Assignment #2: Edge and Corner Detection, and Hough Transform

Due date: 07/04/2016

Implement a Derivative of Gaussian filter and a single thresholding approach to extract candidate edges and implement a Hough transform to detect the main lines from this image.

Contrast the result with that obtained using the corners extracted with a Harris detector

Evaluate the quality of the extracted lines as you vary

1. the parameters of the edge detection algorithm (threshold and sigma);
2. the parameters of the corner detection algorithm (threshold, sigma, and k);
3. the discretization of the Hough space;
4. the number of lines extracted.

Implement a weighted version of the Hough transform where the weight of the votes accumulated in each bin are proportional the response of the derivative of Gaussian filter/ corner detector.

Deliver the code and a report where you study and comment the performance and dependency to the paremeters of the approaches.