The VioFrontendInterface class is an interface for frontends. More...

Inheritance diagram for okvis::VioFrontendInterface:

Public Member Functions
	VioFrontendInterface ()

virtual	~VioFrontendInterface ()


In the derived class, the following methods (and nothing else) have to be implemented:
virtual bool	detectAndDescribe (size_t cameraIndex, std::shared_ptr< okvis::MultiFrame > frameOut, const okvis::kinematics::Transformation &T_WC, const std::vector< cv::KeyPoint > *keypoints)=0
	Detection and descriptor extraction on a per image basis. More...

virtual bool	dataAssociationAndInitialization (okvis::Estimator &estimator, okvis::kinematics::Transformation &T_WS_propagated, const okvis::VioParameters &params, const std::shared_ptr< okvis::MapPointVector > map, std::shared_ptr< okvis::MultiFrame > framesInOut, bool *asKeyframe)=0
	Matching as well as initialization of landmarks and state. More...

virtual bool	propagation (const okvis::ImuMeasurementDeque &imuMeasurements, const okvis::ImuParameters &imuParams, okvis::kinematics::Transformation &T_WS_propagated, okvis::SpeedAndBias &speedAndBiases, const okvis::Time &t_start, const okvis::Time &t_end, Eigen::Matrix< double, 15, 15 > covariance, Eigen::Matrix< double, 15, 15 > jacobian) const =0
	Propagates pose, speeds and biases with given IMU measurements. More...

Detailed Description

The VioFrontendInterface class is an interface for frontends.

Constructor & Destructor Documentation

okvis::VioFrontendInterface::VioFrontendInterface ( )

inline

virtual okvis::VioFrontendInterface::~VioFrontendInterface ( )

inlinevirtual

virtual bool okvis::VioFrontendInterface::dataAssociationAndInitialization	(	okvis::Estimator &	estimator,
		okvis::kinematics::Transformation &	T_WS_propagated,
		const okvis::VioParameters &	params,
		const std::shared_ptr< okvis::MapPointVector >	map,
		std::shared_ptr< okvis::MultiFrame >	framesInOut,
		bool *	asKeyframe
	)

pure virtual

Matching as well as initialization of landmarks and state.

Parameters

	estimator	Estimator.
	T_WS_propagated	Pose of sensor at image capture time.
	params	Configuration parameters.
	map	Current map.
	framesInOut	Multiframe including the descriptors of all the keypoints.
[out]	asKeyframe	Should the frame be a keyframe?

Implemented in okvis::Frontend.

virtual bool okvis::VioFrontendInterface::detectAndDescribe	(	size_t	cameraIndex,
		std::shared_ptr< okvis::MultiFrame >	frameOut,
		const okvis::kinematics::Transformation &	T_WC,
		const std::vector< cv::KeyPoint > *	keypoints
	)

pure virtual

Detection and descriptor extraction on a per image basis.

Parameters

	cameraIndex	Index of camera to do detection and description.
	frameOut	Multiframe containing the frames. Resulting keypoints and descriptors are saved in here.
	T_WC	Pose of camera with index cameraIndex at image capture time.
[in]	keypoints	If the keypoints are already available from a different source, provide them here in order to skip detection.

Implemented in okvis::Frontend.

virtual bool okvis::VioFrontendInterface::propagation	(	const okvis::ImuMeasurementDeque &	imuMeasurements,
		const okvis::ImuParameters &	imuParams,
		okvis::kinematics::Transformation &	T_WS_propagated,
		okvis::SpeedAndBias &	speedAndBiases,
		const okvis::Time &	t_start,
		const okvis::Time &	t_end,
		Eigen::Matrix< double, 15, 15 > *	covariance,
		Eigen::Matrix< double, 15, 15 > *	jacobian
	)		const

pure virtual

Propagates pose, speeds and biases with given IMU measurements.

Parameters

[in]	imuMeasurements	All the IMU measurements.
[in]	imuParams	The parameters to be used.
[in,out]	T_WS_propagated	Start pose.
[in,out]	speedAndBiases	Start speed and biases.
[in]	t_start	Start time.
[in]	t_end	End time.
[out]	covariance	Covariance for GIVEN start states.
[out]	jacobian	Jacobian w.r.t. start states.

Implemented in okvis::Frontend.

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