Method and apparatus for inspecting the cleanliness of top slibers

• US 5,345,515 A
• Filed: 03/18/1992
• Issued: 09/06/1994
• Est. Priority Date: 05/28/1990
• Status: Expired due to Fees

• Current Assignee: Kanebo, Ltd. (HOYU Company Limited)
• Original Assignee: Kanebo, Ltd. (HOYU Company Limited)
• Inventors: Takayama, Shinichi, Muto, Tadashi, Nishi, Noriyuki
• Primary Examiner(s): Boudreau, Leo H.
• Assistant Examiner(s): KELLEY, CHRISTOPHER S
• Application Number:
  US07/820,686
• Time in Patent Office: 902 Days
• Field of Search: 382/1, 382/8, 356/238, 358/106, 358/107, 364/470
• US Class Current: 382/111
• CPC Class Codes: D01G 31/00 Warning or safety devices, e.g. automatic fault detectors, stop motions safety devices of general application F16P; indicating devices of general application G08B
• CPC Class Codes: D01G 31/003 Detection and removal of impurities
• CPC Class Codes: D01G 31/006 On-line measurement and recording of process and product parameters
• CPC Class Codes: G01N 15/1468 with spatial resolution of the texture or inner structure of the particle
• CPC Class Codes: G01N 21/8915 non-woven textile material
• CPC Class Codes: G01N 21/94 Investigating contamination, e.g. dust G01N21/85 takes precedence
• CPC Class Codes: G01N 33/362 material before processing, e.g. bulk cotton or wool
• CPC Class Codes: G06T 2207/30124 Fabrics; Textile; Paper
• CPC Class Codes: G06T 7/0006 using a design-rule based approach
• CPC Class Codes: G06T 7/64 of convexity or concavity

Abstract

A method of inspecting cleanliness of top sliber. The method classifies detected defects into pillwise defects and vegetal defects by introducing roundness, slenderness, gradation, dispersion, area, shape, gradation ratio, and length discernment against image signals of inspected top slibers picked up by an image sensor. An apparatus for inspecting cleanliness of top sliders is utilized. The apparatus initially picks up an image of uniformly spread top slibers conveyed to the apparatus and then processes the image signal with an image processing unit provided therefor. The method involves classifying defects on the top sliber into pillwise defects and vegetal defects in the top sliber based on the generated image signal. The classifying involves the steps of (i) establishing conditional section membership functions for features of defects, (ii) establishing conclusion section membership functions for determining a probability of being a defect against said classified defects, (iii) establishing a fuzzy rule for relating said features of defects with said probability of being a defect against the defects; and (iv) classifying the defects based on fuzzy reasoning derived from the conditional section membership functions, the conclusion section membership functions, and said fuzzy rule.

Claims

1. A method of inspecting the cleanliness of top sliber, comprising the steps of:

(a) providing a generally planar piece of top sliber;

(b) providing an electromagnetic wave source proximate said piece of sliber;

(c) providing an image sensor means for sensing said electromagnetic wave source and for receiving an image of the top sliber;

(d) generating an image signal of said tip sliber with the image sensor;

(e) based on said image signal computing (i) gradation discernment of defects and (ii) gradation ratio discernment of the defects;

(f) classifying the defects into the categories of pillwise defects and vegetal defects mixed in said top sliber based on said computing in step (e)

(g) computing a length discernment of the defects; and

(h) further classifying said categories of pillwise defect and vegetal defect into respective defects based on said computing of length discernment of step (g).

2. A method of inspecting the cleanliness of top sliber comprising the sequential steps of:

(a) generating an image signal by receiving an image of a top sliber with an image sensor;

(b) based on said image signal, computing;

(i) a roundness discernment of a defect on the top sliber, (ii) a slenderness discernment of the defect, (iii) a gradation discernment of the defect, and (iv) a dispersion discernment of the defect;

(c) classifying the defect into the categories of pillwise defects and vegetal defects mixed in said top sliber based on each of said roundness discernment, slenderness discernment, gradation discernment, and dispersion discernment of step (b).

3. A method of inspecting the cleanliness of top sliber, comprising the sequential steps of:

(a) generating an image signal by receiving an image of top sliber with an image sensor;

(b) based on said image signal, computing;

(i) a roundness discernment of a defect on the top sliber, (ii) a slenderness discernment of the defect, (iii) a gradation discernment of the defect, and (iv) a dispersion discernment of the defect;

(c) classifying the defect into the categories of pillwise defects and vegetal defects mixed in said top sliber based on each of said roundness discernment, slenderness discernment, gradation discernment, and dispersion discernment; and

classifying said category of pillwise defect further into slab, nep, and pinpoint defects by applying area discernment, and classifying said category of vegital defect further into trefoil bar, impurities and bar by applying shape discernment, and further classifying said impurities and bar by applying length discernment.

4. A method of inspecting the cleanliness of top sliber, comprising the sequential steps of:

(a) providing a generally planar piece of top sliber;

(b) providing an electromagnetic wave source proximate said piece of sliber;

(c) providing an image sensor means for sensing said electromagnetic wave source and receiving an image of the top sliber;

(d) generating an image signal of the top sliber with the image sensor;

(e) classifying defects on the top sliber into pillwise defects and vegetal defects mixed in said top sliber based on said generated image signal;

(f) said step of classifying includes the steps of (i) establishing conditional section membership functions for features of defects, (ii) establishing conclusion section membership functions for determining a probability of being a particular defect from said defects, (iii) establishing a fuzzy rule for relating said features of defects with said probability of being a particular defect from said defects; and

(iv) classifying the defects based on fuzzy reasoning derived from said conditional section membership functions, said conclusion section membership functions, and said fuzzy rule.