US 5,301,129 AGranted

Video web inspection system employing filtering and thresholding to determine surface anomalies

AssigneeAluminum Company of America

InventorsStephen V. McKaughan, Gary F. Nevers, Joseph W. Landry, John P. Fallon, Paul R. Adomaitis

Life of this patent

Estimating the term…

Loading patent…

Abstract

Method and apparatus for inspecting the surface of a moving web of homogeneous material for anomalies, the apparatus including a light source for illuminating the web surface and a light sensitive detector for receiving data from the web surface. The data is representative of the light reflected from the surface, as the light is affected by surface anomalies and background information. Computational means are provided for receiving output data from the detector in response to the data the detector receives from the web surface, and for processing the output data in a manner that determines the presence of surface anomalies. The computation means includes a plurality of filters for transforming the surface data in a manner that effects substantial elimination of the background information, and means for establishing threshold levels for background information. In addition, automatic means is provided for modifying the threshold levels in response to changes in the reflectance, frequency, or contrast of background information received by the detector and forwarded to the computation means.

Claims11

Dependent claims

In an inspection system for repeatedly and automatically determining the existence of anomalies on a surface of a moving web of non-discrete, homogeneous material, said anomalies existing in the presence of a web surface background that is not anomalous, the system comprising:

a light source for illuminating the web surface;
a light sensitive device for acquiring data from the web surface, and for producing an image of said surface data in the form of pixel elements, and for providing output data that is representative of the web surface data acquired by the light sensitive device, and surface data being representative of the light reflected from the surface to the light sensitive device, as the light is affected by surface anomalies;
and
computational means for receiving the output data from said light sensitive device in response to the data said light sensitive device receives from the web surface, and for processing said output data in a manner that determines the presence of surface anomalies,
said computational means including:
a plurality of multi-directional filters for transforming the web surface data in a manner that effects identification of web surface anomalies of different sizes and shapes and segregates the anomalies of different sizes and shapes from each other and from the web surface background by using edge detection or template matching of an anomaly and integration of anomaly length in the direction of web movement, said filters producing an output that is representative of said pixel elements;
means for establishing a threshold and applying the threshold to the output of said filters in a manner that separates the web surface background in the pixel elements from the web surface data in the pixel elements representing suspect and known anomalies;
and
means for dividing the image of the web surface data provided by the light sensitive device into subframes, the subframes containing pixel elements of varying intensities;
and
means for summing the varying intensities of the pixel elements within a subframe boundary in the direction of web travel to provide an integrated intensity value;
said integrated intensity value being indicative of the existence or non-existence of one or more surface anomalies.

9 more elements

A method of repeatedly and automatically determining the existence of anomalies on a surface of a moving web of non-discrete, homogeneous material, said anomalies existing in the presence of a web surface background that is not anomalous, the method comprising:

illuminating the surface of said web material;
locating a light sensitive detector to receive light reflected from said web surface in response to said illumination, said detector providing an image of the web surface in the form of pixel elements having values that represent web surface data of the moving web of material;
identifying anomalies that may exist on said web surface, and segregating anomalies of different sizes and shapes from each other and from the web surface background by subjecting said pixel element values to a series of computations that employ a plurality of multi-directional filters to effect edge detection or template matching of anomalies and integration of anomaly lengths in the direction of web travel, said filters providing an output that represents pixel element values having varying intensities;
providing a first threshold value and applying the first threshold value to the output of said filters in a manner that separates the web surface background from the web surface data representing anomalies;
and
dividing the image into subframes that have varying intensities;
summing the varying intensities of said subframes to provide an integrated intensity value;
using said integrated intensity value to determine whether or not an anomaly exists on the web surface;
and
automatically modifying the first threshold value in response to changes in the level of the web surface background received from the light sensitive device.

7 more elements

Dependencies are inferred from the claim text.

Facts

Application: 07/875,672Filed 29 Apr 1992 · continuation in part
Priority: 13 Jun 1990
Granted: 5 Apr 1994
Term ends: Not estimatedThe source records don’t support an estimate.

Classification8 codes

G06T 7/001
Industrial image inspection: using an image reference approach
G01N 21/89
Investigating the presence of flaws or contamination: in moving material, e.g. running paper or textiles
G01N 21/93
Investigating the presence of flaws or contamination: detection standards; Calibrating baseline adjustment, drift correction
G01N 2021/936
Detection standards; Calibrating baseline adjustment, drift correction: adjusting threshold, e.g. by way of moving average