Defective object inspection and separation system using image analysis and curvature transformation

• US 5,732,147 A
• Filed: 06/07/1995
• Issued: 03/24/1998
• Est. Priority Date: 06/07/1995
• Status: Expired due to Fees

• Current Assignee: Aqri-Tech, Inc., Grantway LLC
• Original Assignee: Agri-Tech, Inc.
• Inventors: Tao, Yang
• Primary Examiner(s): JOHNS, ANDREW W
• Application Number:
  US08/483,962
• Time in Patent Office: 1,021 Days
• Field of Search: 382/110, 382/154, 382/173, 382/203, 382/256, 382/257, 382/274, 382/276, 348/89, 348/91, 209/576, 209/577, 250/559.45, 250/559.46, 356/237, 356/376, 395/126, 395/127
• US Class Current: 382/110
• CPC Class Codes: B07C 5/3422 using video scanning devices, e.g. TV-cameras

Abstract

Image processing system using cameras and image processing techniques to identify undesirable objects on roller conveyor lines. The cameras above the conveyor capture images of the passing objects. The roller background information is removed and images of the objects remain. To analyze each individual object accurately, the adjacent objects are isolated and small noisy residue fragments are removed. A spherical optical transform and a defect preservation transform preserve any defect levels on objects even below the roller background and compensate for the non-lambertian gradient reflectance on spherical objects at their curvatures and dimensions. Defect segments are then extracted from the resulting transformed images. The size, level, and pattern of the defect segments indicate the degree of defects in the object. The extracted features are fed into a recognition process and a decision making system for grade rejection decisions. The locations in coordinates of the defects generated by a defect allocation function are combined with defect rejection decisions and user parameters to signal appropriate mechanical actions such as to separate objects with defects from those that are defect-free.

Claims

1. A method of identifying defective objects from among a plurality of objects using an image processing system that acquires images of the plurality of objects, the method comprising the steps of:

generating for each of the plurality of objects a plane image;

performing a curvature transform to correct each of the plane images to compensate for varying reflectance levels, thereby forming corrected plane images;

determining ones of the objects that potentially contain defects from the corrected plane images;

separating portions of each of the corrected plane images corresponding to objects that potentially contain defects into object portions and defect portions; and

applying a predetermined threshold to the defect portions to determine whether the corresponding objects constitute defective objects.

2. A method performed by an image processor for grading defective objects, the method comprising the steps of:

acquiring an image of a plurality of objects;

performing a curvature transform on the image to correct the image for differences in gradation caused by differences in light reflectance of the objects;

locating, within the corrected image, defect segments based on differences in gradation caused by differences in light reflectance of the defect segments;

separating the defect segments from normal surfaces of the objects using thresholding on the corrected image; and

assigning grades to the objects corresponding to the defect segments based on characteristics of the defect segments.

3. The method of claim 2 wherein the objects are on a conveyor and the acquiring step includes the substep of:

filtering from the image, pixel data corresponding to the conveyor.

4. An image processor comprising:

means for receiving an image of a plurality of objects;

means for transforming the image into a corrected image, correcting for differences in gradation caused by differences in light reflectance of the objects;

means for locating, within the corrected image, defect segments based on differences in gradation caused by differences in light reflectance of the defect segments; and

means for grading the objects having the defect segments based on characteristics of the defect segments.

5. The image processor of claim 4 further comprising:

means for generating signals to separate the objects having the defect segments based on the grade assigned by the grading means.

6. The image processor of claim 4 wherein the receiving means includes

means for acquiring multiple side-images of the objects as the objects progress through a rotation.

7. A method performed by an image processor for detecting defective objects, the method comprising the steps of:

acquiring an image of an object;

performing a curvature transform on the image to correct the image for differences in gradation caused by differences in light reflectance of the object; and

detecting a defect in the object using the corrected image.

8. The method of claim 7 wherein the detecting step includes the substep of:

locating, within the corrected image, defect segments based on differences in gradation caused by differences in light reflectance of the defect segments.

9. The method of claim 8 further comprising the step of:

assigning a grade to the object corresponding to the defect segments and based on characteristics of the defect segments.

10. A method performed by an image processor for identifying a defect in an object, the method comprising the steps of:

receiving an image of the object;

performing a curvature transform to correct the image to compensate for curvature of the object; and

locating, within the corrected image, the defect in the object.

11. A method performed by an image processor for identifying a defect in an object, the method comprising the steps of:

receiving a pixel image of the object;

identifying a contour of the object from the pixel image;

performing a curvature transform to correct the pixel image to compensate for the contour of the object; and

locating the defect within the corrected pixel image.

12. A method performed by an image processor for identifying a defect in an object using a pixel image of the object, the method comprising the steps of:

identifying a contour of the object from the pixel image;

performing a curvature transform to correct the pixel image to compensate for the contour of the object;

segmenting at least one pixel of the corrected pixel image, which pixel corresponds to the defect; and

applying a threshold to the pixel that corresponds to the defect to thereby identify the defect.

13. A method performed by an image processor for preserving a defect identified in an image including an object, the image being stored in a memory, the method comprising the steps of:

generating a binary image of the image stored in the memory, the binary image having a first value assigned to background and defect pixels of the image and a second value assigned to object pixels of the image;

creating a dilated image of the object, the dilated image having the second value assigned to the background pixels and the first value assigned to the object and defect pixels, and storing the dilated image in the memory; and

combining the binary image with the dilated image to differentiate the defect in the image.

14. A defect preservation apparatus for detecting potential defect data in video data, comprising:

means for receiving the video data;

means for generating binary image data from the received video data, said binary image data having a first value assigned to background and defect portions of the video data and a second value assigned to object portions of the video data;

means for performing a multiple-pass dilation function on the video data, using a plurality of masks, to generate a dilated image in which said second value is assigned to the background portions and said first value is assigned to the object and defect portions; and

means for combining the binary image with the dilated image to detect the potential defect data.

15. An article of manufacture comprising a computer usable medium having computer readable program code means embodied therein for detecting defective objects, the computer readable program code means in the article of manufacture comprising:

computer readable program code means for causing a computer to acquire an image of an object;

computer readable program code means for causing the computer to perform a curvature transform on the image to correct the image for differences in gradation caused by differences in light reflectance of the object; and

computer readable program code means for causing the computer to detect a defect in the object using the corrected image.

16. The article of manufacture of claim 15 wherein the computer readable program code means for causing the computer to detect the defect includes:

computer readable program code means for causing the computer to locate, within the corrected image, a defect segment based on differences in gradation caused by differences in light reflectance of the defect segment.

17. The article of manufacture of claim 16 further comprising:

computer readable program code means for causing the computer to assign a grade to the object corresponding to the defect segment and based on characteristics of the defect segment.

18. An article of manufacture comprising a computer usable medium having computer readable program code means embodied therein for identifying a defect in an object using a pixel image of the object, the computer readable program code means in the article of manufacture comprising:

computer readable program code means for causing a computer to identify a contour of the object from the pixel image;

computer readable program code means for causing the computer to perform a curvature transform to correct the pixel image to compensate for the contour of the object;

computer readable program code means for causing the computer to segment at least one pixel of the corrected pixel image, which pixel corresponds to the defect; and

computer readable program code means for causing the computer to apply a threshold to the pixel that corresponds to the defect to thereby identify the defect.

19. A method for identifying a defect in an object of a plurality of objects using an image processing system that acquires an image of the object, the acquired image including an object image and a background image, the method comprising the steps of:

separating the object image from the background image in the acquired image;

creating a series of rings of the object image to create a contour image, each of the rings relating to a different intensity level of the object due to the object'"'"'s varying reflectance levels;

converting the contour image to a binary image;

forming an inverse image of the binary image; and

identifying the defect in the object by adding the inverse image to the contour image.

20. The method of claim 19, wherein the inverse image forming step includes the substeps of

setting the intensity level for each of the rings to a different uniform level, thereby eliminating any defect from the binary image, and

inverting the intensity level for each of the rings of the binary image.

21. A method for determining the contour of an object using an image processing system that acquires an image of the object, the acquired image including an object image and a background image, the method comprising the steps of:

separating the object image from the background image in the acquired image;

creating a series of rings of the object image, each of the rings relating to a different intensity level of the object due to the object'"'"'s varying reflectance levels;

converting the rings to a binary image;

forming an inverse image of the binary image; and

combining the inverse image with the binary image to determine the contour of the object.