Method of determining an optimum threshold for obtaining binary image signal for use in a pattern recognition system

• US 4,491,962 A
• Filed: 03/23/1982
• Issued: 01/01/1985
• Est. Priority Date: 03/25/1981
• Status: Expired due to Term

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Kashioka, Seiji, Sakou, Hiroshi
• Primary Examiner(s): Masinick, Michael A.
• Assistant Examiner(s): NOT, DEFINED
• Application Number:
  US06/361,479
• Time in Patent Office: 1,015 Days
• Field of Search: 382/8, 382/48, 382/50, 382/52, 382/53, 307/358, 328/116
• US Class Current: 382/270
• CPC Class Codes: G06V 10/28 Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
• CPC Class Codes: G06V 10/443 by matching or filtering

Abstract

A method of determining an optimum threshold for obtaining a binary image signal for use in a pattern recognition system is disclosed in which a pattern of a predetermined shape which exists at one or a plurality of positions on a surface of an object to be imaged is previously stored as a reference pattern. A threshold for obtaining a binary signal is changed stepwise. A video signal is converted into the binary signal on the basis of each value of threshold to obtain a binary image. A plurality of partial patterns which are different in position on the binary image to each other, are successively cut out of the binary image to be compared with the reference pattern. The overall degree of coincidence between the plural partial patterns and the reference pattern is calculated, and a value of threshold at which the overall degree of coincidence is highest, is selected from various values of threshold.

Claims

1. A method of determining an optimum threshold for converting a video signal of an object into a binary signal, suitable for use in a process for recognizing a pattern on said object, comprising the steps of:

storing a pattern expected to exist at one or more positions on a surface of an object as a reference pattern;

converting a video signal of an object into a binary image signal on the basis of each of different values of threshold;

cutting a plurality of partial patterns out of a predetermined region of said binary image signal corresponding to each value of threshold;

detecting the degrees of coincidence each between each of said partial patterns and said reference pattern; and

selecting from said different values of threshold a value of threshold for making highest the overall degree of coincidence between said plurality of partial patterns and said reference pattern, on the basis of aid degrees of coincidence detected at each value of threshold,

wherein an image of said object is divided into a plurality of sections, the highest value of the degrees of coincidence is selected among the degrees of coincidence between a partial pattern and said reference pattern which are obtained for a plurality of partial patterns included in one section, at all of said sections, and the total sum of the highest values of the degrees of coincidence at all of said sections is used as said overall degree of coincidence.

2. A method of determining an optimum threshold according to claim 1, wherein those ones of said highest values of the degrees of coincidence at all of said sections which exceed a predetermined value, are summed up to obtain said overall degree of coincidence.

3. A method of determining an optimum threshold according to claim 1 or 2, wherein the degree of coincidence is detected only for partial patterns cut out of a limited area of said image.

4. A method of determining an optimum threshold according to claim 1 or 2, wherein said different values of threshold are selected from a predetermined range which includes, as a reference value, an optimum threshold used in the pattern recognition processing for a preceding object.

5. A method of determining an optimum threshold according to claim 4, wherein said different values of threshold are varied stepwise in said predetermined range.

6. A method of determining an optimum threshold according to claim 3, wherein said different values of threshold are selected from a predetermined range which includes, as a reference value, an optimum threshold used in the pattern recognition processing for a preceding object.

7. A method of determining an optimum threshold according to claim 6, wherein said different values of threshold are varied stepwise in said predetermined range.

8. A method of determining an optimum threshold for converting a video signal of an object into a binary signal, suitable for use in a process for recognizing a pattern on the object, comprising the steps of:

storing a pattern expected to exist at one or more positions on a surface of an object as a binary reference pattern signal;

converting a video signal of the object into a series of binary image signals each on the basis of a different threshold value;

extracting each binary image signal of a plurality of signals corresponding to respective partial patterns lying within a predetermined region of the pattern;

comparing each partial pattern signal and the reference pattern signal to detect the degree of coincidence therebetween; and

selecting from said different threshold values, one threshold value for making highest the overall degree of coincidence between said plurality of partial pattern signals and said reference pattern signal.

9. A method of determining an optimum threshold according to claim 8, wherein an image of the object is divided into a plurality of sections, the highest value of the degrees of coincidence is selected among the degrees of coincidence between partial patterns and said reference pattern which are obtained for a plurality of partial patterns included in one section, at each of said sections, and the total sum of the highest values of the degrees of coincidence at each of said sections is used as said overall degree of coincidence.

10. A method of determining an optimum threshold according to claim 9, wherein among said highest values of the degrees of coincidence at all of said sections, those which exceed a predetermined value are summed up to obtain said overall degree of coincidence.

11. A method of determining an optimum threshold according to claim 9 or 10, wherein the degree of coincidence is detected only for partial patterns cut out of a limited area of said image.

12. A method of determining an optimum threshold according to claim 8, 9 or 10, wherein said different threshold values are selected from values existing in a predetermined range which is centered on an optimum threshold value found for a preceding object.

13. A method of determining an optimum threshold according to claim 12, wherein said different values of threshold are varied stepwise in said predetermined range.

14. A method of determining an optimum threshold according to claim 11, wherein said different threshold values are selected from values existing a predetermined range which is centered on an optimum threshold value found for a preceding object.

15. A method of determining an optimum threshold according to claim 14, wherein said different values of threshold are varied stepwise in said predetermined range.