[gdal-dev] Heuristics to classify raster data ?

[Dmitriy Baryshnikov]

Hi Even,

The result of classification the map produce less classes than 
aerial/satellites images. This is the way using k-means for you 
purposes. So you need do set the border count of classes to decide is 
this map or imagery. Less classes count - map, otherwise - imagery.

Best regards,
     Dmitry

07.03.2014 22:49, Even Rouault пишет:
> Le jeudi 06 mars 2014 20:50:45, Joaquim Luis a écrit :
>> Even,
>>
>> Did not get it all. You want a method that allows you to tell between a
>> map and aerial/satellite image?
> Yes exactly
>
>> I believe the k-means algorithm would produce quite good results on maps
>> as is expectable that individual clusters would have low variance.
> Hum, I read a bit http://en.wikipedia.org/wiki/K-means_clustering and not
> being familiar with it, it is not obvious how that could lead to the binary
> result I expect ;-) I was hoping for a ready-made solution for my problem. I'm
> kind of demanding tonight ;-)
>
> Thanks anyway,
>
> Even
>