| Re: Supervised classification under IDL [message #55751] |
Fri, 07 September 2007 10:51 |
Jean H.
Messages: 472
Registered: July 2006
|
Senior Member |
|
|
antoine.cottin@ete.inrs.ca wrote:
> Hi,
>
> I have a set of data (several billions points) with 5 statistical
> parameters compute for each points. I also have some subgroups
> corresponding to specific class.
> I like to perform a supervised classification of the complet dataset
> from the subgroups I've got.
> I really don't know how to do this under IDL (ENVI ?).
> Any help on this would be greatly appreciate.
>
> Thanks in advance.
>
> Antoine C.
Antoine,
First, if your data is NOT an image yet, you would have to create one,
with a different band for each of your parameter. Use
envi_write_envi_file. Do not worry about the graphical meaning of the image!
When this is done, transform your subgroup into ROI, Region Of Interest,
that you will use to train the classifier.
... and voila!, go into ENVI, open the classification menu, select your
image with the 5 bands, select the ROI you have defined, your
classification method and enjoy! ... note that you might have to break
each subgroups into smaller ones if they have too much variation within
the groups..
If your original data was NOT an image, you can easily make the
correspondence between the classified image and the original data, as
the Xth pixel class is the one for the Xth entry in your data.
Jean H.
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
