Catch the area outside a contour [message #44120] |
Fri, 20 May 2005 06:23  |
James[1]
Messages: 9 Registered: August 2004
|
Junior Member |
|
|
Hi Folks !
Can anyone help me with the following problem. For example, there is an
area
of the image withing some contour level. How to determine and get
parameters for a region otside that contour level withing the next N
pixels ?
Is there any IDL procedure/function for that ?
Thanks in advance,
James
|
|
|
Re: Catch the area outside a contour [message #44187 is a reply to message #44120] |
Mon, 23 May 2005 07:29  |
James Kuyper
Messages: 425 Registered: March 2000
|
Senior Member |
|
|
James wrote:
>> I'm sorry, I don't seem to understand the problem. Are you trying
to
>> extract how many pixels are outside a given contour level? And by
>> "Fixed Size" do you mean that you want every pixel to represent the
>> same value of area? I may be able to help with some elaboration on
the
>> topic.
>
> Chris Lee gave the right idea ! The question now, what is the name
> of that function, which allows one to expand the area ... Anybody
> can help with that ? I'm trying to find someting in IDL's help, but
> still without a success
I sent the following message on 5/20 using my employer's newsserver,
but I haven't seen any evidence that it's propagated outside that
newsserver. I'll try again using Google.
You want to identify a region consisting of all pixels below the
countour level, that are within a horizontal distance r of the contour?
You can do that with the DILATE command.
Here's an example using one of the test datasets that come with IDL:
@cntour02
contourLevel = 3250
r = 10
above = elev GT contourLevel
x = (INDGEN(2*r+1)-r)#REPLICATE(1.0,2*r+1)
y = TRANSPOSE(x)
disk = x^2+y^2 LE r^2
dilated = DILATE(above, disk)
boundary = dilated GT above
TVSCL,above+2*boundary
It looks like this didn't work properly in some areas near the edge; I
think you have to pad the image with a border equal to half the size
the the structure operator on all sides to avoid that problem.
|
|
|