Computer implemented method of recognizing a position of an object

• US 5,579,415 A
• Filed: 08/29/1994
• Issued: 11/26/1996
• Est. Priority Date: 08/06/1993
• Status: Expired due to Fees

• Current Assignee: Matsushita Electric Industrial Company Limited (Panasonic Holdings Corporation)
• Original Assignee: Matsushita Electric Industrial Company Limited (Panasonic Holdings Corporation)
• Inventors: Itsuzaki, Yoshihiro, Takano, Misuzu, Nakao, Masaya, Horikami, Kinji, Okumura, Kazumasa
• Primary Examiner(s): Couso, Jose L.
• Application Number:
  US08/294,858
• Time in Patent Office: 820 Days
• Field of Search: 382/282, 382/141, 382/142, 382/143, 382/145, 382/151, 382/152, 382/153, 382/154, 382/199
• US Class Current: 382/282
• CPC Class Codes: G06T 2207/30164 Workpiece; Machine component
• CPC Class Codes: G06T 7/74 involving reference images or patches
• CPC Class Codes: G06V 10/751 Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching

Abstract

A method for position recognition, includes the steps of: picking up an image of an object having a circular arc shape as at least part of its profile, by an image pickup device; converting image information of the picked-up image in gray image data and storing the gray image data into a storage device; scanning the stored gray image with a profile scanning window composed of plural gray-level detection areas crossing an imaginary scanning circular profile corresponding to the shape on the gray image and arranged in circular along the scanning circular profile and using it to measure gray-level representative values of an inner- and outer-side gray-level detection areas inside and outside of the scanning circular profile; calculating, at each scanning position during scanning and for each detection area, an absolute value of a difference between the measured gray-level representative values of the inner- and outer-side gray-level detection areas; determining a number of such detection areas in which the absolute value of the difference is larger than a specified gray-level threshold value; detecting a position of the window where the calculated number of detection areas becomes a maximum with respect to their neighborhood, as a position of the shape of the object; and recognizing a position of the object from the detected position of the shape thereof.

Claims

1. A computer implemented method of recognizing a position of an object, comprising the steps of:

picking up an image of an object having a circular arc shape as at least part of its profile using an image pickup device;

converting image information of the picked-up image into gray image data and storing the gray image data in a storage device;

scanning the stored gray image data with a profile scanning window composed of a plurality of gray-level detection areas crossing an imaginary scanning circular profile corresponding to the circular arc shape on the gray image and arranged circularly along the imaginary scanning circular profile and using it to measure a gray-level representative value of an inner-side gray-level detection area and a gray-level representative value of an outer-side gray-level detection area which are inside and outside of the scanning circular profile;

calculating, at each scanning position during scanning and for each gray-level detection area, an absolute value of a difference between the measured gray-level representative value of the inner-side gray-level detection area and the measured gray-level representative value of the outer-side gray-level detection area;

determining a number of gray-level detection areas in which the absolute value of the difference is larger than a specified gray-level threshold value;

detecting a position of the profile scanning window where the calculated number of gray-level detection areas becomes a maximum with respect to their neighborhood, as a position of the circular arc shape of the object; and

recognizing the position of the object from the detected position of the circular arc shape thereof.

2. The method claimed in claim 1, wherein when the object has a plurality of circular arc shapes on its profile, in the scanning step, the scanning operation is carried out by using one profile scanning window composed of the plurality of gray-level detection areas crossing the imaginary scanning circular profile corresponding to the circular arc shapes on the gray image, and in the recognizing step, the object is identified and a position and a direction of the object are recognized based on relative positional relationship among positions of the plurality of circular arc shapes.

3. The method claimed in claim 1, wherein the object is a slot or an open hole for screwing which has the circular arc shape.

4. The method claimed in claim 1, wherein the object is a slot for screwing which has the circular arc shapes.

5. The method claimed in claim 1, wherein

in the picking up step, a first image of a first object which is a slot or an open hole having a circular arc shape as at least part of its profile, and a second image of a second object which is a hole having a circular arc shape as at least part of its profile, are picked up by an image pickup device,

then the converting, scanning, calculating, determining, and detecting steps are carried out on each of the picked-up first and second images, and

positions of the first and second objects are recognized from the detected positions of the circular arc shapes thereof so as to decide whether or not screwing is appropriate, depending on the recognition result.

11. The method claimed in claim 1, further comprising the steps of:

detecting two end points on the scanning circular profile where a sequence of such gray-level detection areas in which the absolute value of the difference is larger than the specified gray-level threshold value cease continuing;

recognizing an opening direction of the circular arc shape from positions of the detected end points; and

recognizing a direction toward which the object is directed, by the recognized opening direction.

12. The method claimed in claim 11, further comprising, after recognizing the direction toward which the object is directed, the steps of:

restricting a search range for a next to-be-recognized position of the object based on the recognized direction;

detecting the next to-be-recognized position from the restricted search range; and

recognizing the position and direction of the object from the position of the circular arc shape and the next to-be-recognized position.

13. The method claimed in claim 3, wherein when the position of the object is recognized from the detected position of the circular arc shape thereof, a direction of the slot or an opening direction of the open hole is also recognized from distributed positions of the gray-level detection areas at the detected position of the circular arc shape.

14. The method claimed in claim 1, wherein the gray-level representative value of the inner-side gray-level detection area is a total value of a specified number of inner-side image gray-level measuring points constituting the inner-side gray-level detection area and the gray-level representative value of the outer-side gray-level detection area is a total value of the same number of outer-side image gray-level measuring points constituting the outer-side gray-level detection area as that of the inner-side image gray-level measuring points.

15. The method claimed in claim 1, wherein the gray-level representative value of the inner-side gray-level detection area is an average value of a specified number of inner-side image gray-level measuring points constituting the inner-side gray-level detection area and the gray-level representative value of the outer-side gray-level detection area is an average value of a specified number of outer-side image gray-level measuring points constituting the outer-side gray-level detection area.

16. The method claimed in claim 1, wherein the gray-level representative value of the inner-side gray-level detection area is a value selected from inner-side image gray-level measuring points constituting the inner-side gray-level detection area and the gray-level representative value of the outer-side gray-level detection area is a value selected from outer-side image gray-level measuring points constituting the outer-side gray-level detection area.

17. The method claimed in claim 3, further comprising the steps of:

calculating a gripping point and direction of a screw to be screwed into the object; and

outputting the coordinates of the gripping point and an angle of the direction of the screw to external equipment.

6. A computer implemented method of recognizing a position of an object, comprising the steps of:

picking up an image of an object having a circular arc shape as at least part of its profile, by an image pickup device;

converting image information of the picked-up image into gray image data and storing the gray image data in a storage device;

scanning the stored gray image data with a profile scanning window composed of a plurality of gray-level detection areas crossing an imaginary scanning circular profile corresponding to the circular arc shape on the gray image and arranged circularly along the imaginary scanning circular profile and using it to measure a gray-level representative value of an inner-side gray-level detection area and a gray-level representative value of an outer-side gray-level detection area which are inside and outside of the scanning circular profile;

calculating, at each scanning position during scanning and for each gray-level detection area, an absolute value of a difference between the measured gray-level representative value of the inner-side gray-level detection area and the measured gray-level representative value of the outer-side gray-level detection area;

summing up the absolute values of all of the gray-level detection areas;

detecting a position of the profile scanning window where the summed absolute values of all of the gray-level detection areas becomes a maximum with respect to their neighborhood, as a position of the circular arc shape of the object; and

recognizing the position of the object from the detected position of the circular arc shape thereof.

7. The method claimed in claim 6, wherein when the object has a plurality of circular arc shapes on its profile, in the scanning step, the scanning operation is carried out by using one profile scanning window composed of the plurality of gray-level detection areas crossing the imaginary scanning circular profile corresponding to the circular arc shapes on the gray image, and in the recognizing step, the object is identified and a position and a direction of the object are recognized based on relative positional relationship among positions of the plurality of circular arc shapes.

8. The method claimed in claim 6, wherein the object is a slot or an open hole for screwing which has the circular arc shape.

9. The method claimed in claim 7, wherein the object is a slot for screwing which has the circular arc shape.

10. The method claimed in claim 6, wherein:

in the picking up step, a first image of a first object which is a slot or an open hole having a circular arc shape as at least part of its profile, and a second image of a second object which is a hole having a circular arc shape as at least part of its profile, are picked up by an image pickup device,

then the converting, scanning, calculating, determining, and detecting steps are carried out on each of the picked-up first and second images, and

positions of the first and second objects are recognized from the detected positions of the circular arc shapes thereof so as to decide whether or not screwing is appropriate, depending on the recognition result.

18. The method claimed in claim 7, further comprising the steps of:

calculating a gripping point and direction of a screw to be screwed into the object; and

outputting the coordinates of the gripping point and an angle of the direction of the screw to external equipment.