project Pane.