Convert Point Spread Function translate to MTF

Hi

I have a guassian point spread function and i would like to find the equivalent MTF in frequency domain.

I have the following code with assistance from : https://www.mathworks.com/matlabcentral/answers/274848-point-spread-function-translate-to-mtf

x=[1,2,3,4,5,6,7,8,9,10,11,12,13]

y=[0,1,3,21,46,156,222,140,111,25,14,3,2]

LSF=[x,y];

OTF = fftshift(fft(LSF)); % OTF

MTF = abs(OTF); % absolute value of OTF

MTF = MTF./max(MTF); % normalize

% correct sampling frequency for conversion on frequency bins to frequency

Size = length(MTF);

spacing = 1.12e-6;% pixel size

fsx = abs(1/spacing); % turn into sampling frequency

a = linspace(-Size/2,Size/2,Size); % form scale for conversion based on frequency bins

conversionx = fsx./Size; % conversion factor for frequency bin units to frequency (unit^-1)

Psi = a.*conversionx; % frequency (unit^-1)

figure

plot(Psi,MTF)

xlim([0,Psi(end)])

xlabel('cycles/unit')

ylabel('MTF')

could someone please help me convert the X axis from cycles/unit to lp/mm , i am not sure how this conversion is peformed

Answer 1

[Drishti]

Hi Zak,

To convert the x-axis from cycles/unit to lp/mm (line pairs/millimetre), you need to first convert the spacing from metres to millimetres.

By multiplying the spatial frequency (Psi) by ‘spacing_mm’, you will obtain the x-axis values in lp/mm.

Refer to the code snippet provided below for better understanding:

% convert spacing from meters to millimeters
spacing_mm = spacing * 1e3; 
% convert to lp/mm
Psi_lpmm = Psi * spacing_mm; 

I hope this resolves the query.