Description
Task 1: (5 Marks) – Function Plotting
Write a MATLAB script which plots the following function:
y = (1 − e
−kt)cos(2πf t) (1)
for:
k = 0.7
f = 3.5
t = [0, 5]
The [0, 5] syntax means “from 0 to 5, including both 0 and 5”.
Your script should plot 1000 data points with the default line style and colours provided by the
plot() function.
For full marks you also need to:
• Write vectorised code. You are not permitted to use any loops or conditional statements. Pay
careful attention to MATLAB element-wise operator behaviour. -1 if for loops etc are
used.
• Comment your code with a description of what the script does. -0.5 marks if missing
• The values of t must include both 0 and 5. -1 mark if t stops close to, but not exactly
on, 5.
• Allow the script to be easily modified with new values for k, f, and t. -1 mark if these are
not clearly initialised and easy to modify (eg: if they are hard coded into the equation
instead of being variables).
• Annotate the plot with axis labels and a title. -1 mark if missing. Read the MATLAB help
documentation on xlabel(), ylabel(), and title() to learn how to do this. The labels
are to be:
– x-label: “Time (s)”
– y-label: “Voltage (V)”
– Title: “Transient”
Page 1 of 3
Project: ENGG1003 – Programming Assignment 2,ENGG1003- Introduction to Procedural Programming
Task 2: (5 Marks) – Noisy Image Restoration
You have been provided with seven images named IMG 5698.JPG to IMG 5704.JPG. These
images were taken in low light conditions with a very high ISO setting which results in “noise” on
the image (random fluctuations in pixel intensity) and “bright pixels” (pixels which appear bright
in every image, despite not being hit by light).
In addition to the seven source images you are also provided with a “dark frame”, dark.png. This
is a photograph taken with the lens cap installed and measures the “bright pixels” intensity.
Write a MATLAB script which performs the following tasks:
1. Calculates the mean of all seven images
2. Subtracts the dark frame from the “mean” image
3. Performs a threshold to blacken “nearly black but not quite” pixels. The threshold value,
assuming a 0-255 intensity range, is to be 25. ie: any intensity value on any pixel channel less
than 25 should be set to zero.
4. Displays the result.
The mean calculation and dark frame subtraction should be calculated on all channels independently
so that the colour balance of the result is not affected.
Marking notes:
• Comments are not required for this task
• The images are all loaded correctly: 1 mark
• The mean calculation is performed correctly: 2 marks
• You will be awarded one of the following for vectorisation:
– Fully vectorised code: 2 marks
– Unvectorised code: 1 mark
NB: a loop may be used to loop over the 7 images but the mean calculation and
dark frame subtraction must be vectorised for full marks.
To save time, you may use the template below to load the seven images and convert the to the
double data type for processing. Remember to convert back to uint8() or scale [0,1] for display
with image().
i1 = double(imread(“IMG_5704.JPG”));
i2 = double(imread(“IMG_5703.JPG”));
i3 = double(imread(“IMG_5702.JPG”));
i4 = double(imread(“IMG_5701.JPG”));
i5 = double(imread(“IMG_5700.JPG”));
i6 = double(imread(“IMG_5699.JPG”));
i7 = double(imread(“IMG_5698.JPG”));
Page 2 of 3 ENGG 1003
Project: ENGG1003 – Programming Assignment 2, ENGG1003- Introduction to Procedural Programming
Task 3: (5 Marks) – Unsharp Mask Image Enhancement
Write a MATLAB script which processes the result of the previous task with an unsharp mask
algorithm.
This algorithm requires the following processing steps:
1. Load the original image and convert it to double() for processing.
2. Create a copy and blur it. See the Week 10 lab for a suitable blurring algorithm or research a
built-in MATLAB function which performs an image blur. Experiment with the “amount” of
blur. 1 mark.
3. Subtract the blurred copy from the original. This is done by applying the formula
y = original − blurred (2)
4. Scale the intensity of y. ie: multiply the intensity of each pixel by, say, 0.1 or 10 or 100.
Experiment with this value.
5. The output image is the blurred and intensity scaled data y plus the original image.
6. Display the result. A conversion to uint8() will clip any negative values or values over 255.
You do not need to worry about this clipping. 3 marks.
7. Compare results between processing the noise-reduced image to an original image. How susceptible is the unsharp mask process to image noise? Does it make the noise worse or suppress
it?
8. Show your demonstrator the result of varying the parameters (blur amount and scaling of y)
and the comparison between a “noisy” image and the “denoised” one. 1 mark.
This effect is easy to over-do but for demonstration purposes the figures below provides an extreme
before & after example. Normally it is used as a subtle enhancement, not an “obvious” effect.
More information about this algorithm, plus examples, can be found on Wikipedia: https://en.
wikipedia.org/wiki/Unsharp_masking
Page 3 of 3 ENGG 1003
