Image processing device and method for detecting the location of the feature of interest in an object image

• US 6,055,334 A
• Filed: 07/25/1995
• Issued: 04/25/2000
• Est. Priority Date: 07/25/1994
• Status: Expired due to Fees

• Current Assignee: Omron Corporation
• Original Assignee: Omron Corporation
• Inventors: Kato, Takayuki
• Primary Examiner(s): Boudreau, Leo H.
• Assistant Examiner(s): Mariam, Daniel G.
• Application Number:
  US08/507,011
• Time in Patent Office: 1,736 Days
• Field of Search: 382/190, 382/192, 382/194, 382/201, 382/205, 382/197, 382/103, 395/142, 395/216, 395/289, 395/294, 395/119
• US Class Current: 382/190
• CPC Class Codes: G06V 10/7515 Shifting the patterns to accommodate for positional errors

Abstract

An image and method for determining the location of known pattern of data in a captured object image wherein a feature analysis is performed upon a representative plurality of pixel locations to determine a starting point, feature analysis is performed upon the pixel locations surrounding the starting point and the starting point is moved to another pixel location if provides a higher feature analysis value than the starting point; the process of performing feature analyses and moving the starting point continues until the starting point is moved a maximum feature analysis value; once the maximum feature analysis value is found, the object image data may be rotated incrementally with respect to the reference pattern data and the feature analysis/starting point selection process repeated for each iteration until an object image location and object angle of rotation is determined.

Claims

1. An image processing device for detecting a target location in an object image of an object pattern that resembles a reference pattern, said device comprising:

an imaging device which captures said object image and outputs object image data which represents said object image;

a memory device which stores said object image data, said reference pattern, and a first search zone on said object image to define a maximum searching area for searching for said object pattern in said object image;

a starting point searching means to determine a selected starting point from a plurality of predetermined representative pixel locations for detecting said target location of said object pattern in said first search zone by calculating a feature analysis value representing a degree of resemblance between said reference pattern and said object pattern in a representative scanning window defined by each of said plurality of predetermined representative pixel locations and determining said selected starting point to be the one of said plurality of predetermined representative pixel locations having the highest feature analysis value;

a feature counting means which takes said selected starting point as a center point of a second searching zone, said second searching zone being smaller than said first searching zone, and calculates said feature analysis value representing a degree of resemblance between said reference pattern and said object pattern defined by each pixel location in said second searching zone; and

a first control device to find a specific pixel location in said second searching zone where said object pattern has the highest feature analysis value, and to determine said specific pixel location to be said target location of said object pattern to be detected if said specific pixel location is at said selected starting point, and if not, to reset said second searching zone to a third searching zone whose center pixel is said specific pixel location having the highest feature analysis value, and repeat said calculating said feature analysis value of said object pattern defined by each pixel location in said third searching zone until said calculated feature analysis value for a pixel location in said third searching area is identical to a feature analysis value for an immediately preceding pixel location, and to determine said pixel location to be said target location of said object pattern to be detected.

2. An image processing device according to claim 1, further comprising a second control device to incrementally rotate said object image with respect to said reference pattern, while said feature counting means calculates said feature analysis, and said first control device finds said specific pixel location, thereby said second control device finding a target angle of said object image with respect to said reference pattern.

3. An image processing device according to claim 1, wherein said feature analysis value representing a degree of resemblance is a normalized correlation value representing a degree of coincidence between said reference pattern and said object pattern.

4. An image processing device according to claim 1, wherein, when there are a plurality of specific pixel locations which have the same highest feature value and one of said plurality of specific pixel locations is said selected starting point in said second searching zone, said first control device finds a sum of said feature analysis values of each object pattern defined by said plurality of specific pixel locations which have the same highest feature value, determines said specific location having the highest sum as a first summed specific pixel location, and said first control device defines said first summed specific pixel location as said target location of said object pattern if said first summed specific pixel location is the same as said selected starting point, or said first control device sets said first summed specific pixel location as said center point of said third searching zone if said first summed specific pixel location is not the same as said selected starting point.

5. An image processing device according to claim 4, wherein, when there are a plurality of specific pixel locations which have the same highest feature value in said second searching zone, said first control device selects one of said plurality of specific location as said first summed specific pixel location based upon a predetermined priority scheme.

6. An image processing device according to claim 1, wherein, when there are a plurality of specific pixel locations which have the same highest feature value and none of said plurality of specific pixel locations is said selected starting point in said second searching zone, said first control device finds a sum of said feature analysis values of each object pattern defined by said plurality of specific pixel locations which have the same highest feature value, and determines said specific location having the highest sum as a second summed specific pixel location, and said second summed specific pixel location set as a center pixel of said third searching zone.

7. An image processing device according to claim 6, wherein, when there are a plurality of specific pixel locations which have the same highest feature value in said second searching zone, said first control device selects one of said plurality of specific location as said first summed specific pixel location based upon a predetermined priority scheme.

8. An image processing device according to claim 2, wherein said second control device calculates said target angle of rotation at which said object image most nearly matches said reference pattern, based on, by means of parabolic approximation, said feature analysis values at said rotated second and third searching zones and each angle of said incremental rotation.

9. An image processing device according to claim 2, further comprising a fine searching means to detect said target location in an object image with a finer precision than one pixel size based on said highest feature analysis value and said specific pixel location found by said first or second control device, and feature analysis feature values at pixel locations surrounding said specific pixel location.

10. An image processing device according to claim 9, wherein said fine searching means detects said target location based on an apex location of a parabolic curve which is calculated by three points defined with pixel locations in vertical and horizontal directions and feature analysis values, said three points in horizontal direction is defined by said specific pixel location in horizontal direction and said highest feature analysis value, and horizontally neighboring two pixel locations and feature analysis values, said three points in vertical direction is defined by said specific pixel location in vertical direction and said highest feature analysis value, and vertically neighboring two pixel locations and feature analysis values.

11. An image processing device according to claim 9, wherein said fine searching means detects said target location by barycentric calculation.