Position and orientation estimation method and apparatus therefor

• US 9,153,030 B2
• Filed: 01/12/2011
• Issued: 10/06/2015
• Est. Priority Date: 02/25/2010
• Status: Active Grant

• Current Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Original Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Inventors: Miyatani, Sonoko
• Primary Examiner(s): Chauhan, Ulka
• Assistant Examiner(s): NGUYEN, ANH TUAN V
• Application Number:
  US13/580,024
• Publication Number:
  US 20120320053A1
• Time in Patent Office: 1,728 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06T 2207/10004 Still image; Photographic image
• CPC Class Codes: G06T 2207/10028 Range image; Depth image; 3D point clouds
• CPC Class Codes: G06T 2207/30164 Workpiece; Machine component
• CPC Class Codes: G06T 7/75 involving models

Abstract

There is provided a method for accurately estimating a position and orientation of an object even if the object is more dispersive in shape than a three-dimensional geometric model with a standard shape. The statistic of deviation of a feature constituting a three-dimensional model representing a three-dimensional standard shape of an object is estimated to determine a reliability for each feature. The amount of deviation is calculated between the feature extracted from observation data obtained by an imaging apparatus and the feature in the three-dimensional model. The three-dimensional position and orientation of the object is estimated based on the amount of deviation and the reliability related to each feature extracted from the three-dimensional model.

Claims

1. A position and orientation estimation method comprising:

obtaining a statistic of deviation of a geometric feature constituting a three-dimensional geometric model of an object for each geometric feature based on tolerance data for the object, the tolerance data representing allowable differences between the object and the three-dimensional geometric model of the object;

setting a reliability for each geometric feature based on the statistic of deviation, wherein the reliability is set higher as the statistic of deviation is smaller;

obtaining an approximate position and orientation of the object;

extracting an image feature of a captured image, wherein the image feature is an edge;

associating the geometric feature with the image feature based on the obtained approximate position and orientation for each geometric feature;

calculating an amount of difference between the geometric feature and the image feature for each geometric feature; and

estimating the position and orientation of the object by correcting the obtained approximate position and orientation based on the reliability and the amount of difference of each geometric feature.

2. The position and orientation estimation method according to claim 1, wherein obtaining a statistic of deviation comprises:

extracting a line segment constituting the three-dimensional geometric model;

extracting a tolerance causing a deviation of the line segment;

calculating the tolerance for each predetermined interval of the line segment; and

generating a vector of a tolerance in the maximum tolerance and taking the vector as the statistic of deviation.

3. The position and orientation estimation method according to claim 1, wherein the statistic of deviation is represented by probability distribution.

4. The position and orientation estimation method according to claim 3, wherein obtaining a statistic of deviation comprises:

extracting a line segment constituting the three-dimensional geometric model;

extracting a tolerance causing a deviation of the line segment;

calculating the tolerance for each predetermined interval of the line segment; and

generating a vector of a normal distribution based on the tolerance and taking the vector as the statistic of deviation.

5. The position and orientation estimation method according to claim 1, wherein obtaining a statistic of deviation comprises:

extracting a surface constituting the three-dimensional geometric model;

calculating a tolerance in the center of gravity of the surface; and

generating a vector of a tolerance in the maximum tolerance and taking the vector as the statistic of deviation.

6. The position and orientation estimation method according to claim 1, further comprising measuring a geometric feature constituting a shape of the object,

wherein calculating a statistic of deviation includes measuring the statistic based on a deviation between the measured result and the three-dimensional geometric model.

7. The position and orientation estimation method according to claim 6, wherein measuring a geometric feature includes performing measurement a plurality of times for each plurality of objects.

8. The position and orientation estimation method according to claim 6, wherein measuring a geometric feature includes performing measurement a plurality of times for each plurality of measurement positions.

9. The position and orientation estimation method according to claim 1, wherein the geometric feature is a point, line segment, or surface.

10. The position and orientation estimation method according to claim 1, wherein calculating the reliability is calculated using a likelihood function.

11. A non-transitory storage medium storing a program for causing a computer to execute the position and orientation estimation method according to claim 1.

12. A position and orientation estimation apparatus comprising:

a statistic obtaining unit configured to obtain a statistic of deviation of a geometric feature constituting a three-dimensional geometric model of the object for each geometric feature based on tolerance data for the object, the tolerance data representing allowable differences between the object and the three-dimensional geometric model of the object;

a reliability calculation unit configured to calculate a reliability for each geometric feature based on the statistic of deviation, wherein the reliability is set higher as the statistic of deviation is smaller;

an obtaining unit configured to obtain an approximate position and orientation of the object;

an image feature extraction unit configured to extract an image feature of a captured image captured by image capture means, wherein the image feature is an edge;

an association unit configured to associate the geometric feature with the image feature based on the obtained approximate position and orientation for each geometric feature;

a deviation-amount calculation unit configured to calculate an amount of difference between the geometric feature and the image feature for each geometric feature; and

position and orientation estimation unit configured to estimate a position and orientation of the object by correcting the obtained approximate position and orientation based on the reliability and the amount of difference of each geometric feature.

13. The position and orientation estimation apparatus according to claim 12, further comprising an image capture unit configured to capture an image of the object.

14. The position and orientation estimation apparatus according to claim 13, wherein the image capture unit captures a range image in a real space as the captured image.