6 replies to this topic

[Hemant]

I have Generated slope and curvature map from Aster 30 meter Dem, but in both slope and curvature map, strips are appearing.
how to remove those striping..
please reply soon..

I am attaching the screen shot of both the maps.

Attached Files

•  Aster_30_meter_slope.jpg   539.36K   33 downloads
•  Aster_30meter_curvature.jpg   282.55K   40 downloads

Edited by Hemant, 15 April 2012 - 01:53 AM.

[NicKV]

I have Generated slope and curvature map from Aster 30 meter Dem, but in both slope and curvature map, strips are appearing.
how to remove those striping..
please reply soon..

You will need to apply a suitable directional filter to the original elevation data. Because the striping has a consistent orientation and spatial frequency, I suggest you use a filter based on the fast Fourier transform.

[Hemant]

Hi, NicKV
Thanx for reply...

But i have no idea about it can you please tell me that how to apply directional filter.
I am using ERDAS Imagine and Geomatica, can u send me the process on my email id : hemantt.gautam@gmail.com


Thanks in advance.

[DaveB]

I wonder if this is a case of the units for the z values and the units for for x/y being different? I know that can cause weird/ugly results when creating hillshades. Maybe the same problem is occurring here?

Here's an article about setting the z factor for hillshades that may help illuminate the issue in that case.

[Hemant]

Hello dave,
My DEM is already In Projected coordinate system......

[SaultDon]

Also check out these two papers in pdf format:

Destriping and Geometric Correction of an Aster Level 1A Image
Identifying and Removing Systematic Errors in USGS DEMS

[NicKV]

Hi, NicKV
Thanx for reply...

But i have no idea about it can you please tell me that how to apply directional filter.
I am using ERDAS Imagine and Geomatica, can u send me the process on my email id : hemantt.gautam@gmail.com


Thanks in advance.

I do not use either ERDAS Imagine or Geomatica so cannot advise your if they support Fourier domain filtering. Surfer has some Fourier domain capabilities and MathLab would certainly have.

As others have implied, the original L1A or L1B ASTER stereo data may not have been preprocessed correctly or sufficiently to remove sensor imparted artifacts.

The best solution would be to have the DEM recreated by the supplier.