Hi, I am unsure about the intent of the Sobel filter; specifically, what does the blurring vector (1, 2, 1)^T contribute? Thank you!!
motoole2
As stated in lecture (on this slide), there are many derivative filters to pick from! Without smoothing in one direction, the derivative could end up being a little more noisy. So this small amount of smoothing can make the Sobel filter more effectively pick out gradients in the presence of noise.
Hi, I am unsure about the intent of the Sobel filter; specifically, what does the blurring vector (1, 2, 1)^T contribute? Thank you!!
As stated in lecture (on this slide), there are many derivative filters to pick from! Without smoothing in one direction, the derivative could end up being a little more noisy. So this small amount of smoothing can make the Sobel filter more effectively pick out gradients in the presence of noise.