Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 1

EE440

1. Perform edge detection via derivative filtering on the image cameraman.tif to highlight the
edges.
a) Read in the image and normalize the pixel values by dividing by 255
b) Use a Sobel operator ([-1 0 1]) to detect vertical edges in original image
c) Use the rotation of that Sobel operator to detect horizontal edges in the original
image
d) Find the norm of the gradients. In other words, combine the horizontal and
vertical edge images using: sqrt(x_edge^2 + y_edge^2)
e) Threshold the combined edge image to generate a binary edge image. Determine
your threshold value to create most reliable edges.

Submit the original, vertical-edge, horizontal-edge, combined-edge, and the final image.

2. Gaussian noise with zero mean and std equal to 15 is added to each pixel value of the
cameraman image. Repeat problem 1.

In reality, we often use the following mask to detect vertical edges.

-1 0 1
-2 0 2
-1 0 1

Then we can rotate this mask to detect horizontal edges.


Repeat problem 1 using the new masks. Compare the results and comment.

3. Load the image circles.png. You will use morphological processing to fill the holes
between circles. Is there any side effect?

Try to fill the holes without any side effects?

You might also like