| Author |
 Topic  |
|
|
menhart@ii
USA
23 Posts |
 Posted - 01/26/2018 : 6:30:09 PM
|
Origin Ver. and Service Release (Select Help-->About Origin):
Operating System:
i wonder how origin smooths 2D data sets, i.e. matrices
origin has a good toolbox including FFT AA SG etc for 1D. but as fgar as i can tell, for matrices all it does is a simple sort of 2x shrink (average) 2x expand (interpolate)
is there any way to use more sophisticated and more controllable smoothing algorithms on 2D data? e.g. n-selected weighted adjacent averaging, convolution and other FFT methods?
|
|
|
easwar
USA
1965 Posts |
Posted - 01/29/2018 : 07:55:45 AM
|
Hello,
Thank you for the inquiry and your suggestions for smoothing methods.
We are currently working on an App for smoothing XYZ/matrix data, and we will look into expanding that to support the methods you suggested. The App will be published when we release 2018b around April of this year.
Easwar
OriginLab
|
 |
|
|
Hideo Fujii
USA
1582 Posts |
Posted - 01/29/2018 : 11:13:35 AM
|
Hi menhart@ii,
Until we get a specialized app which Easwar mentioned, as a work-around, you can try the following method:
1) Convert the matrix to an image by "Image> Conversion> Convert to Image" menu. Calculate the scaling factor
from min and max for both original and the image matrix. (You can get the min/max by copy-and-pasting the
matrix/image data to a worksheet, and running the column statistics.)
2) Apply filter. For example, the sample below was obtained by "Image> Spacial Filters> Median" menu.
3) Convert it back to the data matrix by "Image> Conversion> Convert to Data" menu. Rescale back to the original data range.
In the sample below, Graph1 is the 3D Surface plot of the original matrix, and Graph2 is from the smoothed data.
--Hideo Fujii
OriginLab |
Edited by - Hideo Fujii on 01/30/2018 2:24:40 PM |
|
|
| |
Topic |
|
|
|
matrix smoothing??