open3d.t.pipelines.odometry.rgbd_odometry_multi_scale

open3d.t.pipelines.odometry.rgbd_odometry_multi_scale(source, target, intrinsics, init_source_to_target=(with default value), , depth_scale=1000.0, depth_max=3.0, criteria_list=[OdometryConvergenceCriteria[max_iteration=10, relative_rmse=1.000000e-06, relative_fitness=1.000000e-06]., OdometryConvergenceCriteria[max_iteration=5, relative_rmse=1.000000e-06, relative_fitness=1.000000e-06]., OdometryConvergenceCriteria[max_iteration=3, relative_rmse=1.000000e-06, relative_fitness=1.000000e-06].], method=<Method.Hybrid: 2>, params=OdometryLossParams[depth_outlier_trunc=7.000000e-02, depth_huber_delta=5.000000e-02, intensity_huber_delta=1.000000e-01].)

Function for Multi Scale RGBD odometry.

Parameters

  • source (open3d.t.geometry.RGBDImage) – The source RGBD image.

  • target (open3d.t.geometry.RGBDImage) – The target RGBD image.

  • intrinsics (open3d.core.Tensor) – (3, 3) intrinsic matrix for projection.

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

    (4, 4) initial transformation matrix from source to target. 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

    ()

  • depth_scale (float, optional, default=1000.0) – Converts depth pixel values to meters by dividing the scale factor.

  • depth_max (float, optional, default=3.0) –

  • criteria_list (List[open3d.t.pipelines.odometry.OdometryConvergenceCriteria], optional, default=[OdometryConvergenceCriteria[max_iteration=10, relative_rmse=1.000000e-06, relative_fitness=1.000000e-06], OdometryConvergenceCriteria[max_iteration=5, relative_rmse=1.000000e-06, relative_fitness=1.000000e-06], OdometryConvergenceCriteria[max_iteration=3, relative_rmse=1.000000e-06, relative_fitness=1.000000e-06]]) – List of Odometry convergence criteria.

  • (open3d.t.pipelines.odometry.Method (method) – 2>): Estimation method used to apply RGBD odometry. One of (PointToPlane, Intensity, Hybrid)

  • optional – 2>): Estimation method used to apply RGBD odometry. One of (PointToPlane, Intensity, Hybrid)

  • default=<Method.Hybrid – 2>): Estimation method used to apply RGBD odometry. One of (PointToPlane, Intensity, Hybrid)

  • params (open3d.t.pipelines.odometry.OdometryLossParams, optional, default=OdometryLossParams[depth_outlier_trunc=7.000000e-02, depth_huber_delta=5.000000e-02, intensity_huber_delta=1.000000e-01]) – Odometry loss parameters.

Returns

open3d.t.pipelines.odometry.OdometryResult