open3d.pipelines.odometry.compute_correspondence#

open3d.pipelines.odometry.compute_correspondence(intrinsic_matrix: numpy.ndarray[numpy.float64[3, 3]], extrinsic: numpy.ndarray[numpy.float64[4, 4]], depth_s: open3d.geometry.Image, depth_t: open3d.geometry.Image, option: open3d.pipelines.odometry.OdometryOption = OdometryOption(iteration_number_per_pyramid_level=[20, 10, 5], depth_diff_max=0.03, depth_min=0, depth_max=4)) → open3d.utility.Vector4iVector#

Function to estimate point to point correspondences from two depth images. A vector of u_s, v_s, u_t, v_t which maps the 2d coordinates of source to target.

Parameters:

  • intrinsic_matrix (numpy.ndarray[numpy.float64[3, 3]]) – Camera intrinsic parameters.

  • extrinsic (numpy.ndarray[numpy.float64[4, 4]]) – Estimation of transform from source to target.

  • depth_s (open3d.geometry.Image) – Source depth image.

  • depth_t (open3d.geometry.Image) – Target depth image.

  • option (open3d.pipelines.odometry.OdometryOption, optional, default=OdometryOption( iteration_number_per_pyramid_level=[ 20, 10, 5, ] , depth_diff_max=0.03, depth_min=0, depth_max=4, )) – Odometry hyper parameters.

Returns:

open3d.utility.Vector4iVector