Re: sin profile fitting

Posted by Michael Schmid on Oct 16, 2008; 12:08pm
URL: http://imagej.273.s1.nabble.com/Fwd-area-color-tp3683786p3683794.html

Hi Gabriele,

in case you are only interested in the periodicity or the height:

Make an image out of the data (1*1268 pixels in your case) and use  
Process>FFT>FFT Options to get the power spectrum, then find the  
highest peak.
Finding the correct the spatial frequency is easier if you swap the  
quadrants. Then pixel (0,0) will correspond to a spatial frequency of  
0; (0,1) will correspond to one cycle per image width ("width" means  
the width of the FFT image in pixels).
It won't be very exact, but it should be a good starting point.

The height of the peak will correspond to the square of the amplitude  
of your sine wave, with some arbitrary factor. Note that you have to  
integrate over a few pixels around the maximum.

I don't know whether this will be accurate enough for you, you have  
to try. Also, it only works if all of your data sets have the same  
length (otherwise you have to scale all images to the same image size).


Michael
________________________________________________________________

On 16 Oct 2008, at 10:29, Gabriele Umbriaco wrote:

> Good morning,
> on ANALIZE-->TOOLS-->CURVE FITTING I need to fit a sinusoidal  
> profile for search the value of maximum and minimum of the points  
> to calculate the visibility of the fringes. Can you help me? in  
> curve fitting is not present a sin fitting and I would integrate it.
> Thanks, Gabriele.