a) We compute the gradient on an image using a differentiating kernel that we call Sobel kernel. We also prefer to run a Gaussian filter over the image before we compute the gradient. Explain (i) why...

a) We compute the gradient on an image using a differentiating kernel that we call Sobel

kernel. We also prefer to run a Gaussian filter over the image before we compute the

gradient. Explain (i) why we use a Gaussian filter before computing the gradient, (il)

how we make this double filtering process computationally more efficient, and (ili) how

we compute the gradient direction at a pixel.

We threshold the gradient of the image twice when computing the Canny edges, using

a low threshold and a high threshold. Explain why we use two thresholds instead of

one, and what each threshold changes in the final result.

Jun 07, 2022

Get Answer To This Question