Pattern recognition for sheet metal part models

• US 7,158,853 B2
• Filed: 10/25/2004
• Issued: 01/02/2007
• Est. Priority Date: 10/25/2004
• Status: Expired due to Fees

• Current Assignee: Amada Company Limited
• Original Assignee: Amada Company Limited
• Inventors: Sakai, Satoshi, Ishii, Masanobu
• Primary Examiner(s): Picard; Leo
• Assistant Examiner(s): KASENGE, CHARLES R
• Application Number:
  US10/971,124
• Publication Number:
  US 20060089738A1
• Time in Patent Office: 799 Days
• Field of Search: 700/95, 700/97, 700/98, 700/182, 700/117, 700/145, 700/165, 700/174, 378/4, 378/901, 708/131, 703/1, 703/6, 703/7
• US Class Current: 700/182
• CPC Class Codes: G06V 10/424 Syntactic representation, e.g. by using alphabets or grammars
• CPC Class Codes: G06V 10/435 Computation of moments

Abstract

Analysis is performed on sheet metal products to determine characteristics (or feature vectors) of the sheet metal products. Multiple feature vectors can be obtained for each sheet metal product. Exemplary analyses include a first order moment and a second order moment. The characteristics can be compared with characteristics of other sheet metal part shapes to determine how similar the shapes are.

Claims

1. A method of specifying characteristics of a sheet metal product, comprising:

obtaining feature vectors derived from at least one characteristic extraction method, at least one feature vector having a size and a direction;

comparing a feature vector of the sheet metal product with a feature vector of another sheet metal product; and

determining a level of similarity between the sheet metal products based upon the comparison.

2. The method of claim 1, wherein the feature vector is derived by applying a moment as the characteristic extraction method.

3. The method of claim 2, in which the moment comprises a second order moment.

4. The method of claim 2, in which the moment comprises a first order moment.

5. The method of claim 2, in which the applying further comprises applying the moment to a 3D sheet metal part shape.

6. The method of claim 2, in which the applying further comprises applying the moment to a 2D sheet metal pad shape.

7. The method of claim 2, in which the applying further comprises applying the moment to a 3D sheet metal part shape and applying the moment to a 2D sheet metal part shape.

8. The method of claim 1, further comprising:

obtaining at least one other feature vector of the sheet metal product; and

comparing the feature vectors of the sheet metal product with a plurality of corresponding feature vectors of the other sheet metal product to determine a refined degree of similarity between the sheet metal products.

9. The method of claim 8, in which a user selects the feature vectors, and assigns a weight to each selected feature vector, the refined degree of similarity being determined based upon the assigned weights and the selected feature vectors.

10. The method of claim 1, wherein the feature vector is derived by performing a sheet metal CT scan cross section as the characteristic extraction method.

11. The method of claim 10, wherein the feature vector is derived by performing a UZ analysis as the characteristic extraction method.

12. A computer readable medium storing a program for specifying characteristics of a sheet metal product, comprising:

a feature vector code segment that obtains a first plurality of feature vectors derived from at least one characteristic extraction method, at least one feature vector having a size and a direction;

a comparing code segment that compares the at least one feature vector of the sheet metal product with a feature vector of another sheet metal product; and

a similarity code segment that determines a level of similarity between the sheet metal products based upon the comparison.

13. The medium of claim 12, in which the feature vector is derived by applying a moment as the characteristic extraction method.

14. The medium of claim 13, in which the moment comprises a second order moment.

15. The medium of claim 13, in which the moment comprises a first order moment.

16. The medium of claim 12, in which the obtaining code segment further obtains at least one other feature vector of the sheet metal product, the comparing code segment further compares the at least one feature vector and the at least one other feature vector, and the similarity code segment further determines a refined level of similarity based on the further comparing, the further comparing comprising comparing the plurality of feature vectors of the sheet metal product with a plurality of corresponding feature vectors of another sheet metal product.

17. The method of claim 16, further comprising an input receiving code segment that receives a user'"'"'s selection of the plurality of the feature vectors, and assignment of a weight to each selected feature vector, the refined degree of similarity being determined based upon the assigned weights and the selected feature vectors.

18. The medium of claim 12, wherein the at least one feature vector is derived by performing a sheet metal CT scan cross section as the characteristic extraction method.

19. The medium of claim 18, wherein the at least one feature vector is derived by performing a UZ analysis as the characteristic extraction method.

20. A method for searching for similar sheet metal part shapes, comprising:

inputting a freehand sketch representing a sheet metal part shape;

determining at least one feature vector of the input sheet metal part shape; and

comparing the at least one feature vector of the input sheet metal part shape with at least one corresponding feature vector of another sheet metal part shape to determine a degree of similarity between the sheet metal part shapes.

21. The method of claim 20, in which the inputting further comprises inputting a flat pattern outline.

22. The method of claim 20, in which the inputting further comprises inputting cross section lines.

23. The method of claim 20, in which the determining further comprises determining a plurality of feature vectors of the input sheet metal part shape, and

the comparing further comprises comparing the plurality of feature vectors of the input sheet metal part shape with a plurality of corresponding feature vectors of the other sheet metal part shape to determine a refined degree of similarity between the sheet metal part shapes.

24. A computer readable medium storing a program for searching for similar sheet metal part shapes, comprising:

an input code segment that accepts input of a freehand sketch representing a sheet metal part shape that is used as a key part shape;

a feature vector code segment that determines at least one feature vector of the key part shape; and

a similarity analysis code segment that compares the at least one feature vector of the key part shape with at least one corresponding feature vector of another sheet metal part shape to determine a degree of similarity between the sheet metal part shapes.

25. The medium of claim 24, in which the input code segment further accepts input of a flat pattern outline.

26. The medium of claim 24, in which the input code segment further accepts input of cross section lines.

27. The medium of claim 24, in which the feature vector code segment further determines a plurality of feature vectors of the key part shape, and

the similarity analysis code segment further compares the plurality of feature vectors of the key part shape with a plurality of corresponding feature vectors of the other sheet metal part shape to determine a refined degree of similarity between the sheet metal part shapes.