Surface analysis system and method

• US 5,809,162 A
• Filed: 12/22/1995
• Issued: 09/15/1998
• Est. Priority Date: 12/22/1995
• Status: Expired due to Term

• Current Assignee: Lucent Technologies, Inc. (Nokia Corporation)
• Original Assignee: Lucent Technologies, Inc. (Nokia Corporation)
• Inventors: Csipkes, Andrei, Mufti, Muid Ur-Rehman, Palmquist, John Mark
• Primary Examiner(s): Mancuso, Joseph
• Assistant Examiner(s): Bobys, Marc
• Application Number:
  US08/577,947
• Time in Patent Office: 998 Days
• Field of Search: 65/376-8, 250/559.01, 250/559.04-559.06, 250/559.08, 250/559.4, 250/559.42, 250/559.44-559.48, 348/125-134, 356/73.1, 356/239, 356/240, 382/141-150, 382/152, 382/173, 382/180, 382/254, 382/260, 382/266, 382/276, 382/307, 382/308, 382/108
• US Class Current: 382/108
• CPC Class Codes: G02B 6/3807 Dismountable connectors, i.e. comprising plugs
• CPC Class Codes: G02B 6/3818 of a low-reflection-loss type G02B6/3813 takes precedence

Abstract

A surface analysis system can contactlessly, automatically, and rapidly detect, classify, and evaluate a surface of an object, particularly, an optical fiber end face, for discontinuities to derive a single pass/fail conclusion regarding the surface. The surface analysis system has a scope for capturing an image of the end face. A computer is connected to the scope. A machine vision system is associated with the computer for receiving the image. A surface analysis program is associated with the computer for driving the machine vision system. The program searches for any discontinuities in the image by analyzing each pixel and a corresponding pixel structure of pixels to determine whether a discontinuity resides at each pixel. Discontinuities are classified as one of the following: binary thresholds, local gradients, and directional gradients. The pixel structure includes a plurality of pixels that were previously analyzed. The program determines whether the surface is continuous based upon any discovered discontinuities.

Claims

1. A method for analyzing a surface of an optical fiber end face, comprising the steps of:

(a) producing a digitized image of said optical fiber end face, said image having a plurality of pixels;

(b) searching for discontinuities in said image by successively analyzing a plurality of contiguous pixels, said analysis at each said pixel including examining a pixel structure corresponding with each said pixel to determine whether a discontinuity resides at each said pixel, said pixel structure including a plurality of pixels that were previously analyzed;

(c) analyzing structure memory elements that are contiguous and that are situated contiguous with a pixel that is currently analyzed, and referred to hereafter as a structure working element;

(d) analyzing satellite elements that are contiguous with at least one of said memory elements and that are noncontiguous with respect to said structure working element;

(e) determining whether said structure working element is a local gradient and a directional gradient based upon said analyses of said satellite elements and said structure memory elements, wherein;

said local gradient is a pixel having an intensity that differs by at least a first predefined amount from intensities associated with surrounding pixels, wherein said local gradient corresponds with a surface texture discontinuity

said directional gradient is a plurality of pixels having intensities that differ by at least a second predefined amount from intensities associated with pixels situated about said directional gradient, wherein said directional gradient corresponds with a scratch;

(f) combining intensities of said satellite elements to derive a satellite element average;

(g) comparing said satellite element average to said structure working element;

(h) when said average is unequivalent to said structure working element within a first predefined threshold, then identifying said structure working element as a tentative gradient and performing the following steps;

(1) comparing each said structure memory element to said structure working element;

(2) when said structure memory element is equivalent to said structure working element as defined by a second predefined threshold, then identifying said structure working element as a directional gradient and identifying said structure memory element as a directional gradient if not already identified as such;

(3) when said structure memory element is unequivalent to said structure working element as defined by said second predefined threshold, then identifying said tentative local gradient as a local gradient; and

(i) when said average is equivalent to said structure working element as defined by said first predefined threshold, then identifying said structure working element as a nongradient.

2. The method of claim 1, further comprising the step of scanning said pixels of said image with said pixel structure in a raster format.

3. The method of claim 1, further comprising the step of analyzing each said pixel only a single time during said searching step.

4. The method of claim 1, further comprising the step of searching for any binary thresholds in said image, said binary threshold being a plurality of pixels having intensities that differ by at least a third predefined amount from intensities associated with an average of pixel intensities in said image.