Digital Image Processing

Note: Use MATLAB Software or ‘C' only

Set I

1. Make a BMP image file with a white square in center in paintbrush. Display the magnitude of its 2-D Fourier Transform in the form of the image. Bring the (0, 0) - frequency coordinate to the center of this display. Change the size of the square and observe the change in the corresponding DFT display. Comment on the observation. Rotate the image by some angle (say 45°) and observe the change in FT display. Explain your observation with mathematical equations.

2. Load an image. Take its FFT. Reconstruct the image using only

a) magnitude information of FFT

b) Phase information of FFT

Comment on the observation.

3. Load an image in MATLAB in gray shades. Change the number of gray levels to 64, 16, 8, 4, 2 to display the same image. Comment on the observation. Now change the spatial resolution of this image. Explain the effect of sampling error.

4. Take an image having only two point sources. Pass it through a low pass filter defined

(a) by a mask in spatial domain

(b) in frequency domain

Observe the result. Change the cut-off frequency of the filter and comment on the result.

5. Take an image that has black and white alternating vertical strips. Pass it through a high pass filter defined in

(a) in frequency domain

(b) by a mask in spatial domain.

Observe the result. Change the cut-off frequency of the filter and comment on the result.

6. Take an image. Carry out edge detection in both directions using

(a) Sobel operator

(b) Prewitt operator

(c) Roberts operator

Do not use in-built commands. Use the corresponding masks and then carry out edge detection.