{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

# Lab7 - that has absolute values smaller than p set them to...

This preview shows page 1. Sign up to view the full content.

EE440 – Introduction to Digital Imaging Systems Lab. 7 (50%) 7_1.bmp is a blurred image, emulating a movement of camera during the shooting of the LENNA image. The mathematical expression for the blurring is i.e., the average of ten consecutively shifted images, where B is the blurred image and A is the original. Restore the image A using pseudo-inverse filtering. Submit the restored image. Hints: 1. Use a filter size of 10 x 10, that is, h = eye(10)/10 . 2. Zero pad both your image and your filter. For example, if your image is NxN, and your filter is MxM, both should be zero-padded to the size (N+M-1)x(N+M-1). In this case, M = 10. 3. Get the Fourier transforms of both the filter and the blurred image. Let’s denote them by H and B . 4. Set a threshold p . For elements in H
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: that has absolute values smaller than p , set them to p . 5. Get the Fourier transform of the restored image using A = B./H 6. Apply inverse Fourier transformation to A to get a , the original image. 7. Adjust p to get better results. It is usually around 0.01. 8. Useful and valid MATLAB commands: fft2 ,ifft2, eye, abs, real (50%) Now add in some white Gaussian noise, with zero mean and standard deviation equals 12, on the motion blurred image B(x,y) to create a new image C(x,y). ) , ( ) , ( ) , ( y x y x B y x C Use inverse filtering and Wiener filtering (where 12 ) , ( v u P ) to restore the image. You can also use the simplified version, i.e., K v u H v u H v u Q 2 * | ) , ( | ) , ( ) , ( , with various values of K, to restore the image....
View Full Document

{[ snackBarMessage ]}

Ask a homework question - tutors are online