ENGT5111 Digital Signal
Processing
2019-20
Filter Design and Image
Processing
(Resit Assignment)
(Maximum Marks: 50)
Issue date: 04/07/2020
Submission
deadline: 04/09/2020
Feedback: 25/09/2020
What to submit:
Your coursework must be submitted as a report in Word format to Turnitin. All MATLAB scripts used must be included
as appendices at the end of the report and numbered according to the
corresponding questions. All code lines must be commented to explain them. All
sources used in the report must be cited. All references must be in IEEE
format. Please note that 5 marks
will be allocated to the presentation and organisation of your report.
Part
I (Filter design and signal analysis)
Problem 1
1)
Use MATLAB to design a bounded-input bounded-output (BIBO)
stable lowpass IIR filter with sampling frequency 8192 Hz and cut-off frequency
300 Hz. Plot the magnitude response, phase response, and pole-zero plot of the
filter. Discuss your results.
(10 marks)
2)
Apply your filter to the MATLAB built-in audio signal ‘handel’. Listen
to the original signal and to the filtered signal and comment.
(5 marks)
3)
Use the MATLAB function spectrogram to plot the spectrogram of
the original ‘handel’ signal and the spectrogram of the filtered signal. Your
plots must show time on the x-axis. Discuss the plots.
(8 marks)
4)
Repeat part 2) and part 3) for the MATLAB built-in signal
‘chirp’.
(2
marks)
5)
Find a simple mathematical expression for the output signal if your
filter is applied to the signal x[n]=16sin5(0.02πn)-20sin3(0.02πn)+5sin(0.02πn).
(5 marks)
Part II (Image
Processing)
Problem 2
1)
Write a MATLAB function
that takes as input arguments a 512x512 grayscale image and an integer n
(0<n<65), (1) partitions the image into 8x8 blocks, (2) computes the 2D-discrete
cosine transform (DCT) of each block, (3) orders the DCT coefficients in each
block according to the zig-zag scan, (4) sets the last n coefficients in each
block to zero, (5) computes the inverse 2D-DCT of each block of coefficients,
and (6) displays the resulting image. (10 marks)
2)
Plot the peak
signal-to-noise ratio (PSNR) for n=1, 2, 4, 8, 16, 32, 64 when the input image in
part 1) is ‘boat.512’ (Fishing Boat) from
http://sipi.usc.edu/database/database.php?volume=misc. Show the
reconstructed image each time. Discuss the results.
(5 marks)
0 comments:
Post a Comment