Automated
Inspection of
Optics using
Quentin Turchette and Trey Turner
Automated surface- quality mea- surement instrument.
ISO Specifications
REO
Representatives from Research Electro-Optics Inc., a precision optics manufacturer,
explain how to quickly and accurately identify surface defects using an automated
microscope-based system that follows the ISO 10110-7 specification.
It is difficult for manufacturers of pre- cision optical components to quickly
and accurately detect µm-scale surface
defects. This is often because it’s time-consuming for an operator to examine
each piece; the resolution needed to measure these defects limits the area that can
be examined at one time; and, of course,
there is the potential for human error during data translation. A possible solution
for high-volume facilities such as ours is
to implement an automated microscope-based system for surface quality inspection using ISO 10110-7 specification.
Automated inspection
using ISO 10110-7
ISO 10110-7 defines allowable defect
sizes and the total area that they can
cover, but it does not specify how measurements are to be performed. Since
the smallest defect can be on the order
of micrometers, manufacturers cannot
rely on visual inspection. A calibrated,
high-magnification microscope can be
used to detect size defects on this scale.
However, an operator-based microscope
system has a small field of view, making
it time consuming to examine the entire
surface of an optical component.
At Research Electro-Optics Inc., we use a
conservative image-processing approach that
provides an upper-limit
estimate of defect size.
An automated microscope-based
system can speed up inspection and
eliminate operator-dependent biases.
Our system uses dark field microscopy
with a charge-coupled device (CCD)
detector, image-processing software and
a motorized stage that can test multiple
sites on a single optic. The CCD only
images the light scattered from optical
defects: It appears as bright spots against
a dark background. A computer controls
the camera and microscope stage motion,
while simultaneously receiving and processing data.
The system’s image processing soft-
ware measures the size of each defect by
finding the location of its edge. This is an
uncomplicated process for defects that
are substantially larger than the resolving
power of the microscope because the
intensity profiles typically have wide,
flat distributions with a small transition
region at the edge. Therefore, the exact
location of the edge does not significant-
ly change the measured size.
14 | OPN Optics & Photonics News
www.osa-opn.org
