[GRASS-user] i.gensig

[Manuel Seeger]

Markus Neteler schrieb:
> On Fri, Oct 24, 2008 at 10:57 AM, Manuel Seeger <seeger at uni-trier.de> wrote:
>   
>> Hello all,
>>
>> I using grass 6.3.0 on a 64 machine (ubuntu 7.10)
>> When applying i.gensig on a set of 20 rasters for defining 6 different
>> groups I get the following output:
>> Finding training classes ...
>> 6 classes.
>> Calculating class means ...
>> Calculating class covariance matrices...
>>
>> Signature [1] not invertible.
>>
>> Signature [2] not invertible.
>>
>> Signature [3] not invertible.
>>
>> Signature [4] not invertible.
>>
>> Signature [5] not invertible.
>>
>> Signature [6] not invertible.
>>
>> Writing signature file [DRP_Zemmer] ...
>>
>> Done.
>>
>> The problem is:
>> the signature file is empty.
>> it may be related to the info Signature [1] not invertible. But I don't know
>> what GRASS wants to tell me.
>>     
>
> I am not sure what happens - apparently
>  imagery/i.gensig/can_invert.c
> fails. Could you try to reduce the number of input maps for
> a test?
>
> Markus
>
>   
Its long time passed, excuse me
I tried reducing the number of rastermaps I get the same message.

When I transform the rasters first with PCA I get no error!

What is the problem?

Best regards
Manuel

-- 

_______________________________________________________________________
Dr. Manuel Seeger
Wiss. Assistent 			Scientific Assistant
Physische Geographie 			Dpt. of Physical Geography
FB VI - Geographie/Geowissenschaften 	Geography/Geosciences
Universität Trier			University of Trier
D - 54286 Trier
Tel.:	+49-651-201 4557
Fax:	+49-651-201 3976
Web:	http://www.uni-trier.de/index.php?id=18521
_______________________________________________________________________

EGU General Assembly 2009
SSS17. Experimental Methods in Soil Erosion Studies
Convener: 	  	Seeger, M.
Co-Convener: 	  	Kuhn, N.; Quinton, J.
http://meetingorganizer.copernicus.org/EGU2009/session/930
_______________________________________________________________________