[GRASS-user] Interpolation method for parallel samples

GRASS GRASS user
1

Hello!

I have airborne gamma measurements obtain in lines. The spatial distance between the ‘points’ along the line is 50m (e-w direction) and the spatial distance between the lines is 1000m (n-s direction)

I am currently experimenting with the v.surf.bspline function - setting ew_step=100 and ns_step=1000. I am having trouble setting the ew_step any lower…? I am also using the smoothing parameter lambda_i.

If anyone have some interpolation tips with this type of data - I would be happy to hear about it.

Martin

2

On Fri, Dec 12, 2014 at 1:43 PM, Martin Album Ytre-Eide
<Martin.Album.Ytre-Eide@nrpa.no> wrote:

Hello!

I have airborne gamma measurements obtain in lines. The spatial distance
between the 'points' along the line is 50m (e-w direction) and the spatial
distance between the lines is 1000m (n-s direction)

I am currently experimenting with the v.surf.bspline function - setting
ew_step=100 and ns_step=1000. I am having trouble setting the ew_step any
lower.....? I am also using the smoothing parameter lambda_i.

If anyone have some interpolation tips with this type of data - I would be
happy to hear about it.

While I am not familiar with this kind of data I wonder if anisotrophy
is involved.
If yes, then you may also check

http://grass.osgeo.org/grass70/manuals/v.surf.rst.html
with
theta=float Anisotropy angle (in degrees counterclockwise from East)
scalex=float Anisotropy scaling factor

Markus

3

Martin Album Ytre-Eide wrote:

Hello!

I have airborne gamma measurements obtain in lines. The spatial distance
between the 'points' along the line is 50m (e-w direction) and the spatial
distance between the lines is 1000m (n-s direction)

I am currently experimenting with the v.surf.bspline function - setting
ew_step=100 and ns_step=1000. I am having trouble setting the ew_step any
lower.....?

What exactly is the trouble with setting ew_step any lower? In theory,
it should be possible to set ew_step to any value.

Markus M

I am also using the smoothing parameter lambda_i.

If anyone have some interpolation tips with this type of data - I would be
happy to hear about it.

Martin

_______________________________________________
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user

4

Thanks for your answer Markus!

Here are my thoughts and trials:

The normal screen output when running v.surf.bspline looks something like this:

Reading values from attribute table...
Initializing output...
100%
Load masking map
100%
Processing subregion 1 of 25...
  90%

##
..but when running with ew_step=50 and ns_step=1000, I get:

Reading values from attribute table...
Initializing output...
100%
Load masking map
100%
100%
100%
100%
  10%

... and it does not seem to finish at any point.

Even setting the ew_step=90 generates the same behaviour.

My result with 100 and 1000 is not that bad actually - I might say I'm happy with that.

I am experimenting with v.surf.rst, but for some reason it ignores a lot of points (to dence). I also don't know if I understand the theta and scalex parameters. For me, it seems they are optimizing the results from a "feature" that is anisotropic. Does this "feature" in my case mean my area with samples that I wont to interpolate? My area is somewhat rectangular - being about twice the size in the nort-south direction, but my data - as I mentioned , are spaced by 50 m in the e-w direction and 1000 m in the n-s directions. So what combo of theta and scalex should I use?

Thanks!

Martin

_______________________________________
Fra: Markus Metz [markus.metz.giswork@gmail.com]
Sendt: 17. desember 2014 22:11
Til: Martin Album Ytre-Eide
Kopi: grass-user@lists.osgeo.org
Emne: Re: [GRASS-user] Interpolation method for parallel samples

Martin Album Ytre-Eide wrote:

Hello!

I have airborne gamma measurements obtain in lines. The spatial distance
between the 'points' along the line is 50m (e-w direction) and the spatial
distance between the lines is 1000m (n-s direction)

I am currently experimenting with the v.surf.bspline function - setting
ew_step=100 and ns_step=1000. I am having trouble setting the ew_step any
lower.....?

What exactly is the trouble with setting ew_step any lower? In theory,
it should be possible to set ew_step to any value.

Markus M

I am also using the smoothing parameter lambda_i.

If anyone have some interpolation tips with this type of data - I would be
happy to hear about it.

Martin

_______________________________________________
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user

5

On Mon, Jan 5, 2015 at 11:51 AM, Martin Album Ytre-Eide
<Martin.Album.Ytre-Eide@nrpa.no> wrote:

Thanks for your answer Markus!

Here are my thoughts and trials:

The normal screen output when running v.surf.bspline looks something like this:

Reading values from attribute table...
Initializing output...
100%
Load masking map
100%
Processing subregion 1 of 25...
  90%

##
..but when running with ew_step=50 and ns_step=1000, I get:

Reading values from attribute table...
Initializing output...
100%
Load masking map
100%
100%
100%
100%
  10%

Oops.

... and it does not seem to finish at any point.

That is a bug. Can you please open a ticket for this and add an
example using the North Carolina sample dataset? I guess this bug will
appear with any input data as long as ns_step and ew_step are
sufficiently different.

Markus M

Even setting the ew_step=90 generates the same behaviour.

My result with 100 and 1000 is not that bad actually - I might say I'm happy with that.

I am experimenting with v.surf.rst, but for some reason it ignores a lot of points (to dence). I also don't know if I understand the theta and scalex parameters. For me, it seems they are optimizing the results from a "feature" that is anisotropic. Does this "feature" in my case mean my area with samples that I wont to interpolate? My area is somewhat rectangular - being about twice the size in the nort-south direction, but my data - as I mentioned , are spaced by 50 m in the e-w direction and 1000 m in the n-s directions. So what combo of theta and scalex should I use?

Thanks!

Martin

_______________________________________
Fra: Markus Metz [markus.metz.giswork@gmail.com]
Sendt: 17. desember 2014 22:11
Til: Martin Album Ytre-Eide
Kopi: grass-user@lists.osgeo.org
Emne: Re: [GRASS-user] Interpolation method for parallel samples

Martin Album Ytre-Eide wrote:

Hello!

I have airborne gamma measurements obtain in lines. The spatial distance
between the 'points' along the line is 50m (e-w direction) and the spatial
distance between the lines is 1000m (n-s direction)

I am currently experimenting with the v.surf.bspline function - setting
ew_step=100 and ns_step=1000. I am having trouble setting the ew_step any
lower.....?

What exactly is the trouble with setting ew_step any lower? In theory,
it should be possible to set ew_step to any value.

Markus M

I am also using the smoothing parameter lambda_i.

If anyone have some interpolation tips with this type of data - I would be
happy to hear about it.

Martin

_______________________________________________
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user