Public Member Functions
bool	operator() (CameraParameters const &camParams, LedPtrList const &leds, BeaconStateVec const &beacons, std::vector< BeaconData > &beaconDebug, Eigen::Vector3d &outXlate, Eigen::Quaterniond &outQuat, int skipBrightsCutoff=-1, std::size_t iterations=5)
	Perform RANSAC-based pose estimation. More...

bool	operator() (EstimatorInOutParams const &p, LedPtrList const &leds)
	Perform RANSAC-based pose estimation and use it to update a body state (state vector and error covariance) More...

Member Function Documentation

§ operator()() [1/2]

bool osvr::vbtracker::RANSACPoseEstimator::operator()	(	CameraParameters const &	camParams,
		LedPtrList const &	leds,
		BeaconStateVec const &	beacons,
		std::vector< BeaconData > &	beaconDebug,
		Eigen::Vector3d &	outXlate,
		Eigen::Quaterniond &	outQuat,
		int	skipBrightsCutoff = `-1`,
		std::size_t	iterations = `5`
	)

Perform RANSAC-based pose estimation.

Parameters

[out]	outXlate	translation output parameter
[out]	outQuat	rotation output parameter
	skipBrightsCutoff	If positive, the number of non-bright LEDs seen that will trigger us to skip using bright LEDs in pose estimation.

Effectively invert the image points here so we get the output of a coordinate system we want.

Make a vector of the inlier beacon IDs.

Now, we will sort that vector of inlier beacon IDs so we can rapidly binary search it to flag the LEDs we used.

Get to Eigen via OpenCV's Affine transform class.

bool osvr::vbtracker::RANSACPoseEstimator::operator()	(	EstimatorInOutParams const &	p,
		LedPtrList const &	leds
	)

Perform RANSAC-based pose estimation and use it to update a body state (state vector and error covariance)

Parameters

[out] state Tracked body state that will be updated if a pose was estimated

Call the main pose estimation to get the vector and quat.

OK, so if we're here, estimation succeeded and we have valid data in xlate and quat.

Zero things we can't measure.

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