Web inspection system

• US 20020109112A1
• Filed: 02/09/2001
• Published: 08/15/2002
• Est. Priority Date: 02/09/2001
• Status: Active Grant

• Current Assignee: Wintriss Engineering Corporation
• Original Assignee: Wintriss Engineering Corporation
• Inventors: Becker, Robin D., Guha, Sujoy D., Kiraly, Chris M.
• Granted Patent: US 6,750,466 B2
• Time in Patent Office: Days
• US Class Current: 250/559.46
• CPC Class Codes: G01N 2021/891 Edge discrimination, e.g. by signal filtering
• CPC Class Codes: G01N 21/89 in moving material, e.g. running paper or textiles G01N21/90, G01N21/91, G01N21/94 take precedence
• CPC Class Codes: G01N 21/8903 using a multiple detector array
• CPC Class Codes: G01N 21/8921 Streaks
• CPC Class Codes: G06T 2207/30124 Fabrics; Textile; Paper
• CPC Class Codes: G06T 7/0004 Industrial image inspection

Abstract

A web inspection system provides detection of web flaws along the machine direction and cross direction of a web. The detectable percent contrast between good web material and bad web material in one embodiment approaches noise level. The web inspection system utilizes a multiple of smart cameras connected to a host computer via an ethernet hub. Each smart camera includes a line scan camera for producing digital pixels, a means for lighting and pixel correction on a pixel by pixel basis, a web edge detector for monitoring the edges of a web, a multi-pipeline flaw detection pre-processor for detecting very small changes in the web material, a run length encoder for generating data regarding the location of each group of potential flaws in a cross direction, a 2D blob detector and analyzer for generating data regarding the location of block flaws along a machine direction, and an inspect/reject analysis for determining the actual flaw data from the potential flaw data. A low contrast web inspection system provides a balanced and distributed architecture that handles high defect rates, and that is easily integrated with an existing web manufacturing system.

Claims

1. A system for web inspection of a web, the system comprising:

a plurality of smart cameras, each smart camera for detecting a plurality of web flaws from a streaming video signal, each smart camera having means for generating flaw image data and flaw location data;

a host computer for controlling the low contrast web inspection system and for accepting and displaying the flaw image data and the flaw location data; and

an ethernet for connecting the plurality of smart cameras to the host computer.

2. The system of claim 1, wherein each smart camera of the plurality of smart cameras comprises:

a line scan camera for generating a pixel representation of a portion of the web;

a lighting uniformity and pixel sensitivity correction means for correcting each pixel of the pixel representation and for providing a corrected pixel representation;

a web edge detector for detecting at least one edge of the web;

a multi-pipeline pre-processor for filtering the corrected pixel representation, the multi-pipeline preprocessor generating a prioritized data stream of potential flaws;

a run length encoder for generating location data regarding a location of each group of the potential flaws in a cross direction;

a blob detector for generating block data regarding the location of blocks of the potential flaws along a machine direction; and

an inspect/reject analyzer for determining actual flaw data from the prioritized data stream of potential flaws.

3. The system of claim 2, wherein the multi-pipeline processor comprises:

a plurality of filters for averaging the corrected pixel representation over a distance of the web along a machine direction of the web;

a plurality of adaptive background subtraction channels connected to the plurality of filters;

a plurality of thresholders, each thresholder of the plurality of thresholders connected to an output of an adaptive background subtraction channel of the plurality of adaptive background subtraction channels, each thresholder for grouping a subtracted pixel representations; and

a priority logic circuit for prioritizing the outputs of each of the plurality of thresholders.

4. The system of claim 3, wherein the plurality of filters comprises:

a background filter;

a machine direction streak filter;

a cross direction streak filter; and

a small flaw filter.

5. The system of claim 3, wherein the plurality of thresholders comprises:

a single pixel flaw detector;

a uniformity detector;

a machine direction streak detector;

a cross direction streak detector; and

a small flaw detector.

6. The system of claim 1, wherein each smart camera of the plurality of smart cameras detects the plurality of web flaws from the streaming video signal at a contrast approaching a signal noise level.

7. A method for low contrast web inspection of a web, the method comprising the steps of:

providing at least on smart camera for inspecting at least a portion of the web;

generating flaw image data and flaw location data;

transmitting the flaw image data and flaw location data over an ethernet;

displaying the flaw image data and flaw location data.

8. The method of claim 7, wherein the step of generating flaw image data and flaw location data comprises the steps of:

generating a pixel representation of a portion of the web;

correcting the pixel representation for a lighting uniformity and a pixel sensitivity;

filtering the corrected pixel representation utilizing a plurality of filters;

grouping the filtered corrected pixel representations to generate a plurality of potential flaw data streams;

generating a prioritized data stream from the plurality of potential flaw data streams;

generating cross direction location data regarding a location of the prioritized data stream;

generating block data regarding the location of blocks of the prioritized data stream along a machine direction; and

determining actual flaw data from the prioritized data stream of potential flaws utilizing the cross direction location data and the block data.

9. A method for generating a prioritized image data stream from a digitized video stream of a web, the method comprising the steps:

averaging the digitized video stream over a distance of the web to generate an averaged background signal;

averaging the digitized video stream over a distance of the web along a machine direction of the web to generate a filtered machine direction signal;

averaging the digitized video stream over a distance of the web along a cross direction of the web to generate a filtered cross direction signal;

subtracting the averaged background signal from the filtered machine direction signal to generate a first pixel representation;

subtracting the averaged background signal from the filtered cross direction signal to generate a second pixel representation;

grouping the first and second pixel representations to generate at least two data streams of potential flaws; and

prioritizing the at least two data streams of potential flaws to generate the prioritized image data stream.