Methods of detecting, classifying and quantifying defects in optical fiber end faces

• US 5,179,419 A
• Filed: 11/22/1991
• Issued: 01/12/1993
• Est. Priority Date: 11/22/1991
• Status: Expired due to Fees

• Current Assignee: Fitel USA Corporation
• Original Assignee: AT&T, Inc.
• Inventors: Schmidt, Arthur T., Shahraray, Behzad, Palmquist, John M.
• Primary Examiner(s): McGraw, Vincent P.
• Application Number:
  US07/796,204
• Time in Patent Office: 417 Days
• Field of Search: 356/73.1, 356/237, 382/1, 382/8
• US Class Current: 356/73.1
• CPC Class Codes: G01N 21/952 Inspecting the exterior surface of cylindrical bodies or wires G01N21/956 takes precedence
• CPC Class Codes: G02B 6/3843 with auxiliary facilities for movably aligning or adjusting the fibre within its ferrule, e.g. measuring position or eccentricity testing the alignment of axes, including eccentricity, G01B11/27
• CPC Class Codes: G02B 6/385 Accessories for testing or observation of connectors means for centering or aligning the light guide within the ferrule with auxiliary facilities for movably aligning or adjusting the fibre within its ferrule, measuring position, eccentricity G02B6/3843; mechanical features associated with the optical testing of optical fibres G01M11/088
• CPC Class Codes: G02B 6/3863 fabricated by using polishing techniques grinding of the fibre ends B24B19/226

Abstract

Methods are provided to detect, classify and quantify defects such as chips, pits, scratches and cracks in a polished end surface (31) of optical fiber and specifically in an end face of an optical fiber terminated by a ferrule (34). Images of the end face are acquired at each of three focal positions which collectively include all the features of interest. Information from these images is combined into a single image which is processed further. Discrepancies between the images are used to discriminate between cracks and scratches. Morphological processing is used to segment the fiber from its ferrule and a Hough transform is used to estimate the center and radius of the optical fiber, which facilitates the isolation of the fiber. Chips and pits in the fiber end face are detected and quantified by thresholding and morphological processing of the isolated fiber. Edge detection is used to detect edge segments resulting from scratches and cracks. Then the line segments are classified into scratches and cracks. The detected segments are matched to provide scratches in a final image. Quantitative measures then are used to establish standards for the quality of the polished, terminated optical fibers.

Claims

1. A method of detecting, classifying and quantifying defects in a polished end face of an optical fiber, said method comprising the steps of:

acquiring images of an end face of a fiber at an optimum focal position and at auxiliary positions behind and in front of said optimum focal position;

generating a combined image of the fiber end face which depicts any features of interest which exist only in the end face of the optical fiber;

identifying features of interest which exist only in the end face of the optical fiber; and

comparing quantitative data generated for each of the features of interest with acceptable values to determine the acceptability of the polished end face of the optical fiber.

2. A method of detecting, classifying and quantifying defects in a polished end face of an optical fiber, which is terminated in a connector component, said method comprising the steps of:

acquiring images of an end face of a fiber, which has been terminated by a connector component, at an optimum focal position and at auxiliary positions behind and in front of said optimum focal position;

generating a combined image of the terminated optical fiber end face which depicts any feature of interest which exist only in said terminated end face;

isolating the end face of the optical fiber from the materials of the connector component which terminates the fiber to provide an image which depicts any features of interest which exist only in the end face of the optical fiber;

identifying the features of interest which exist only in the end face of the optical fiber; and

comparing quantitative data generated for each of the features of interest with acceptable values to determine the acceptability of the terminated optical fiber.

3. A method of detecting, classifying and quantifying defects a polished end face of a terminated optical fiber, said method comprising:

generating a focus function of the end face of a ferrule in which the optical fiber is terminated;

determining a location of the focus function where the focus of the terminated end face of the fiber is optimum;

acquiring a primary image of the end face of the fiber at the optimum focus location;

acquiring auxiliary images of the end face of the fiber behind and in front of the optimum focus location;

storing the acquired primary and auxiliary images;

providing a difference image which is the absolute value of the difference between the auxiliary images;

thresholding the difference image;

refining the difference image to discard portions of the difference image which result from minor variations between the two auxiliary images;

smoothing the refined difference image to change binary lines in said refined difference image to lines having a generally gradually varying distribution;

comparing pixels between the primary image and smoothed difference image to provide a combined image which depicts any scratches and cracks in the fiber end face;

isolating the fiber portion of the primary image from materials of the ferrule to provide an image which depicts any feature of interest which exists only in the end face of the optical fiber;

detecting and quantifying chips and pits in the isolated combined image;

detecting and determining edge strength and direction of lines in said isolated combined image;

identifying cracks which exist in the end face of the optical fiber and which appear in said isolated combined image;

tracing edges in the isolated combined image and decomposing them into segments;

matching line segments in the isolated combined image; and

comparing quantitative data generated for each of the features of interest with acceptable values to determine the acceptability of the terminated optical fiber.

4. The method of claim 3, wherein said method includes the steps of:

acquiring an image of the end face of the terminated optical fiber;

analyzing the high frequency content of each of a plurality of subregions of the image;

verifying a distribution of the focus function and adjusting the focus, if necessary, to establish an optimum, primary focus position;

acquiring an image at a first auxiliary position spaced from said optimum focus position; and

acquiring an image at a second auxiliary postition spaced from the first auxiliary position with the primary focus position therebetween.

5. The method of claim 3, wherein said step of providing a combined image includes the step of;

comparing the smoothed difference image with the primary image to provide a combined image each pixel of which is equal to the greater of the corresponding pixels in the smoothed and primary images.

6. The method of claim 3, wherein said step of isolating the fiber includes the steps of:

thresholding the combined image;

morphologically opening the thresholded primary image to separate a ring of adhesive material, which is disposed between said optical fiber and its terminating ferrule to hold the fiber in the ferrule, into discrete areas;

filtering the discrete areas of adhesive material and discarding the discrete areas which are less than a predetermined size;

estimating the center of the end face of the optical fiber;

determining the radius of the end face of the optical fiber; and

defining a circular mask to isolate the end face of the optical fiber.

7. The method of claim 3, wherein said step of detecting chips and pits includes the steps of:

thresholding the portion of the combined image which falls within the isolated optical fiber end face to provide an image which contains chips and pits;

morphologically closing the image to merge adjacent ones of the chips and pits;

morphologically opening the morphologically closed image to remove noise and relatively small components;

identifying individual chips and pits; and

quantifying the chips and pits which appear in the end face of the optical fiber.

8. The method of claim 3, wherein said step of detecting and determining edge strength and direction of lines includes the steps of:

stretching the intensity of the isolated portion of the combined image to provide an image in which features appearing in the fiber end face are enhanced;

smoothing the enhanced image to reduce noise and to shape the intensity profile of edges in the enhanced image;

computing derivatives in horizontal, vertical and diagonal directions;

determining the strength and direction of edges in the smoothed enhanced image to provide edge strength and direction images; and

thinning edges to have a width of a single pixel.

9. The method of claim 8, which also includes tracing and decomposing of edges which includes the steps of:

thresholding the edge strength image to discard edges having a strength less than a predetermined value;

labeling connected components from said edge strength image;

thresholding the labeled component image to remove noise and relatively short segments; and

identifying cracks from the thresholded images by determining the correspondence with non-zero pixels in the binary difference image.

10. The method of claim 9, wherein said lines which appear in said thresholded, labeled component image and which are not cracks are processed to match line segments and to identify scratches.

11. The method of claim 10, wherein the step of line matching includes the step of providing a matrix which for the pairing of each two lines which appear in said thresholded image, a matching coefficient is provided.

12. The method of claim 11, wherein a corrected matching coefficient which is equal to the product at the matching coefficient and a correction factor, which provides an indication of the separation between line segments, is used in matching line segments.

13. The method of claim 12, wherein said step of line matching comprises an iterative process.

14. The method of claim 13, wherein a final image includes chips, pits, scratches and cracks of interest which are identified.

15. The method of claim 14, wherein any defects which appear in the final image are quantified and the quantification compared to predetermined values to decide whether or not the polished end face of the terminated optical fiber is acceptable.