The purpose of this homework is to experiment with edge detection and openCV
Your program should be able to do the following:
1. Edge Detection [5 points]
o Add edge detection for grey level images to your choice of options.
o Use Sobel operator (3by3) and (5×5) to compute dx and dy, compute gradient amplitude, compute edge
o Your program should
operate within specified ROIs (up to three ROIs as in previous assignments)
generate display of the amplitude of the gradient operator as intensity image
generate binary edge image derived from amplitude of the gradient operator by thresholding.
generate binary edge image by further thresholding the above output using direction information
(e.g. display only horizontal (+/- 10 degree) edges or 45 degree +/- 10 degree edges etc)
o Test your program on some grey level images
2. Utilization of OpenCV [5 points]
o replicate first part of this assignment using OpenCV calls and compare performance
o utilize Canny module of OpenCV and compare to Sobel results
o implement histogram equalization using OpenCV and compare to you histogram stretching
o Implement Otsu algorithm using OpenCV and compare to your optimal thresholding algorithm
o Combine operations by applying histogram equalization only to foreground as determined by Otsu.
Make sure that you have complete report for this assignment (not just few comments).
Include input and output images (use several gray level images as appropriate).
Discuss performance of edge detection on grey level images.
Discuss performance and utility of OpenCV
Discuss performance of Canny edge detector (vs Sobel) as edge detection.
Discuss performance of combined operations
How to submit
• Submit paper report in class on the due date
• See TA help desk for instruction on program submission and testing.