Method for optical testing of samples

• US 5,185,810 A
• Filed: 04/26/1991
• Issued: 02/09/1993
• Est. Priority Date: 04/26/1990
• Status: Expired due to Fees

• Current Assignee: Carl-Zeiss-Stiftung
• Original Assignee: Carl-Zeiss-Stiftung
• Inventors: Freischlad, Klaus
• Primary Examiner(s): Boudreau, Leo H.
• Assistant Examiner(s): KELLEY, CHRISTOPHER S
• Application Number:
  US07/692,194
• Time in Patent Office: 655 Days
• Field of Search: 356/345, 356/359, 356/376, 356/371, 382/8
• US Class Current: 382/100
• CPC Class Codes: G01B 11/2441 using interferometry
• CPC Class Codes: G06T 2207/30108 Industrial image inspection
• CPC Class Codes: G06T 7/0004 Industrial image inspection

Abstract

The optical testing method of the invention permits the reliable elimination of discontinuities, i.e., the determination of the integer values of modulo 2π phase differences, even when the measured data includes seriously noisy signals. The intensity signals from a two-dimensional raster of camera image points are used to compute differences in phase values between adjacent image points in lines, as well as between adjacent image points in columns. The phase differences are summed along a closed path around each of a plurality of partial fields which are delineated by a predetermined set of adjacent image points. A filter (21) searches the vector field of these phase differences; and those partial fields where vortexes exist, i.e., each field for which the sign-correct sum along its closed path is not zero, are masked, thereby creating holes in the image pattern. Additional closed paths are summed around these holes and, by masking further image points between these holes, the masked partial fields are combined until the path integral around each of these combined holes disappears. The final phase value data used for display of a phase map of the sample is calculated by path integration of the unmasked image points; and inconsistent measured values, e.g., due to noisy signals, do not corrupt the phase map.

Claims

1. In a method for reducing the effects of noise in the optical testing of samples in which information pertaining to the sample in the form of a pattern of light recorded by a camera is evaluated along a two-dimensional raster of image points, said method having the steps of:

using the brightness of the pattern to compute a corresponding phase value, expressed in modulo 2π

, for each image point;

calculating the differences, in modulo 2π

, between the phase values of adjacent image points in a line and between the phase values of adjacent image points in a column for said entire raster;

computing a first sign-correct sum of said modulo 2π

differences along a closed path around each of a plurality of partial fields, each such partial field being delineated by a predetermined set of adjacent image points; and

masking all image points in each partial field for which said sign-correct sum is not zero so that each said masked partial field forms a hole in said image pattern, said holes being omitted from further evaluation of said pattern image;

the improvement comprising the further steps of;

computing a second sign-correct sum of said modulo 2π

differences along a closed path delineated by image points positioned around each said hole, and, for those closed paths for which said second sign-correct sum is not zero,

masking the image points between said holes to combine said holes until a second sign-correct sum of said modulo 2π

differences along a closed path of image points positioned around said combined holes becomes zero.

2. The method of claim 1 wherein, when masking the image points between said holes to combine said holes, said final masking step comprises the further step of choosing the shortest path between said masked partial fields.

3. The method of claim 1 wherein said masking step further comprises masking any image point which is not part of at least one of said partial fields delineated by a predetermined set of adjacent image points.

4. The method of claim 1 wherein said second sum computing step comprises the further steps of:

computing a second sign-correct sum of said modulo 2π

differences along a closed path delineated by outermost unmasked image points of said two-dimensional raster, and, if said computed second sum for said outermost image points does not equal zero,

combining at least one of said outermost image points with at least one of said masked partial fields by masking the image points between said outermost image point and said masked partial field until a second sign-correct sum of said modulo 2π

differences along a closed path which includes image points positioned adjacent said combined partial field becomes zero.

5. The method of claim 1 comprising the further step of:

determining a final phase value for each unmasked image point by summing the modulo 2π

differences between the first computed phase values along a path of unmasked image points interconnecting each said unmasked image point with a common reference point.

6. The method of claim 5 comprising the further step of:

using said final phase value of said unmasked points to determine a final phase value for masked image points adjacent said unmasked points, by adding the modulo 2π

differences between the final phase value determined for each said unmasked point and said initially computed phase value for said adjacent masked point.

7. The method of claim 1 wherein said pattern of light recorded by the camera is an interferogram.