image processing - How to find the edges of different scales?

Question

The question is:

Briefly suggest how blurring and edge detection be combined to select edges of a particular scale (for example, only the edges of large objects)

This is from a mock question paper prepared by our professor.

I understand Sobel Filter has the effect of both blurring (Gaussian) and detecting edges in an image. However, what confuses me is how does that combination(of any filter(s)) help in detecting edges of a particular scale.

Is this related to Difference of Gaussian (DoG)? If yes, would it help with finding large object edges (or equivalently small object edges?

This is one of the results of DoG with gaussians with sigma 1.0 and 2.5

Seems like it has detected all types of edges.

Original Image:

Another hypothesis is a method used in SIFT to find features at different scale. By scaling the blurred image at multiple levels and finding differences between different blurred images. Would this help in finding large object edges (or equivalently small object images?

Please let me know if any additional details are required.

question from:https://stackoverflow.com/questions/65889975/how-to-find-the-edges-of-different-scales

Answer 1

Waitting for answers