Notes for Lab 8: Filtering and Edge Detection of Images

Announcements:

Display all of your results for the supplement with show_img (both the images and the results of fft2 applied to the images).
When using hist, a second argument may be supplied for the number of histogram bins. The default for this is 10; you may find that a higher number serves you better (100 or 1000). Also, you can simplify your histogram graphs by using hist(xx(:)).
The result of fft2 will be the two-dimensional spectrum of the signal, by oddly arranged: each axis (rows or columns) has starts with low frequencies at all of the corners, and increases to the nyquist frequency (sampling frequency / 2) and it's negative in the center. Below are examples of the magnitude and phase of such spectra:

Magnitude Angle

Here's another example, specifically for the lenna image:

Note that this result will be a matrix of complex values. That's why you need to take its magnitude or phase in order to graph it.

I want to combine two images, but they're of different sizes! You can just take as much data as is in the smaller one with smaller = larger(1:length(small), 1:length(small)).
My histograms are multicolored and jagged! The hist function is computing a separate histogram for each row and overlaying them; try putting all the data from the 2-D matrix into a 1-D vector, using a command like hist(data(:), 100). The "(:)" makes a vector of all the data.


Magnitude	Angle