Scribd
Upload
28 views2 pages

SIMG-782 Digital Image Processing: Homework 5

This document outlines 5 exercises for a digital image processing homework assignment. Exercise 1 involves analyzing properties of a Gaussian function, including its Fourier transform and periodic extensions. Exercise 2 involves sampling the Gaussian to relate sample spacing to bandwidth and duration. Exercise 3 asks about sampling requirements for an image and properties of its discrete Fourier transform (DFT). Exercise 4 instructs the student to filter an image with varying Butterworth filters and examine the results in both frequency and spatial domains.

Uploaded by

Chintamaneni Vijayalakshmi
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
28 views2 pages

SIMG-782 Digital Image Processing: Homework 5

This document outlines 5 exercises for a digital image processing homework assignment. Exercise 1 involves analyzing properties of a Gaussian function, including its Fourier transform and periodic extensions. Exercise 2 involves sampling the Gaussian to relate sample spacing to bandwidth and duration. Exercise 3 asks about sampling requirements for an image and properties of its discrete Fourier transform (DFT). Exercise 4 instructs the student to filter an image with varying Butterworth filters and examine the results in both frequency and spatial domains.

Uploaded by

Chintamaneni Vijayalakshmi
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

SIMG-782 Digital Image Processing

Homework 5
Due November 8, 2005
Ex. 1 Dene the function It can be shown that
2

f (x) = ex

f (x)dx = 1

(a) Show that Ef =

1 f 2 (x)dx = 2

(This is called the energy in f (x). (b) Show that f (x) has the Fourier transform F (u) = eu (c) Consider the periodic function

fp (x) =
k=

f (x kL)

Find an expression for the Fourier coecients in

fp (x) =
n=

cn ei2nx/L

(d) Explore what happens when fp (x) is approximated by


N0 1

fp (x) =
n=N0 +1

cn ei2nx/L

with dierent choices for L and N0 . (e) What does the sum
N0 1

c2 0

+2
n=1

|cn |2

represent? What value do you expect it to approach for large L and large N0 ? Ex. 2 We want to represent f (x) of problem 1 by a list of sample values fn = f (nd), n = N, N + 1, . . . , N 1, N . (a) Determine values for N and d such that the error E= 1 2 Ef d fn 0.01 Ef n=N
N

Relate the quantities N, W, L, d where W is a measure of the bandwidth and L is a measure of the duration. (b) Calculate the DFT from the sample set that you choose. This will produce a set of Fourier coecients. Compare the coecient values to those you found in Problem 1.

Ex. 3 Suppose that an image of dimensions 4 6 inches has detail to the frequency of 300 dots per inch in each direction. (a) How many samples are required to preserve the information in the image? (b) How many values are contained in the DFT F (a, b) of the image? (c) Suppose that the image is sampled at a frequency that corresponds to detail up to 600 dots per inch (but in reality the detail only goes to 300 dots per inch). What is the eect on the DFT? Ex. 4 Here we are going to examine ltering of an image in the frequency domain. Obtain the image barb.png from the images directory. Investigate the eect of ltering the image with a Butterworth lter with parameter p = 2 and D0 = 10, 20, 40 and 80. (See lecture 12, page 20). Explain the results in terms of both frequency domain and spatial domain concepts.

You might also like