`mrpt.tfest`

mrpt-tfest Python API

Package Contents

class mrpt.tfest.TMatchingPair

class mrpt.tfest.TMatchingPair(globalIdx: int, localIdx: int, global_pt: mrpt.math.TPoint3Df, local_pt: mrpt.math.TPoint3Df)

class mrpt.tfest.TMatchingPair(globalIdx: int, localIdx: int, global_x: float, global_y: float, global_z: float, local_x: float, local_y: float, local_z: float)

A pair of corresponding points (float coordinates) in two point sets.

errorSquareAfterTransformation: float

globalIdx: int

global_pt: mrpt.math.TPoint3Df

localIdx: int

local_pt: mrpt.math.TPoint3Df

class mrpt.tfest.TMatchingPairList

A list of TMatchingPair.

push_back(arg0: TMatchingPair) → None: Appends a pair.

size() → int: Returns the number of pairs.

class mrpt.tfest.TMatchingPairList_d

A list of TMatchingPair_d.

push_back(arg0: TMatchingPair_d) → None: Appends a pair.

size() → int: Returns the number of pairs.

class mrpt.tfest.TMatchingPair_d

class mrpt.tfest.TMatchingPair_d(globalIdx: int, localIdx: int, global_pt: mrpt.math.TPoint3D, local_pt: mrpt.math.TPoint3D)

class mrpt.tfest.TMatchingPair_d(globalIdx: int, localIdx: int, global_x: float, global_y: float, global_z: float, local_x: float, local_y: float, local_z: float)

A pair of corresponding points (double coordinates) in two point sets.

errorSquareAfterTransformation: float

globalIdx: int

global_pt: mrpt.math.TPoint3D

localIdx: int

local_pt: mrpt.math.TPoint3D

class mrpt.tfest.TPotentialMatch

C++ API: struct mrpt::tfest::TPotentialMatch

A potential pairing of two points, passed to user_individual_compat_callback.

property idx_other: int: Index of the point in the other set

property idx_this: int: Index of the point in this set

class mrpt.tfest.TSE2RobustParams

C++ API: struct mrpt::tfest::TSE2RobustParams

Parameters of se2_l2_robust().

property max_rmse_to_end: float: Stops when a solution has an RMSE below this value. 0 (default) means 2*normalizationStd.

property probability_find_good_model: float: Probability of finding a good model, for the adaptive number of iterations (default: 0.999).

property ransac_algorithmForLandmarks: bool: Use the Mahalanobis distance (true, default) or the Euclidean distance (false).

property ransac_fuseByCorrsMatch: bool: If true (default), modes are fused when their subsets of pairs match; otherwise, when their means are closer than ransac_fuseMaxDiffXY and ransac_fuseMaxDiffPhi.

property ransac_fuseMaxDiffPhi: float: Maximum difference in phi to fuse two modes, in radians (default: 0.1 deg).

property ransac_fuseMaxDiffXY: float: Maximum difference in x and y to fuse two modes (default: 0.01).

property ransac_mahalanobisDistanceThreshold: float: Threshold to accept a pair into a consensus set (default: 3.0).

property ransac_maxSetSize: int: Number of pairs tested for consensus in each iteration (default: 20). Set it to the number of correspondences to test all of them.

property ransac_minSetSize: int: Minimum size of a consensus set (default: 3).

property ransac_min_nSimulations: int: Minimum number of iterations with the adaptive method (default: 1500).

property ransac_nSimulations: int: Number of iterations, or 0 (default) for an adaptive number, from probability_find_good_model.

property user_individual_compat_callback: Callable[[TPotentialMatch], bool]: Optional function(TPotentialMatch) -> bool: return True to accept a candidate pairing, False to reject it (None: accept all)

property verbose: bool: Prints debug information.

class mrpt.tfest.TSE2RobustResult

C++ API: struct mrpt::tfest::TSE2RobustResult

Result of se2_l2_robust().

property largestSubSet: TMatchingPairList: The consensus subset of pairs with the lowest RMSE.

property ransac_iters: int: Number of iterations executed.

property transformation: mrpt.poses.CPosePDFSOG: The estimated transformations, as a sum of Gaussians (mrpt.poses.CPosePDFSOG).

class mrpt.tfest.TSE3RobustParams

C++ API: struct mrpt::tfest::TSE3RobustParams

Parameters of se3_l2_robust().

property forceScaleToUnity: bool: Estimates a rigid transformation, with scale 1 (default: true).

property ransac_threshold_ang: float: Maximum rotation difference for a solution to match a candidate, in radians (default: 1 deg).

property ransac_threshold_scale: float: Maximum scale difference for a solution to match a candidate (default: 0.03).

property user_individual_compat_callback: Callable[[TPotentialMatch], bool]: Optional function(TPotentialMatch) -> bool: return True to accept a candidate pairing, False to reject it (None: accept all)

property verbose: bool: Prints debug information.

ransac_maxSetSizePct: float

ransac_minSetSize: int

ransac_nmaxSimulations: int

ransac_threshold_lin: float

class mrpt.tfest.TSE3RobustResult

C++ API: struct mrpt::tfest::TSE3RobustResult

Result of se3_l2_robust().

property transformation: mrpt.poses.CPose3D: Estimated SE(3) transform as CPose3D (converted from internal CPose3DQuat).

inliers_idx: list[int]

scale: float

mrpt.tfest.se2_l2(arg0: TMatchingPairList) → tuple: Least-squares SE(2) estimation. Returns (ok, mrpt::math::TPose2D)

mrpt.tfest.se2_l2_robust(in_correspondences: TMatchingPairList, in_normalizationStd: float, in_ransac_params: TSE2RobustParams) → tuple: Robust SE(2) estimation using RANSAC, returning a sum of Gaussians. in_normalizationStd is the standard deviation of the point coordinates. Returns (ok, TSE2RobustResult).

mrpt.tfest.se2_l2_with_covariance(correspondences: TMatchingPairList) → tuple: Least-squares SE(2) estimation, also returning its normalized 3x3 covariance, which must be multiplied by the variance of the point coordinates. Returns (ok, mrpt.math.TPose2D, mrpt.math.CMatrixDouble33). ok is false for less than two pairs, or if the rotation is undefined.

mrpt.tfest.se3_l2(in_correspondences: TMatchingPairList, forceScaleToUnity: bool = False) → tuple
mrpt.tfest.se3_l2(in_correspondences: TMatchingPairList_d, forceScaleToUnity: bool = False) → tuple
mrpt.tfest.se3_l2(in_points_this: list[mrpt.math.TPoint3D], in_points_other: list[mrpt.math.TPoint3D], forceScaleToUnity: bool = False) → tuple: Least-squares SE(3) estimation (Horn’s method) from two lists of corresponding points of the same length. Returns (ok, mrpt.poses.CPose3D, scale).

mrpt.tfest.se3_l2_robust(arg0: TMatchingPairList, arg1: TSE3RobustParams) → tuple: Robust SE(3) estimation using RANSAC. Returns (ok, TSE3RobustResult)

mrpt.tfest

Package Contents

`mrpt.tfest`