Download as pdf or txt
Download as pdf or txt
You are on page 1of 2

VIETNAMESE - GERMAN UNIVERSITY

Electrical Engineering & Information Technology


Signals & Systems

HOMEWORK 3 (optional)
EEIT-2018 Intake

Due: Monday, May 24th, 2021

Problem 1
Given a casual second-order process which follows the transfer function

and initial inputs and


no dead time.
a) Calculate the noise-free output , .
b) Assuming the process is not known and only measured input and output samples found
in part a) are given, identify the process by applying the Least squares fit and estimating the
unknown parameters of .
c) What can you conclude after estimating the unknown parameters in part b)?
Problem 2
Given the analog filter prototype

Calculate the frequency response of the filter, ensuring the stability and minimum time-
delay properties for the system.

Problem 3
Design a first-order high pass IIR filter with cut-off frequency . Use the Butterworth analog
filter prototype.

Problem 4
Consider the noisy signal , where is the white noise. Your task is to
design the system that filters to yield . Use the concept of Wiener filtering learned
in class to calculate the optimal Wiener filter that achieves the task, i.e. specify the transfer
function of the filter.
Problem 5. Image restoration using inverse filtering
Complete Matlab code 'image_restoration_inverse.m' to do the following:
a. Read in image 'trui.png' using function imread and take 2-D centered Fourier transform using
functions fftshift(fft2( )). Display the image and its Fourier transform. Comment on how the
Fourier transformed image looks like.
b. Blur the given image by convolving with the motion-blur filter (or multiplying in the Fourier
domain). Hint: Simply run the provided code section.
c. Do inverse filtering avoiding small values when dividing the blurred image with the filter H in
the Fourier domain. Hint: Simply run the provided code section.
d. Add Gaussian noise with mean 0, variance 0.002 using the function imnoise and repeat
inverse filtering. Comment on the image restoration quality.

Problem 6. Image restoration using Wiener filtering


Complete Matlab code 'image_restoration_wiener.m' to do the following:
a. Read in image 'trui.png' again using function imread. Convert to the double format using the
function double() and save as the variable I.
b. Now we will simulate the blurred and noisy image by performing the following steps:
 Define the camera motion effect by creating PSF variable using function fspecial, which
represents the linear motion of a camera by 21 pixels, with an angle of 11 degrees in a
counterclockwise direction. Add the camera motion effect to the original image by
filtering the image with the created PSF. Hint: run the command imfilter(I,PSF,'circular').
Using function randn, generate noise variable having standard normal distribution with
mean 0, variance 15.
 Using function randn, generate noise variable having standard normal distribution with
mean 0, variance 15. Add this noise to the blurred image created in the previous step.

c. Apply the Wiener filtering to the blurred and noisy image by running the provided code. The
main steps in the code are calculations of the noise power spectrum and signal power spectrum
and then performing Wiener filtering using function deconvwnr. Comment on the restoration
quality compared to the inverse filtering in Problem 5.

You might also like