open3d.t.pipelines.registration.multi_scale_icp

open3d.t.pipelines.registration.multi_scale_icp(source, target, voxel_sizes, criteria_list, max_correspondence_distances, init_source_to_target=(with default value), estimation_method=TransformationEstimationPointToPoint, save_loss_log=False)

Function for Multi-Scale ICP registration

Parameters

• source (open3d.t.geometry.PointCloud) – The source point cloud.

• target (open3d.t.geometry.PointCloud) – The target point cloud.

• voxel_sizes (open3d.utility.DoubleVector) – o3d.utility.DoubleVector of voxel sizes in strictly decreasing order, for multi-scale icp.

• criteria_list (List[open3d.t.pipelines.registration.ICPConvergenceCriteria]) – List of Convergence criteria for each scale of multi-scale icp.

• max_correspondence_distances (open3d.utility.DoubleVector) – o3d.utility.DoubleVector of maximum correspondence points-pair distances for multi-scale icp.

• init_source_to_target (open3d.core.Tensor, optional) –

  Initial transformation estimation Default value:

  [[1.0 0.0 0.0 0.0], [0.0 1.0 0.0 0.0], [0.0 0.0 1.0 0.0], [0.0 0.0 0.0 1.0]] Tensor[shape={4, 4}, stride={4, 1}, Float64

  ()

• estimation_method (open3d.t.pipelines.registration.TransformationEstimation, optional, default=TransformationEstimationPointToPoint) – Estimation method. One of (TransformationEstimationPointToPoint, TransformationEstimationPointToPlane, TransformationEstimationForColoredICP, TransformationEstimationForGeneralizedICP)

• save_loss_log (bool, optional, default=False) – When True, it saves the iteration-wise values of fitness, inlier_rmse, transformaton, scale, iteration in loss_log_ in regsitration_result. Default: False.

Returns

open3d.t.pipelines.registration.RegistrationResult