osvr::vbtracker::TrackingSystem Class Reference

Classes
struct	Impl
	Private implementation structure for TrackingSystem. More...

Public Member Functions
ConfigParams const &	getParams () const
void	setUseIMU (bool useIMU)
bool	haveCameraPose () const
void	setCameraPose (Eigen::Isometry3d const &camPose)
Eigen::Isometry3d const &	getCameraPose () const
	This gets rTc - the pose of the camera in the room.
Eigen::Isometry3d const &	getRoomToCamera () const
	This gets cTr - the inverse of the camera pose, transforms from the room coordinate system to the camera coordinate system. More...
bool	isRoomCalibrationComplete ()
void	calibrationHandleIMUData (BodyId id, util::time::TimeValue const &tv, Eigen::Quaterniond const &quat)
	Called by TrackedBody::incorporateNewMeasurementFromIMU() if room calibration is not complete. More...
Setup and Teardown
	TrackingSystem (ConfigParams const &params)
	~TrackingSystem ()
TrackedBody *	createTrackedBody ()
Runtime methods
Perform the initial phase of image processing. This does not modify the bodies, so it can happen in parallel/background processing. It's also the most expensive, so that's handy.
ImageOutputDataPtr	performInitialImageProcessing (util::time::TimeValue const &tv, cv::Mat const &frame, cv::Mat const &frameGray, CameraParameters const &camParams)
LedUpdateCount const &	updateLedsFromVideoData (ImageOutputDataPtr &&imageData)
	This is the second phase of the video-based tracking algorithm - the part that actually changes LED state. More...
BodyIndices const &	updateBodiesFromVideoData (ImageOutputDataPtr &&imageData)
	The combined second and third phases of the video-based tracking algorithm. More...
BodyIndices const &	processFrame (util::time::TimeValue const &tv, cv::Mat const &frame, cv::Mat const &frameGray, CameraParameters const &camParams)
	All parts of the tracking algorithm combined for convenience. More...
Accessors
std::size_t	getNumBodies () const
bool	isValidBodyId (BodyId i) const
TrackedBody &	getBody (BodyId i)
TrackedBody const &	getBody (BodyId i) const
TrackedBodyTarget *	getTarget (BodyTargetId target)
TrackedBodyTarget const *	getTarget (BodyTargetId target) const

Friends
class	TrackingDebugDisplay

Member Function Documentation

calibrationHandleIMUData()

void osvr::vbtracker::TrackingSystem::calibrationHandleIMUData	(	BodyId	id,
		util::time::TimeValue const &	tv,
		Eigen::Quaterniond const &	quat
	)

Called by TrackedBody::incorporateNewMeasurementFromIMU() if room calibration is not complete.

getParams()

ConfigParams const& osvr::vbtracker::TrackingSystem::getParams ( ) const

inline

Todo:

refactor;

getRoomToCamera()

Eigen::Isometry3d const & osvr::vbtracker::TrackingSystem::getRoomToCamera

(

)

const

This gets cTr - the inverse of the camera pose, transforms from the room coordinate system to the camera coordinate system.

isRoomCalibrationComplete()

bool osvr::vbtracker::TrackingSystem::isRoomCalibrationComplete

(

)

Todo:

should just be able to do this by checking the state of the calibrator.

If this is true, we know it's true. If it's false, we must go check.

Update the cached value.

processFrame()

BodyIndices const& osvr::vbtracker::TrackingSystem::processFrame ( util::time::TimeValue const &  tv, cv::Mat const &  frame, cv::Mat const &  frameGray, CameraParameters const &  camParams  )

inline

All parts of the tracking algorithm combined for convenience.

Returns

A reference to a vector of body indices that were updated with this latest frame.

setUseIMU()

void osvr::vbtracker::TrackingSystem::setUseIMU ( bool  useIMU )

inline

Todo:

just for debugging

updateBodiesFromVideoData()

BodyIndices const & osvr::vbtracker::TrackingSystem::updateBodiesFromVideoData

(

ImageOutputDataPtr &&

imageData

)

The combined second and third phases of the video-based tracking algorithm.

The third phase uses the updated LED data stored in each target to arrive at updated pose estimates.

These two phases are combined in one call to ensure preconditions - it would be invalid to call the third phase without immediately previously calling this second phase, but no state handover needs to take place.

Parameters

imageData

Output from the first step - please std::move() the output of the first step into this step.

Returns

A reference to a vector of body indices that were updated with this latest frame.

Do the second phase of stuff

Do the third phase of tracking.

Trigger debug display, if activated.

updateLedsFromVideoData()

LedUpdateCount const & osvr::vbtracker::TrackingSystem::updateLedsFromVideoData

(

ImageOutputDataPtr &&

imageData

)

This is the second phase of the video-based tracking algorithm - the part that actually changes LED state.

Parameters

imageData

Output from the first step - please std::move() the output of the first step into this step.

If you don't require updated poses, you can stop after this step, not proceeding to the third and final phase, and still keep track of which beacons are which.

Returns

a reference to an internal map of body IDs to counts of used LED measurements for debugging.

Clear internal data, we're invalidating things here.

Update our frame cache, since we're taking ownership of the image data now.

Go through each target and try to process the measurements.

---

The documentation for this class was generated from the following files:

• plugins/unifiedvideoinertialtracker/TrackingSystem.h
• plugins/unifiedvideoinertialtracker/TrackingSystem.cpp