[GRASS-dev] change linear back to bilinear in grass7 ?

[Helena Mitasova]

Martin I read the discussion post (I should have mentioned it, I did not realize that the change has been made until now) 

to all - please read the wikipedia entry linked in my post (read the whole thing, not just the first paragraph), 

"Contrary to what the name suggests, the bilinear interpolant is not linear; nor is it the product of two linear functions.
Alternatively, the interpolant can be written as
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 539ce24cab74f38d7c3bc3f59ab0c916.png
Type: image/png
Size: 738 bytes
Desc: not available
URL: <http://lists.osgeo.org/pipermail/grass-dev/attachments/20140127/281e5899/attachment.png>
-------------- next part --------------
"

so the equation is not linear, the function is a hyperbolic surface

and even if you go with linear, then the examples in all manual pages should be updated -
currently it is a mess which could have been avoided just by keeping he original names of the parameters.

Helena

I have not checked the code for the bicubic option, but based on the manual entries referring to cubic convolution
(see again the wikipedia entry for bicubic interpolation) I am guessing it is bicubic. I will look at it - but the easiest 
is to look at the implementation and compare with the equation in wikipedia. 



Helena Mitasova
Associate Professor
Department of Marine, Earth, and Atmospheric Sciences
2800 Faucette Drive, Rm. 1125 Jordan Hall
North Carolina State University
Raleigh, NC 27695-8208
hmitaso at ncsu.edu

"All electronic mail messages in connection with State business which are sent to or received by this account are subject to the NC Public Records Law and may be disclosed to third parties.? 

On Jan 27, 2014, at 5:29 AM, Hamish wrote:

> Helena wrote:
>>> As we are going through the class material I have noticed 
>>> that in GRASS7  in r.resample.interp method bilinear was
>>> changed to linear and bicubic was changed to cubic.
> 
> Martin:
>> for the record, there was a discussion about that a while
>> ago [1]. So probably we will to open this discussion again.
> ...
>> [1] http://lists.osgeo.org/pipermail/grass-dev/2013-May/063855.html
> 
> 
> Hi all and happy new year.
> 
> for my 2c wrt interpolations I seem to come across "bilinear" in the world-beyond grass more commonly, and to me it makes more sense.
> Is the "bicubic" truly considering x and y separately? i.e. is it actually correct to use the "bi-" prefix with it? If it's more of a buffer operation, cubic might be better.
> 
> but whatever is chosen, consistency in use is good to help with the learning curves. I wouldn't object to bilinear + cubic though if each was considered best in its class.
> 
> 
> regards,
> Hamish
>