Identification of occluded edge regions from 3D point data

• US 7,843,448 B2
• Filed: 11/21/2005
• Issued: 11/30/2010
• Est. Priority Date: 11/21/2005
• Status: Active Grant

• Current Assignee: Leica Geosystems AG (Hexagon AB)
• Original Assignee: Leica Geosystems AG (Hexagon AB)
• Inventors: Wheeler, Mark Damon, Bukowski, Richard William, Kung, Jonathan Apollo
• Primary Examiner(s): Tung; Kee M
• Assistant Examiner(s): CRADDOCK, ROBERT J
• Application Number:
  US11/285,582
• Publication Number:
  US 20070136408A1
• Time in Patent Office: 1,835 Days
• Field of Search: 345/419, 345/420, 702/151, 702/152, 702/155
• US Class Current: 345/419
• CPC Class Codes: G01S 17/89 for mapping or imaging
• CPC Class Codes: G06T 17/00 Three dimensional [3D] modelling, e.g. data description of 3D objects
• CPC Class Codes: G06T 2200/24 involving graphical user interfaces [GUIs]

Abstract

An improved interface and algorithm(s) can be used to simplify and improve the process for locating an occluded edge from a series of points in a point cloud. An interface can allow the user to select a hint point thought to be near an edge of interest, which can be used to generate an initial edge profile. An interface can allow the user to adjust the fit of the initial profile in cross-section, then can use that profile to generate a profile of the entire edge. A moving fit window can use an imaginary plane to provide an additional constraint, and can utilize a moving average to extend the edge and determine proper end locations. An interface then can display the results of the fit to the user and allow the user to adjust the fit, such as by adjusting the end points of the calculated edge. Such a process can be used to fit linear or curvilinear occluded edges, and can fit a number of irregular shapes as well as regular shaped edges such as “v-shaped” edges.

Claims

1. A method of aligning a view along an occluded edge corresponding to data points in a point cloud, comprising the steps of:

receiving a seed point selected from the data points contained in the point cloud that is indicated to lie near the edge;

generating a covariance matrix using adjacent points near the seed point in the point cloud and calculating the eigenvectors of the covariance matrix;

determining surface data points near the seed point that are likely to lie along a surface adjacent to the occluded edge;

determining an average vector using a surface vector from each surface data point to the seed point;

determining the edge direction using a smallest eigenvector of the eigenvectors and the average vector; and

displaying points near the seed point in a view that corresponds to a cross-section of the edge near the seed point based on the edge direction.

2. A method of creating an initial surface profile corresponding to data points in a point cloud, comprising the steps of:

receiving a seed point selected from the data points contained in the point cloud that is indicated to lie near the edge;

generating a covariance matrix using adjacent points near the seed point in the point cloud and calculating the eigenvectors of the covariance matrix;

determining surface data points near the seed point that are likely to lie along a surface adjacent to the occluded edge;

determining an average vector using a surface vector from each surface data point to the seed point; and

determining the edge direction using a smallest eigenvector of the eigenvectors and the average vector.

3. A method according to claim 2, further comprising:

determining a line segment having an acceptable fit to the adjacent points orthogonal to the edge direction.

4. A method according to claim 3, further comprising:

extending the line segment to determine whether the extended line segment passes near the seed point, the line segment forming the initial surface profile when passing near the seed point.

5. A method of modeling an occluded edge corresponding to data points in a point cloud, comprising the steps of:

receiving a seed point selected from the data points contained in the point cloud that is indicated by a user to lie near the occluded edge;

determining a direction of extension of the occluded edge;

determining a line segment having an acceptable fit to the adjacent points orthogonal to the direction of extension;

generating an orthogonal plane that is orthogonal to the line segment and intersects the line segment near the seed point; and

modeling the occluded edge using constraints imposed by the line segment and orthogonal plane,

wherein the step of determining a direction of extension of the occluded edge includes generating a covariance matrix using adjacent points near the seed point in the point cloud and calculating the eigenvectors of the covariance matrix, determining surface data points near the seed point that are likely to lie along a surface adjacent the occluded edge, determining and average vector using a surface vector from each surface data point to the seed point, and determining the edge direction using a smallest eigenvector of the eigenvectors and the average vector.

6. A method according to claim 5, wherein:

the step of modeling the occluded edge includes using a random consensus algorithm.

7. A method according to claim 5, wherein the point cloud excludes data points on an occluded side of the edge.

8. A method according to claim 5, wherein the seed point is a first seed point, the method further comprising:

receiving a second seed point selected from the data points contained in the point cloud that is indicated by a user to lie near the occluded edge.

9. A method according to claim 8, wherein the first seed point lies near a substantially linear portion of the edge.

10. A method according to claim 8, wherein the second seed point lies near an end of the edge.

11. A method according to claim 8, wherein the second seed point lies near a discontinuity of the edge.