Inheritance diagram for gtsam::LocationRecovery:

Public Member Functions
	LocationRecovery (const LevenbergMarquardtParams &lmParams)
	Construct with LM parameters.
	LocationRecovery ()=default
	Default constructor.
NonlinearFactorGraph	buildGraph (const DirectionEdges &edges, bool bilinear=true) const
	Build factor graph from direction measurements.
void	addAnchorPrior (Key anchorKey, NonlinearFactorGraph *graph, const SharedNoiseModel &priorNoiseModel=noiseModel::Isotropic::Sigma(3, 0.01)) const
	Add a prior pinning one key to the origin.
Values	initializeRandomly (const std::set< Key > &keys, size_t numEdges, bool bilinear, std::mt19937 *rng, const Values &initialValues=Values()) const
	Random initialization of Point3 keys and BATA scale variables.
Values	initializeRandomly (const std::set< Key > &keys, size_t numEdges, bool bilinear, const Values &initialValues=Values()) const
	Version of initializeRandomly with a fixed seed.

Public Types
using	DirectionEdges = std::vector<BinaryMeasurement<Unit3>>

Static Protected Member Functions
static SharedNoiseModel	convertNoiseModel (const SharedNoiseModel &unit3NoiseModel)
	Convert Unit3 (2D manifold) noise to Point3 (3D ambient) noise.

Protected Attributes
LevenbergMarquardtParams	lmParams_

Member Function Documentation

◆ addAnchorPrior()

void LocationRecovery::addAnchorPrior	(	Key	anchorKey,
		NonlinearFactorGraph *	graph,
		const SharedNoiseModel &	priorNoiseModel = noiseModel::Isotropic::Sigma(3, 0.01) ) const

Add a prior pinning one key to the origin.

Fixes 3 translation DOF. Does not assume the key is a camera or landmark — it just pins a Point3.

NonlinearFactorGraph LocationRecovery::buildGraph	(	const DirectionEdges &	edges,
		bool	bilinear = true ) const

Build factor graph from direction measurements.

If bilinear is true, uses BilinearAngleTranslationFactor (BATA) with per-observation scale variables Symbol('S', i). Otherwise uses the chordal TranslationFactor.

Parameters

edges	Unit3 direction measurements between Point3 unknowns.
bilinear	if true, use BATA factors (default: true).

Values LocationRecovery::initializeRandomly	(	const std::set< Key > &	keys,
		size_t	numEdges,
		bool	bilinear,
		std::mt19937 *	rng,
		const Values &	initialValues = Values() ) const

Random initialization of Point3 keys and BATA scale variables.

Point3 keys are initialized uniformly in [-1, 1]^3. If bilinear is true, also creates numEdges scale variables initialized to 1.0.

Parameters

keys	Point3 keys to initialize.
numEdges	number of edges (for scale variable count).
bilinear	whether to create BATA scale variables.
rng	random number generator.
initialValues	optional seed values for specific keys.

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