Optics: measuring and testing – Inspection of flaws or impurities
Reexamination Certificate
2001-10-16
2003-03-11
Font, Frank G. (Department: 2877)
Optics: measuring and testing
Inspection of flaws or impurities
C356S239400, C356S239100
Reexamination Certificate
active
06532064
ABSTRACT:
FIELD OF INVENTION
The invention relates to an automated inspection apparatus and method for use in the optical inspection of 3-dimensional (3D) translucent objects such as fish, to simultaneously identify surface, embedded and semi-transparent edge anomalies and distinguishing features.
BACKGROUND ART
Translucent objects, such as fish, often contain defects or conditions which lead to or cause contamination of the object. It is necessary to eliminate these, contaminated areas or objects and, as such, a reliable method for detecting these features or anomalies has been the subject matter of certain prior art.
Reliable detection of optical characteristics of a translucent object poses difficulties due to illumination, which is used during inspection, being either adsorbed by the object or scattered unevenly over the object. Conventional illumination can be too bright in certain areas or too dim in others. Another problem associated with detection of foreign objects is the creation of shadows or extreme brightness which results in an image capture device “stopping-up” or “stopping-down” the aperture to prevent under or overexposure.
Typical light sources range from conventional light fixtures to ultra violet spectrum. As an example, halogen lighting can be used to provide an overall light source for the inspection of an object such as a fish fillet. The use of UV is standard for the purposes of inspecting glass.
Surface, embedded and semi-transparent edge anomalies in translucent objects such as fish are characteristically only manifested under appropriate dark field or bright field illumination. As such, dark and bright field illumination are typically addressed individually due to light path propagation during compensation for individual manifestation characteristics of the anomalies.
For example, bright field lighting is a lighting technique which directs specular or diffuse reflections of light to the camera. Surface defects, such as blood and skin are detected with this technique. Dark field lighting is a lighting technique which directs back scattered light from the surface of an object to the camera. Embedded, anomalies, such as a parasite, and semi-transparent edge anomalies, such as transparent bones, in a fish fillet are detectable. with this technique. However, use of either technique presents known difficulties as summarized hereinafter. Changing light levels or the creation of geometrical shadows can distort accurate sensor detection.
Conventional methods used for anomaly detection for fish processing entail placing a fish on a light table having a surface illuminated. At least one lamp is used to illuminate the target area in bright field light. Operators visibly inspect a fish fillet to identify surface, embedded and semi-transparent edge anomalies.
Inherent to this method is visual fatigue, inconsistent visual perception by an operator and optical disparity between operators. Attempts have been made to overcome these difficulties by known art.
Examples of art that discuss several ways of properly illuminating and inspecting an object. include U.S. Pat. Nos. 4,585,315; 5,845,002; 6,022,124; 6,049,379; 5,493,123; international patents PCT/US95/11318 (equivalent to U.S. Pat. No. '002); PCT/US97/20058; Japanese patents 3165534A2; 8201222A2; and 11108637A2.
U.S. Pat. No. '315 discusses a bright field and dark field microscope illuminator with two axicon mirrors, a third plane mirror and shutters positioned in the paths of the light beams. Through use of the disclosed embodiments, simultaneous bright and dark field illumination or alternatively singularly bright or dark field illumination is achieved through the opening and closing of the appropriate shutters. However, the advantages of even illumination of a translucent object would technically obviate from detection of translucent or transparent anomalies that would manifest under variant light intensities. Simultaneous detection of surface, embedded and semitransparent edge anomalies is not possible with the use of only dark or bright field illumination.
U.S. Pat. No. '002 teaches a method of statistical evaluation of a translucent object by scanning graphic images of an object and processing incident light frequencies enabling a pixel by pixel analysis of topographic surface features of a fruit. A selected frequency, or combinations of frequencies, of light is directed at the fruit according to porosity of the peel. A computerized optical scanner having two light sources disposed at approximately 120 degrees from a vertical plane emit incident light towards an object, or to mirrors, to effectively scatter incident light within the fruit and cause the fruit to “glow”. A frequency spectrum is selected based on maximum, minimum and standard deviation of the intensity of the entire pixel pattern constituting the image. Both hemispheres of a fruit are analyzed; an algorithm assists in eliminating portions of the graphic information that are not relevant (such as bright field illumination or reflected light sources), and which do not constitute “glow” from the fruit. Sharp transitions are evaluated by filtering the image and comparing aberrations pixel by pixel. International application WO 96/14169 is derived from U.S. Pat. No. '002. Application of the above methodology prohibits bright field illumination techniques that enable detection of transparent anomalies such as skin and bones embedded or on the surface of a fish fillet.
U.S. Pat. No. '124 discloses a ring-light source and reflective ring focusing element wherein LED's are strung in one or more circular rows and strobe (or pulse) light to ring reflectors provides uniform lighting of an object. Light emissions from the LED's approach the object at an angle oblique or perpendicular to the optical axis. The effective dark field illumination patterns are intended to minimize light from the illumination source from entering the camera. However, use of only a dark field technique during inspection of a translucent object, such as a fish, severely diminishes the manifestation of those anomalies, thus requiring bright field techniques.
U.S. Pat. No. '379 shows a method of scanning multiple images of a translucent object and applying brightness ratios to the scanned images for the detection of flaws in the target area. A glass bottle is disclosed wherein light intensity readings are analyzed based on an acceptable range of ratios determined by a target area and a control area. The allowable range is calculated by analyzing an object with known defects and an object that is known-to be non-defective. The disclosure notes that “blind spots” may arise during inspection for defects.
The techniques of U.S. Pat. No. '123 involves the use of ultra violet radiation during a predetermined inspection period of glass. Use of ultra violet radiation during the inspection of organic materials for consumption is not desirable.
International patent PCT/US97/20058 discloses an automated inspection system with bright field and dark field illumination. The detection of “macro-defects” such as scratches, incomplete photoresist coverage and non-uniform edge bead removal on a semiconductor wafer are detectable through the use of simultaneous bright and dark field illumination. Image data acquisition is achieved through illumination of an object by at least two light sources wherein light striking the patterned surface of a wafer respectively propagates dark and light field light paths which are collected by an imaging lens. Light sensitive sensors are positioned behind the lenses which concentrate the light passing through them on the light receiving surfaces of the sensors. Data captured by the light sensors is output in a form of digital data streams. The streams of digital data are processed for creation of a gray level deviation map from which an absolute difference image is used to detect bright field and dark field, defects. Although the above addresses nanometer topographical defects, subtleties of light degradation impede detection of tran
Hearn Paul M.
Reimer Ernest M.
Baader-Canpolar Inc.
Fincham Ian
Font Frank G.
McFadden Fincham
Nguyen Sang H.
LandOfFree
Automatic inspection apparatus and method for simultaneous... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Automatic inspection apparatus and method for simultaneous..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Automatic inspection apparatus and method for simultaneous... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3040410