Method and apparatus for calculating position and orientation by combination of features of partial shapes

• US 4,803,735 A
• Filed: 10/09/1986
• Issued: 02/07/1989
• Est. Priority Date: 10/11/1985
• Status: Expired due to Term

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Nishida, Yoshie, Tsuji, Masao, Hata, Seiji
• Primary Examiner(s): Boudreau, Leo H.
• Application Number:
  US06/917,053
• Time in Patent Office: 852 Days
• Field of Search: 382/8, 382/34, 382/48, 382/45
• US Class Current: 382/151
• CPC Class Codes: G06T 7/73 using feature-based methods
• CPC Class Codes: G06V 10/40 Extraction of image or video features

Abstract

There are disclosed a method and an apparatus for properly recognizing the position and orientation of a whole object article from its partial shape. The method comprises the steps of first extracting an image of a partial shape from a two-dimensional picture image representing at least one portion of the object article whose position and orientation are to be recognized, then calculating at least one of a center of gravity, an intersection of straight lines and a linear equation with respect to each of the extracted feature shapes, thereby obtaining at least one feature parameter, and calculating the position and orientation of the whole object article by arbitrarily combining such feature parameters.

Claims

1. A position and orientation calculating method comprising the steps of:

inputting a two-dimensional picture image of an object article obtained from an image detector;

extracting at least two two-dimensional images from at least two windows defining image processing ranges corresponding to two portions of the object article in accordance with coordinates of each of said two windows, and extracting data describing coordinates of a sequence of points from a sequence of segments approximated as a polygon for an image designating a contour of a portion of the object article between each of said two windows;

calculating a center of gravity from said data with respect to a pattern adapted for extraction of an entirety of a partial shape from said two two-dimensional images and calculating a straight line by a linear equation from said data with respect to the pattern adapted for extraction of the partial shape from said windows; and

calculating calculated data describing position and orientation of the object article from said center of gravity and calculating the orientation of said object article from a combination of said center of gravity and the intersection of said straight line and a perpendicular line extending thereto from said center of gravity.

2. A position and orientation calculating method according to claim 1, wherein the actual position and orientation are defined by computing differences between said calculated data and reference dictionary data.

3. A position and orientation calculating method comprising the steps of:

inputting a two-dimensional picture image of an object article obtained from an image detector;

extracting at least two two-dimensional picture images from at least two windows defining image processing ranges corresponding to two portions of the object article in accordance with coordinates of each of said two windows, and extracting data describing coordinates of a sequence of points from a sequence of segments approximated as a polygon for an image designating a contour of a portion of the object article between each of said windows;

calculating at least two centers of gravity from said data with respect to a pattern adapted for extraction of entire partial shapes by each of said windows; and

calculating as calculated data describing the position of said object article from one of said centers of gravity and the orientation of said object article from said two centers of gravity.

4. The method as defined in claim 3, wherein, the actual position and orientation are defined by computing differences between said calculated data and reference dictionary data.

5. A position and orientation calculating method comprising the steps of:

including a two-dimensional picture image of an object article obtained from an image detector;

extracting at least two two-dimensional picture images from at least two windows defining image processing ranges corresponding to portioins of the object article in accordance with the coordinates of each said two windows, and extracting data describing coordinates of a sequence of points in a sequence of segments approximated as a polygon for an image designating a contour of a portion of the object article between each of said windows;

calculating at least two straight lines by a linear equation from said data with respect to patterns adapted for extraction of partial shapes from said windows; and

calculating as calculated data describing the position of said object article from the intersection of said two straight lines and orientation of said object article from the inclination of one of said straight lines.

6. The method as defined in claim 5, wherein the actual position and orientation are defined by computing differences between said calculated data and reference dictionary data.

7. The method as defined in claim 6, wherein the position and orientation of said object article are calculated according to an arbitrary combination of said four patterns.

8. A position and orientation calculating apparatus, comprising:

an image detector for inputting a two-dimensional picture image of an object article;

an image processor including;

extracting means for extracting at least two twodimensional picture images from at least two windows defining image processing ranges corresponding to poritons of the object article in accordance with coordinates for each of said two windows, and for extracting extractable data describing coordinates for a sequence of points from a sequence of segments approximating a polygon for an image designating the contour of portions of the object article between each of said windows, and

calculating means for calculating information describing a center of gravity from said extractable data with respect to a pattern adapted for extraction of an entirety of a partial shape from said two images provided by said windows and calculating a straight line by a linear equation from said extractable data with respect to a pattern adapted for extraction of the partial shape by said windows, and calculating information describing a position of said object article from said center of gravity and said information describing the position and orientation of said object article from a combination of said center of gravity and the intersection of said straight line and a perpendicular line extending thereto from said center of gravity; and

correcting means for providing corrected data by correcting the difference between position and orientation calculated and reference dictionary data defining said coordinates of each of said two windows on the basis of the calculation position and orientation information.

9. The apparatus as defined in claim 8, wherein the apparatus is adapted for controlling a robot for moving the object article in accordance with corrected data provided by said correcting means.