FFT of greyscale 1D periodic gratings image

I'm trying to relate the size of the pixels in the image to an actual length in a physical unit of measure of the period of the gratings in the image. More specifically, I know the value of the period in microns and I want to relate that to the value for the number of pixels per unit of physical distance.

How can I do this using the fft function in matlab?

Thanks

Answer 1

[Image Analyst]

The Fraunhofer diffraction image is the FFT of the grating waveform. The light diffracts out at certain angles and looks pretty close to the Fraunhofer pattern after a short while. So as you move the image plane away from the grating and the Fresnel zone, the pattern expands and gets bigger even though the shape and pattern are pretty much the same. You know the grating equation. The distance between spikes in the scene depends on how far away the image plane is from the grating.

To spatially calibration your image, see the attached demo.

[James]

Still not understanding how it is calculating the pixels per unit of physical distance. Also, how can I equate the Period to the number of pixels?

Thanks

[James]

Also what is the rgbValues improfile(1000) mean?

[Image Analyst]

You calculate the distance, using improfile or imdistline(), then tell it how many real world units that is. You have to know that. I'm not going to know if your image was taken from a microscope and the distances are in micrometers or if you took it though a telescope and the distances are in kilometers.

improfile(1000) says to take 1000 intensity samples between the endpoints of the line.