Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Interferometric Video Autofocus

IP.com Disclosure Number: IPCOM000062790D
Original Publication Date: 1986-Dec-01
Included in the Prior Art Database: 2005-Mar-09
Document File: 2 page(s) / 14K

Publishing Venue

IBM

Related People

Glaser, I: AUTHOR [+3]

Abstract

In interferometric microscopy, the tolerances on focus error are much tighter than in normal microscopy. These tight tolerances increase the need for an autofocus capability. However, normal autofocus techniques are inadequate.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 2

Interferometric Video Autofocus

In interferometric microscopy, the tolerances on focus error are much tighter than in normal microscopy. These tight tolerances increase the need for an autofocus capability. However, normal autofocus techniques are inadequate.

To understand how one can use interferometric data to obtain the location of the "best" focus, it is useful to consider a three dimensional image space. Each x-y (fixed z0) section of this space represents the image obtained through the EXTRA interferometer at one focus position. The z axis is the "focus" or interferometric path difference. Thus, for a given coordinate pair (x0, y0) we get, as a function of z, a signal similar to that of a classical Michelson interferometer. However, acquiring and analyzing the complete 3-D is hardly practical - for a region 100mm thick and a visible wavelength of about 0.5mm, 1000 z slices are necessary, each containing, typically, 5122=262,144 pixels, or a total of N 2 x 108 pixels.

A number of subsets of this data set can be developed and used to determine the best focus. In particular, one can look at fringe contrast in a number of successive images at different focus points. This corresponds to studying horizontal slices in the above-mentioned data set. A second alternative is to look at vertical columns of the data set. In either case, one has to determine the location of maximum fringe contrast to identify the optimal focus. The vertical fringe approach has several advantages and will be explained in more detail.

If we select a fixed point on the x-y plane, and look at the z distribution of the intensity, we get a pattern similar to that observed in a classical Michelson interferometer. To obtain the optimum "focus" we have to detect the envelope of that function. A number of approaches can be used to detect this fringe envelope. Whatever approach is chosen, it must be fast and insensi...