| IDL Morph structuring element [message #93268] |
Wed, 25 May 2016 07:18 |
David B
Messages: 18
Registered: January 2015
|
Junior Member |
|
|
So I am working with lines of pixels, which is about as exciting as one may expect.
An interesting and somewhat bothersome problem is that IDL morph_'whatever' does not seem* to support 'don't care' situations within the masks.
For example, morph_hitormiss would require:
Result = MORPH_HITORMISS (Image, HitStructure, MissStructure)
where the Hit/Miss structures are something like:
HitStructure = replicate(1, 3, 3)
producing:
1 1 1
1 1 1
1 1 1
Or you could have something radical like:
1 0 1
1 1 1
1 1 1
But what you cannot have is:
0 - -
0 1 0
0 0 0
where the dashes now represent null values or NANs. For the purposes of these morphological operations, this is useless as this then leads to no flexibility in the application of these algorithms where one does not care about the structure that is produced by these dashed areas under spatial transforms.
Perhaps someone at Harris could address this issue. Not that after the implementation in 5.3 any updates to these routines seem to have appeared.
*seem as though me staring at it for hours confirms this.
Doh!
David
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
