wtr
Full Member level 5
Hi Guys,
I've implemented different types of edge detectors and what i find they have in common is a symmetrical window (3by3). I was wondering if anyone knows of any work or papers where people use asymmetric windows.
That is to say a sobel window like the below
Becomes something akin to
This way, I'm hoping that i could somehow grab forward or backwards and thereby grow an edge. At the moment I'm not getting edge closure on my signal.
I've used morphological filters and am familiar with erosion and dilation.
The problem is after the different types of edge detection i use i get something like the below
-------x-----
Where the x is a gap in the line and therefore i can't close it to identify a rectangle.
Many thanks as always,
Wes
I've implemented different types of edge detectors and what i find they have in common is a symmetrical window (3by3). I was wondering if anyone knows of any work or papers where people use asymmetric windows.
That is to say a sobel window like the below
1 | 0 | -1 |
2 | 0 | -2 |
1 | 0 | -1 |
Becomes something akin to
2 | 1 | 0 | -1 |
4 | 2 | 0 | -2 |
2 | 1 | 0 | -1 |
This way, I'm hoping that i could somehow grab forward or backwards and thereby grow an edge. At the moment I'm not getting edge closure on my signal.
I've used morphological filters and am familiar with erosion and dilation.
The problem is after the different types of edge detection i use i get something like the below
-------x-----
Where the x is a gap in the line and therefore i can't close it to identify a rectangle.
Many thanks as always,
Wes