Method for inspecting garments for holes having a contrasting background

• US 5,283,443 A
• Filed: 12/03/1992
• Issued: 02/01/1994
• Est. Priority Date: 04/17/1990
• Status: Expired due to Fees

• Current Assignee: De Montfort University
• Original Assignee: De Montfort University
• Inventors: Abbasszadeh, Siavash, Norton-Wayne, Leonard
• Primary Examiner(s): Nelms, David C.
• Assistant Examiner(s): LE, QUE TAN
• Application Number:
  US07/937,872
• Time in Patent Office: 425 Days
• Field of Search: 250/572, 250/562, 250/563, 356/392, 356/393, 356/394, 358/106, 358/107, 382/48, 382/51, 382/55, 382/8, 382/44, 382/45, 382/46
• US Class Current: 250/559.42
• CPC Class Codes: D06H 3/08 by photo-electric or television means D06H3/12, D06H3/14, D06H3/16 take precedence
• CPC Class Codes: G01N 2021/8887 based on image processing techniques
• CPC Class Codes: G01N 21/8983 for testing textile webs, i.e. woven material
• CPC Class Codes: G01N 2201/102 Video camera

Abstract

A method for inspecting garments for holes includes front-lighting the garment against a contrasting background and forming a pixel image thereof, isolating garment pixels from background pixels by automatically selecting an optimum binarization threshold grey level, binarizing the image at that level, determining the garment boundary, testing for consistency of grey level within the garment boundary on the binarized image along rows and columns of pixels within the garment boundary and noting inconsistent pixels, and identifying as holes only pixels which have been noted as inconsistent in both horizontal and vertical tests.

Claims

1. A method for inspecting garments for holes comprising:

front-lighting the garment against a contrasting background and forming a pixel image thereof,

isolating garment pixels from background pixels by automatically selecting an optimum binarization threshold grey level,

binarizing the image at that level,

determining the garment boundary,

testing for consistency of grey level within the garment boundary on the binarized image along rows and columns of pixels within the garment boundary and noting inconsistent pixels, and

identifying as holes only pixels which have been noted as inconsistent in both horizontal and vertical tests.

2. A method according to claim 1, which comprises selecting the threshold grey level on the basis of data from a part of the image.

3. A method according to claim 2, which comprises using a single row or column of pixels for said data.

4. A method according to any one of claims 1 to 3, which comprises selecting the threshold grey level by thresholding the image at all possible grey levels to yield a histogram in which the number of segmented pixels is plotted against the threshold grey level, said histogram having two peak values corresponding to object and background, and taking the optimum binarization threshold grey level to be that level at which the histogram has a minimum value between the peaks.

5. A method according to claim 1, which comprises determining the garment boundary in the binarized image by tracing the boundary with an eight-directional Freeman chain code robot.

6. A method according to claim 1 which comprises comparing the length of the determined garment boundary to a predetermined minimum length to eliminate spurious measurements of noise clusters in the background.

7. A method according to claim 1, which comprises producing a fresh, binarized image for the grey level consistency testing by a low pass filter operation on the original pixel image, and binarizing said filtered image using the selected threshold grey level.

8. A method according to claim 1, which comprises indicating any inconsistent pixel group discovered within the garment boundary on a displayed image by a superimposed marker.

9. A method according to claim 1, which comprises making the pixel image by a video camera.

10. A method according to claim 1, in which the garment comprises a knitted garment.

11. A method according to claim 10, wherein the garment comprises a boarded garment.