OPTICS
How MTF curves aid lens selection
How do users choose the best lens for an imaging system? Edmund Optics’Dr Boris Lange opts for modulation transfer functions to inform the decision
T
en to 15 years ago, when VGA sensors were considered high-
quality and pixels were 10µm, it was pixel size that limited the performance of an imaging system. Today, with smaller pixels, it’s optics that can limit resolution. Edmund Optics offers 11
different 25mm C-mount lenses as standard products, along with additional custom lenses. Tis is due to progress made in sensor technology; one standard lens is not enough to cover the range of image sensors available. How do users choose the best lens for a certain application? Te answer is to use modulation transfer function (MTF) curves, which are a good comparison tool – they give a measure with which to compare a €50 lens with a €150 lens or a €1,000 lens. An MTF curve plots contrast
along the y-axis and resolution in line pairs per millimetre along the x-axis. All MTF curves start in the upper left corner of the graph at 100 per cent contrast (0lp/mm equates to two objects that are infinitely far away and will always have 100 per cent contrast) and, as objects are brought closer together, the contrast drops. Multiple curves are required
to get a representative plot of how a lens will perform over the entire sensor – in the middle and at the edges – and at tangential and sagittal focus. Te curves can also indicate how the lens
will perform at a certain working distance, at the F#, and at the wavelength of light used.
Parameters for imaging performance Contrast is an important metric for image quality. Contrast by itself, however, is useless if a spatial metric is not specified alongside it. Two objects through a lens will still have high contrast when the objects are far apart, but when the objects are moved closer together, at some point every optical system will have trouble resolving the detail and
‘MTF curves give a measure with which to compare a €50 lens with a €150 lens or a €1,000 lens’
the two objects will eventually blur. Te spatial metric tells the user how far away two objects are from each other – this is what optical engineers call resolution. Te resolution of a camera
can be expressed in megapixels. But the resolution of an imaging system, or of the optics, is always a spatial metric, usually specified in micrometres or the distance apart of critical features. Optical designers may also use the inverse of micrometres, which is line pairs per millimetre (lp/
34 IMAGING AND MACHINE VISION EUROPE FEBRUARY/MARCH 2020
mm): 145lp/mm equates to a pixel size of 3.45µm; 100lp/mm would correspond to a 5µm pixel. Another consideration the
user must make is the position on the sensor where contrast and resolution are measured; the centre of the image or at the edges. Te user can alter the focus to sacrifice performance in the middle of the image to gain performance at the edge. In addition, resolution charts
will specify a tangential and sagittal focus. Wavelength and the F# of the lens will also influence performance of the imaging system.
Advice for using MTF MTF curves are not set in stone. Just as the user can focus a real lens, this can also be done using optical design software. In this software, the user can sacrifice performance in the middle of the image for performance at the edge. In addition, MTF curves don’t include manufacturing tolerances, so make sure there’s enough buffer between what you want and what is seen in the MTF curves. Tis might be a 10 or 15 per cent MTF point drop,
going from nominal design in the perfect world to the real world. Keep in mind specific
application parameters when using MTF curves. Contrast will change depending on the working distance – you can get 30 per cent contrast at infinity, but 10 per cent contrast at a working distance of 75mm at 230lp/mm. Te same is true with F#. Te best way is to think
about the application: specify the contrast and the resolution needed for the feature size extracted in the centre or edge of the image, and then relate that to the MTF performance. O
This article is based on Dr Boris Lange’s presentation given for our Future Vision webcast,
www.imveurope.com/webcast. Lange is imaging Europe manager at Edmund Optics. The firm has an introduction to modulation transfer function on its website.
Would you like to write about the optical design of an imaging system you’ve developed? Email:
greg.blackman@
europascience.com
@imveurope |
www.imveurope.com
Vladimir Arndt/
Shutterstock.com
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40
