0
$\begingroup$

I am investigating processing of SAR images right now, it is new for me. From Magellan images (1024x1024), I have to determine the Doppler Centroid. For that, I have to compute the azimuth spectrum (several times because it changes a little bit depending on the range) for each range line. I use this code:

clear all; close all; clc

f = fopen('ff01.img');
image = fread(f, [1024 1024], '*uint8');
figure
imshow(image,[])
rect=[0 0 20 1024];
J = imcrop(image,rect);
figure
J_image=imshow(J)
rect2=[0 0 20  250];
figure
J2 = imcrop(image,rect2);
J2_image=imshow(J2);
 J3=fft(J2);
 figure
 j4=imshow(J3)

but the problem is that what I find clearly doesn't look like a spectrum... I take the absolute value and it gives nothing interesting, I think that there is something I don't understand....

Thanks/

$\endgroup$
  • $\begingroup$ You are going to have supply more details if you want a good answer. However, SAR images are complex-valued and I see no conversion to complex after you read in the data. Also, radar images typically need to be displayed using a log-scale and not linear. Why are you cropping the images? What domain are they in? $\endgroup$ – AnonSubmitter85 Feb 13 at 15:04
  • $\begingroup$ imshow(fftshift(abs(J3))). I am afraid that the way this question is phrased makes it too open. $\endgroup$ – A_A Feb 17 at 10:23