www.edmundoptics.co.uk/imaging 169
resource guide fixed focal length telecentric liquid lens / specialty filters/accessories
cameras illumination targets
microscopy /
objectives
able resolution may be less than what the MTF curve suggests.
Calculations from an MTF test bench are, by definition, contrast as a
function of spatial frequency within a one, two, or three-dimensional
coordinate system.
Slanted-Edge MTF Testing
Slanted-edge MTF testing obtains the same system level information
as MTF testing but more quickly, with more adaptability, and on less
expensive equipment. Where regular MTF testing uses the Airy disk
of a point source, magnified by an objective to fit the full sensor to
remove any external MTF contributions, slanted-edge MTF testing
uses a high contrast edge target positioned at an angle of several
degrees. If only the MTF contribution of the lens is required, then
all contributions from other components must be removed from the
system level MTF by dividing them out (because contributions are
multiplicative). The first step in obtaining the slanted-edge MTF is to
measure the sigmoidal-shaped edge-spread function and its derivative
to determine the line-spread function, which is then filtered, and
Fourier transformed into the MTF curve.
Figure 12.3
Figure 12.3: Showing a slanted edge MTF measurement. Multiple
measurement areas can be tested concurrently.
This process only works if the contrast transition of the edge target
occurs of a scale four times smaller than the Nyquist limit. If an edge
target has a transition of 100μm and the resolution goal is 100lp/mm,
(5μm Nyquist sample size), then the edge target would be enough, if
the magnification (m) is less than 0.0125X (see calculation).
(Transition Width)Image space μm < Nyquist Width μm ÷ 4
(Transition Width)Object Space μm × m < Nyquist Width μm ÷ 4
100μm × m < 5μm ÷ 4
5μm
100μm
m < ÷ 4
m < 0.0125
Camera Tests
Camera test is an umbrella term for all tests that use a camera. The
slanted-edge MTF test is one specific camera test but it is not the
only common camera test for resolution. Camera test methods can
be adapted, using different techniques or equipment, to any real-word
application to obtain system-level, environmentally relevant resolution
information. These tests often use test targets that allow for a
specific set of performance metrics to be tested at the correct imagefield
positions and in the test environment.
As sensors continue to have a greater number of smaller sized
pixels, the optical and mechanical requirements of lenses made to
interface with them must also improve. System and machine vision
end users often misunderstand the effects that govern vision system
performance. Understanding what information each test method provides,
as well as the pros and cons for each, will ensure success.
Method Pros Cons
Rear
Projection
Fast Qualitative
Inexpensive Poor diagnostic capabilities
Operator friendly
Poor ability to control spectral content
(response is photopic)
Tests multiple field points concurrently Hard to maintain accurate coordinate system
Quickly establish limiting resolution Difficult to use for higher magnifications
MTF
Off the shelf measurement
instruments available
Calculated at one image point at a time
Higher-order aberrational
content can be diagnosed
Insensitive to stray light and other
environmental contrast degrading phenomena
High accuracy and precision
Poor reproduction of most
real-world illumination environments
Most broadly-applicable
“general use” test method
Expensive, easily misused equipment
Slanted
Edge
MTF
Tests multiple field points concurrently
Target selection can be difficult for
common magnification range
(~0.05 - 0.01X)
Free and commercial
software readily available
Sensitive to illumination uniformity
Can use a standard (ISO 12233) or tweak
to suit specific needs
System-level contribution from non-lens
components can lead to large error bars
(especially at high frequencies)
Can be used to measure system level MTF
including lens, camera, illumination, and
any image processing algorithms
Easily misused
Coordinate system of cameras shifts
from tangential and sagittal
performance to horizontal and vertical
Camera
Tests
Tests multiple field points concurrently Easily misused
The most application-specific approach Little to no outside support
Endlessly adaptable Difficult to correlate
Can be used to measure system level
performance including lens,
camera, illumination, and
any image processing algorithms
Table 12.1: The pros and cons of some of the most common lens
performance test methods.
/imaging