| Re: IDL's built-in function DILATE and ERODE doesn't work as described in help [message #50830 is a reply to message #50676] |
Mon, 16 October 2006 13:21  |
gqshen2008@gmail.com
Messages: 9
Registered: October 2005
|
Junior Member |
|
|
Technically speaking, we can do that, but why we use IDL at all if we
have to put extra effort like that? I am sure many people here can
write most of IDL's routines by themselves. What IDL should do is
actually lessen the programming burden, not increase that.
Jean H. wrote:
> Could one simply add a border around its image, process the
> "label_region" and delete the border?
>
> Jean
>
> Karsten Rodenacker wrote:
>> Jo, I meant with deleting border points 'set to integer 0' which is
>> infact the background.
>> To define connected components there has background to be defined in
>> advance,
>> hence the background is NOT in the result a labelled component, even
>> not necessarily connected!
>> Connected compenents start with index 1! This differs with the results
>> of contour path where each connected border gets an entry in the info
>> list with mark inside or outside.
>>
>> However, the deletion of the border in LABEL_REGION is nasty and
>> superfluos today!
>> Regards
>> KR
>>
>> Am Fri, 13 Oct 2006 10:46:31 +0200 schrieb Jo Klein <jo_kln@yahoo.co.uk>:
>>
>>>> LABEL_REGION
>>>> Although not directly MM related it is one of the jewels of IDL
>>>> (like where and histogram). The disturbing thing is (again) the
>>>> border point behavior. It is deleting a one bit border.
>>>
>>> Hi Karsten,
>>> At the risk of repeating myself - it's even worse than that. It
>>> doesn't delete the border, but puts it into the first component. What
>>> if you're actually interested in that one, say, to calculate
>>> background features? It will be contaminated with portions of the
>>> other components - not nice.
>>> Cheers,
>>> Jo
>>
>>
>>
>>
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
