Product appearance inspection methods and apparatus employing low variance filter

• US 5,537,670 A
• Filed: 08/24/1994
• Issued: 07/16/1996
• Est. Priority Date: 08/08/1991
• Status: Expired due to Term

• Current Assignee: Philip Morris USA Incorporated (Altria Group, Inc.)
• Original Assignee: Philip Morris USA Incorporated (Altria Group, Inc.)
• Inventors: Maher, Robert J., Dante, Henry M., Cox, Kenneth A.
• Primary Examiner(s): Mancuso, Joseph
• Assistant Examiner(s): NOT, DEFINED
• Application Number:
  US08/295,185
• Time in Patent Office: 692 Days
• Field of Search: 382/8, 382/9, 382/30, 382/33, 382/34, 382/14, 382/141, 382/143, 382/159, 382/209, 382/217-221, 356/237, 348/129, 348/130
• US Class Current: 382/219
• CPC Class Codes: G06T 2207/20036 Morphological image processing
• CPC Class Codes: G06T 2207/30144 Printing quality
• CPC Class Codes: G06T 2207/30164 Workpiece; Machine component
• CPC Class Codes: G06T 7/001 using an image reference approach

Abstract

Images such as product images are identified as substantially similar to one or more reference images by finding pixels which have substantially the same value or values in all or substantially all of the reference images. Image erosion and/or dilation may be used in processing the reference image data to help identify pixels which can always be expected to have the same value in all acceptable images. The values associated with the corresponding pixels in a product image are combined and compared to an expected value. The values associated with pixels which correspond to regions not similar in substantially all of the reference images are combined to produce a discriminant function. The product image is identified as substantially similar to the reference image or images only if the combined values from the product image compare favorably with the expected value, and the dot product between the product image and the discriminant function is equal to a predetermined value.

Claims

1. A system for determining whether a particular product has a predetermined appearance comprising:

a filter;

first means for defining a low variance region in said filter from a first region of each of a plurality of training products having said predetermined appearance, each first region of the plurality of training products comprising subdivisions, each one of the subdivisions of each first region of the plurality of training products being substantially free of variation from a set of corresponding subdivisions of each of the remaining training products in a predetermined image characteristic;

means for identifying a first region comprising subdivisions of the particular product corresponding to the first region of the plurality of training products;

means for defining a composite value of the first region of the particular product, the composite value defined as a composite of numerical values associated with the subdivisions of the first region of the particular product and indicative of a predetermined image characteristic;

means for detecting an amount of variation by which (a) said composite value of said particular product deviates from (b) a predetermined composite value and for identifying said particular product as not having said predetermined appearance if said amount of variation exceeds a predetermined value;

second means for defining a high variance region in said filter from a second region having said predetermined appearance, each second region of the plurality of training products comprising subdivisions not identified in the first reason of the plurality of training products, substantially every one of the subdivisions of each second region of the plurality of training products being subject to high variation from the set of corresponding subdivisions of each of the remaining training products in said predetermined image characteristic;

means for defining discriminant values associated with the second reason of the training products, the discriminant values defined as a combination of numerical values associated with the subdivisions of the second region of the training products and indicative of a predetermined image characteristic;

means for computing a dot product between the defined discriminant values and numerical values associated with the subdivisions of the second region of the particular product and indicative of a predetermined image characteristic; and

means for detecting a measure by which said dot product deviates from a predetermined dot product value range and for identifying said particular product as having said predetermined appearance if said measure is within a predetermined range.

2. The system defined in claim 1 further comprising:

a video camera; and

means for positioning said particular product in the field of view of said video camera so that the video camera can capture an image of said particular product for processing by said system.

3. The system defined in claim 2 further comprising:

means for illuminating said particular product in the field of view of said camera.

4. The system defined in claim 2 wherein said means for positioning comprises:

a conveyor for conveying particular products one after another through the field of view of said video camera.

5. The system defined in claim 4 wherein said means for positioning further comprises:

means for detecting when a particular product on said conveyor is at a predetermined location in the field of view of said video camera; and

means responsive to said means for detecting for utilizing the output signal of said video camera concurrent with detection that a particular product is at said predetermined location as said image of said particular product.

6. The system defined in claim 5 wherein said means for positioning further comprises:

means responsive to said means for detecting for momentarily illuminating said particular product at said predetermined location.

7. Apparatus for determining whether a particular image, comprising a plurality of particular pixels, has a predetermined appearance, said apparatus comprising:

means for identifying a first region of pixels in a plurality of training images, each training image having the predetermined appearance and comprising a plurality of pixels, the first region identified such that every one of the pixels in the first region is substantially identical to each corresponding pixel in each of the training images;

means for identifying a first region in the particular image comprised of particular pixels corresponding to the pixels of the first region of said training images;

means for associating a numerical value indicative of a predetermined image characteristic with each of the corresponding pixels of the first region of the particular image;

means for combining these associated numerical values to produce a composite value of the pixels of the first region of the particular image; and

means for determining whether the particular image does not have the predetermined appearance if an amount of deviation of the produced composite value from a predetermined composite value exceeds a predetermined value;

means for identifying a second region in said training images, the second region comprised only of pixels of said training images not identified in the first region of said training images;

means for identifying a second region in the particular image comprised of particular pixels corresponding to the pixels of the second region of said training images;

means for associating a numerical value indicative of a predetermined image characteristic with each of the pixels Of the second region of said plurality of training images and with each of the pixels of the second region of the particular image;

means for combining the numerical values associated with the pixels of the second region of said plurality of training images to produce discriminant values associated with the second region of said plurality of training images;

means for computing a dot product between the discriminant values and the numerical values associated with corresponding pixels of the second region of the particular image; and

means for determining whether the particular image has the predetermined appearance if this computed dot product is within a predetermined dot product range.