Open3D (C++ API)  0.18.0+80ae047
Data Structures | Enumerations | Functions
open3d::t::pipelines::registration Namespace Reference

Data Structures

class  ICPConvergenceCriteria
 Class that defines the convergence criteria of ICP. More...
 
class  RegistrationResult
 
class  RobustKernel
 
class  TransformationEstimation
 
class  TransformationEstimationPointToPoint
 
class  TransformationEstimationPointToPlane
 
class  TransformationEstimationForColoredICP
 
class  TransformationEstimationForDopplerICP
 

Enumerations

enum class  RobustKernelMethod {
  L2Loss = 0 , L1Loss = 1 , HuberLoss = 2 , CauchyLoss = 3 ,
  GMLoss = 4 , TukeyLoss = 5 , GeneralizedLoss = 6
}
 
enum class  TransformationEstimationType {
  Unspecified = 0 , PointToPoint = 1 , PointToPlane = 2 , ColoredICP = 3 ,
  DopplerICP = 4
}
 

Functions

core::Tensor ComputeFPFHFeature (const geometry::PointCloud &input, const utility::optional< int > max_nn, const utility::optional< double > radius)
 
core::Tensor CorrespondencesFromFeatures (const core::Tensor &source_features, const core::Tensor &target_features, bool mutual_filter=false, float mutual_consistency_ratio=0.1)
 Function to find correspondences via 1-nearest neighbor feature matching. Target is used to construct a nearest neighbor search object, in order to query source. More...
 
RegistrationResult EvaluateRegistration (const geometry::PointCloud &source, const geometry::PointCloud &target, double max_correspondence_distance, const core::Tensor &transformation=core::Tensor::Eye(4, core::Float64, core::Device("CPU:0")))
 Function for evaluating registration between point clouds. More...
 
RegistrationResult ICP (const geometry::PointCloud &source, const geometry::PointCloud &target, const double max_correspondence_distance, const core::Tensor &init_source_to_target=core::Tensor::Eye(4, core::Float64, core::Device("CPU:0")), const TransformationEstimation &estimation=TransformationEstimationPointToPoint(), const ICPConvergenceCriteria &criteria=ICPConvergenceCriteria(), const double voxel_size=-1.0, const std::function< void(const std::unordered_map< std::string, core::Tensor > &)> &callback_after_iteration=nullptr)
 Functions for ICP registration. More...
 
RegistrationResult MultiScaleICP (const geometry::PointCloud &source, const geometry::PointCloud &target, const std::vector< double > &voxel_sizes, const std::vector< ICPConvergenceCriteria > &criteria_list, const std::vector< double > &max_correspondence_distances, const core::Tensor &init_source_to_target=core::Tensor::Eye(4, core::Float64, core::Device("CPU:0")), const TransformationEstimation &estimation=TransformationEstimationPointToPoint(), const std::function< void(const std::unordered_map< std::string, core::Tensor > &)> &callback_after_iteration=nullptr)
 Functions for Multi-Scale ICP registration. It will run ICP on different voxel level, from coarse to dense. The vector of ICPConvergenceCriteria(relative fitness, relative rmse, max_iterations) contains the stopping condition for each voxel level. The length of voxel_sizes vector, criteria vector, max_correspondence_distances vector must be same, and voxel_sizes must contain positive values in strictly decreasing order [Lower the voxel size, higher is the resolution]. Only the last value of the voxel_sizes vector can be {-1}, as it allows to run on the original scale without downsampling. More...
 
core::Tensor GetInformationMatrix (const geometry::PointCloud &source, const geometry::PointCloud &target, const double max_correspondence_distance, const core::Tensor &transformation)
 Computes Information Matrix, from the transfromation between source and target pointcloud. It returns the Information Matrix of shape {6, 6}, of dtype Float64 on device CPU:0. More...
 

Enumeration Type Documentation

◆ RobustKernelMethod

Enumerator
L2Loss 
L1Loss 
HuberLoss 
CauchyLoss 
GMLoss 
TukeyLoss 
GeneralizedLoss 

◆ TransformationEstimationType

Enumerator
Unspecified 
PointToPoint 
PointToPlane 
ColoredICP 
DopplerICP 

Function Documentation

◆ ComputeFPFHFeature()

core::Tensor open3d::t::pipelines::registration::ComputeFPFHFeature ( const geometry::PointCloud input,
const utility::optional< int >  max_nn = 100,
const utility::optional< double >  radius = utility::nullopt 
)

Function to compute FPFH feature for a point cloud. It uses KNN search (Not recommended to use on GPU) if only max_nn parameter is provided, Radius search (Not recommended to use on GPU) if only radius parameter is provided, and Hybrid search (Recommended) if both are provided.

Parameters
inputThe input point cloud with data type float32 or float64.
max_nn[optional] Neighbor search max neighbors parameter. [Default = 100].
radius[optional] Neighbor search radius parameter. [Recommended ~5x voxel size].
Returns
A Tensor of FPFH feature of the input point cloud with shape {N, 33}, data type and device same as input.

◆ CorrespondencesFromFeatures()

core::Tensor open3d::t::pipelines::registration::CorrespondencesFromFeatures ( const core::Tensor source_features,
const core::Tensor target_features,
bool  mutual_filter = false,
float  mutual_consistency_ratio = 0.1 
)

Function to find correspondences via 1-nearest neighbor feature matching. Target is used to construct a nearest neighbor search object, in order to query source.

Parameters
source_feats(N, D) tensor
target_feats(M, D) tensor
mutual_filterBoolean flag, only return correspondences (i, j) s.t. source_features[i] and target_features[j] are mutually the nearest neighbor.
mutual_consistency_ratioFloat threshold to decide whether the number of correspondences is sufficient. Only used when mutual_filter is set to True.
Returns
(K, 2, Int64) tensor. When mutual_filter is disabled: the first column is arange(0, N) of source, and the second column is the corresponding index of target. When mutual_filter is enabled, return the filtering subset of the aforementioned correspondence set where source[i] and target[j] are mutually the nearest neighbor. If the subset size is smaller than mutual_consistency_ratio * N, return the unfiltered set.

◆ EvaluateRegistration()

RegistrationResult open3d::t::pipelines::registration::EvaluateRegistration ( const geometry::PointCloud source,
const geometry::PointCloud target,
double  max_correspondence_distance,
const core::Tensor transformation = core::Tensor::Eye(4, core::Float64core::Device("CPU:0")) 
)

Function for evaluating registration between point clouds.

Parameters
sourceThe source point cloud. (Float32 or Float64 type).
targetThe target point cloud. (Float32 or Float64 type).
max_correspondence_distanceMaximum correspondence points-pair distance.
transformationThe 4x4 transformation matrix to transform source to target of dtype Float64 on CPU device.

◆ GetInformationMatrix()

core::Tensor open3d::t::pipelines::registration::GetInformationMatrix ( const geometry::PointCloud source,
const geometry::PointCloud target,
const double  max_correspondence_distance,
const core::Tensor transformation 
)

Computes Information Matrix, from the transfromation between source and target pointcloud. It returns the Information Matrix of shape {6, 6}, of dtype Float64 on device CPU:0.

Parameters
sourceThe source point cloud. (Float32 or Float64 type).
targetThe target point cloud. (Float32 or Float64 type).
max_correspondence_distanceMaximum correspondence points-pair distance.
transformationThe 4x4 transformation matrix to transform source to target.

◆ ICP()

RegistrationResult open3d::t::pipelines::registration::ICP ( const geometry::PointCloud source,
const geometry::PointCloud target,
const double  max_correspondence_distance,
const core::Tensor init_source_to_target = core::Tensor::Eye(4, core::Float64core::Device("CPU:0")),
const TransformationEstimation estimation = TransformationEstimationPointToPoint(),
const ICPConvergenceCriteria criteria = ICPConvergenceCriteria(),
const double  voxel_size = -1.0,
const std::function< void(const std::unordered_map< std::string, core::Tensor > &)> &  callback_after_iteration = nullptr 
)

Functions for ICP registration.

Parameters
sourceThe source point cloud. (Float32 or Float64 type).
targetThe target point cloud. (Float32 or Float64 type).
max_correspondence_distanceMaximum correspondence points-pair distance.
init_source_to_targetInitial transformation estimation of type Float64 on CPU.
estimationEstimation method.
criteriaConvergence criteria.
voxel_sizeThe input pointclouds will be down-sampled to this voxel_size scale. If voxel_size < 0, original scale will be used. However it is highly recommended to down-sample the point-cloud for performance. By default original scale of the point-cloud will be used.
callback_after_iterationOptional lambda function, saves string to tensor map of attributes such as "iteration_index", "scale_index", "scale_iteration_index", "inlier_rmse", "fitness", "transformation", on CPU device, updated after each iteration.

◆ MultiScaleICP()

RegistrationResult open3d::t::pipelines::registration::MultiScaleICP ( const geometry::PointCloud source,
const geometry::PointCloud target,
const std::vector< double > &  voxel_sizes,
const std::vector< ICPConvergenceCriteria > &  criteria_list,
const std::vector< double > &  max_correspondence_distances,
const core::Tensor init_source_to_target = core::Tensor::Eye(4, core::Float64core::Device("CPU:0")),
const TransformationEstimation estimation = TransformationEstimationPointToPoint(),
const std::function< void(const std::unordered_map< std::string, core::Tensor > &)> &  callback_after_iteration = nullptr 
)

Functions for Multi-Scale ICP registration. It will run ICP on different voxel level, from coarse to dense. The vector of ICPConvergenceCriteria(relative fitness, relative rmse, max_iterations) contains the stopping condition for each voxel level. The length of voxel_sizes vector, criteria vector, max_correspondence_distances vector must be same, and voxel_sizes must contain positive values in strictly decreasing order [Lower the voxel size, higher is the resolution]. Only the last value of the voxel_sizes vector can be {-1}, as it allows to run on the original scale without downsampling.

Parameters
sourceThe source point cloud. (Float32 or Float64 type).
targetThe target point cloud. (Float32 or Float64 type).
voxel_sizesVectorDouble of voxel scales of type double.
criteria_listVector of ICPConvergenceCriteria objects for each scale.
max_correspondence_distancesVectorDouble of maximum correspondence points-pair distances of type double, for each iteration. Must be of same length as voxel_sizes and criterias.
init_source_to_targetInitial transformation estimation of type Float64 on CPU.
estimationEstimation method.
callback_after_iterationOptional lambda function, saves string to tensor map of attributes such as "iteration_index", "scale_index", "scale_iteration_index", "inlier_rmse", "fitness", "transformation", on CPU device, updated after each iteration.