Image Processing: Dteremining the MTF of an optical system from a PSF or LSF

[blueeyedpop]

Hi,

I am designing an automated MTF test bench to determine the optical
quality of a lens system. This will be expressed in terms of MTF. The
imaged target will be a horizontal slit being imaged by a vertical
line scanner, thus producing a point.

How do I determine the MTF for arbitrary frequencies ?

A post on sci.image.processing suggested that I
"compute the 2-D Fourier transform of the image,
compute the modulus from real & imaginary parts of FT,
normalize to 1.0 at zero frequency,
frequency should come out as cycles per unit distance, where pixel
width
is the unit distance"

but I have no idea how to express this in labview.



Thanks.

Mike

[iangbell]

In effect - getting MTF by this route is the same as frequency response testing using impulse response. If you want togo into the maths you will find that the inverse FFT of a systems frequency response is the systems impulse response in the time domain.

This means that given the time domain impulse repsonse of a system you can use an FFT to get the Frequency response.

In your case the for time domain read spatial domain. You will measure a 1d signal which is the optical systems response to your impulse. My advice is to put your signal into a waveform datatype - enter the spatial sample interval as dT and then you can use the FFT Spectrum (Mag-Phase).vi fromt he Waveform Measurements palette. This will automatically give the magnitude spectrum which you ca
n then normalise.

So basically use the FFT functions in LV.

IanB