Matlab Out of memory in step aps_powerlaw(3,3)

[jayd...@gmail.com]

Dear David,

When I run the aps_powerlaw(3,3), the following error came out. My matlab version is Matlab 2012. The physical memory of my computer is 8 GB. In my view, it is big enough. After running the PS processing with StaMPS, there is just 7112 PS points left from 10 SAR images. The value of powerlaw_spatial_bands in parms_aps.mat is [2000, 4000]. Do you know how to solve this problem? 

Thank you very much!

>> aps_powerlaw(3,3)

********************************************************************************************************** 

D.P.S. Bekaert, A.J. Hooper and T.J. Wright,

A spatially-variable power-law tropospheric correction technique for InSAR data, JGR,  doi:10.1029/2014JB011558

**********************************************************************************************************

Step 3: Bandfiltering of the regular grid and converting back to a local grid 

***Bandfiltering***

Using butterworth filter degree: 3

Normalise the butterworth filter. 

norm_filter_flag =

     0

n_degree_butterworth =

    10

Perform mirroring to reduce filtering effects on edges 

By half the maximum filter length at each edge. 

??? Out of memory. Type HELP MEMORY for your options.

Error in ==> bandfiltering at 280

            z_grid_new(:,:,k) = padarray(z_grid(:,:,k),[n_mirror_y

            n_mirror_x],'symmetric');

Error in ==> aps_powerlaw at 463

    bandfiltering(Z_regular,xy_res(1),xy_res(2),spatial_bands,save_path,ifg_based_correction) 

[David Bekaert]

Hi,

The Fourier filtering requires a regular grid for each interferogram.

The resolution to which this regular grid is created is given by the value of powerlaw_xy_res in meter units.

Believe the default is 30 m, which be changed such it is not be smaller than the merge_resample_size in stamps.

To avoid edge effects during band filtering this area is further expanded with the maximum band filter size by mirroring the data.

It is at this stage where the code crashes for you, giving the out of memory message.
If your study region is larger than 100 km, the number of grid points can rapidly increase with a higher resolution.
It might help if you relax the value of powerlaw_xy_res, by lowering the resolution.

Cheers,

David

--
You received this message because you are subscribed to the Google Groups "TRAIN - Toolbox for Reducing Atmospheric InSAR Noise" group.
To unsubscribe from this group and stop receiving emails from it, send an email to TRAIN_suppor...@googlegroups.com.
To post to this group, send email to TRAIN_...@googlegroups.com.
Visit this group at http://groups.google.com/group/TRAIN_support.
To view this discussion on the web visit https://groups.google.com/d/msgid/TRAIN_support/15d883f1-589a-45f3-8765-340544dc215c%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[jayd...@gmail.com]

Hi David, 

Thanks for your reply. 

You said that "Believe the default is 30 m, which be changed such it is not be smaller than the merge_resample_size in stamps." But, in the TRAIN munual, the description of powerlaw_xy_res is: "The regular grid resolution [x y] in m, to which the data will be interpolated. This should not be made larger than the resolution of the data."  Do the merge_resample_size and the resolution of the data have the same meaning? If so, is it contradictory?

Best wishes,

Jie

[jayd...@gmail.com]

Hi David,

I just tested it by modifying the value of powerlaw_xy_res to [100 100] with the value of merge_resample_size 100. It really works. Thanks!

Jie

[David Bekaert]

Hi Jie,

The wording of the manual is a  bit confusion, I have changed it for the next release.

powerlaw_xy_res = The regular grid resolution [x y] in m, to which the data will be interpolated. This value should not be better (smaller) than the data resolution, i.e. powerlaw_xy_res >= data_res .

Cheers,

David

--

You received this message because you are subscribed to the Google Groups "TRAIN - Toolbox for Reducing Atmospheric InSAR Noise" group.
To unsubscribe from this group and stop receiving emails from it, send an email to TRAIN_suppor...@googlegroups.com.
To post to this group, send email to TRAIN_...@googlegroups.com.
Visit this group at http://groups.google.com/group/TRAIN_support.

To view this discussion on the web visit https://groups.google.com/d/msgid/TRAIN_support/a0f11f8d-8880-4934-8179-e5b8b164909f%40googlegroups.com.