Re: Metereological Cloud Size Detection [message #34472] |
Fri, 21 March 2003 05:28 |
David Fanning
Messages: 11724 Registered: August 2001
|
Senior Member |
|
|
Colin (colinduff65@hotmail.com) writes:
> i am looking for an IDL routine that will automatically scan an image
> and detect different cloud regions (image is a cloudmask ie pixel is
> either cloud or clear(0 or 1) ) and calculate their size. I can use
> the Find_Boundary routine but this works on only one region - the
> Where function was used to detect the cloud pixels for the whole image
> but only one cloud boundary can be detected.
I think you are looking for LABEL_REGION. In practice,
it is usually necessary to do a bit of image processing
first (say a round or two of erosion/dilation) to
eliminate a lot of single pixel regions. What you do
will depend upon your image and your application, of course.
In my application, which automatically finds "blobs",
I've had to be quite careful in the corners of images,
too, so these have a special processing step. In fact,
now that I think about it, be sure to put a border of
empty pixels around your image before you use LABEL_REGION,
because otherwise this routine will strip the edge
pixels off the region, giving erroneous area and perimeter
values. I remember this one costing me a LOT of time! :-(
Cheers,
David
--
David W. Fanning, Ph.D.
Fanning Software Consulting, Inc.
Phone: 970-221-0438, E-mail: david@dfanning.com
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Toll-Free IDL Book Orders: 1-888-461-0155
|
|
|