[STATSGRASS] compare several maps

[Jarek Jasiewicz]

Carlos "Guâno" Grohmann wrote:
> Sure, sorry about that.
>
> I am analyzing various methods of calculation of surface roughness. So
> I will have several maps, produced with different methods, at
> different scales and with different datasets.
> All maps are calculated with moving-window procedures. I'll have maps
> calculated  with neighborhoods of 3,5,7,9,11,.. pixels, over DEMs with
> 5,10,25,50,90 meters resolution, using probably three different
> methods.
>
> So I'll end up with tens of maps! I was wondering what would be a good
> approach to compare them. I was thinking to rescale all maps to a
> common range (say, 0-255) first, and then compare.
>
> I guess it makes sense to compare like this:
>
> DEM_10m - Window_Size_3x3 - method A/B/C
> DEM_10m - Window_Size_5x5 - method A/B/C
>
> and so on for each dem resolution and window size.
>
> carlos
>
>
> On 8/7/07, Jarek Jasiewicz <jarekj at amu.edu.pl> wrote:
>   
>> Carlos "Guâno" Grohmann wrote:
>>     
>>> What do you guys think is a good approach to compare several raster
>>> maps at the same time? correlation matrix? some graphical output?
>>> density plots?
>>>
>>> cheers
>>>
>>> carlos
>>>
>>>
>>>       
>> well, give us more detail what do you need
>>
>>     
>
>
>   
difficult question, correlation matrix seems not to be good method, if 
you want to compare data produced with different resolution and 
different moving window (i.e pixel with greater area will have in all 
paces teh same value will cover  group of small pixels with compleatly 
deferent walue)

maybe autocorellation in individual map will  be  interesting parameter
density plots also seems interesting