Problem running Region Group tool

[ClintCabanero]

Hello all,

I'm simply trying to run the RegionGroup Spatial Analyst extension tool in ArcToolbox.  For some reason, when running this against a large raster (northern California), the tools runs 'successfully' - meaning it completes without error - except that the output has a data anomaly - whereby there are striped bands across the output.  The cells in these striped bands also have large integer values (e.g. 1991 while all other values are 1, 2, 3, 4, etc. for each unique group).  This tool problem has been repeated by different people on my team.

I'm seeing this same RegionGroup tool problem in ArcGIS Desktop 10 AND ArcGIS Desktop 10 Service Pack 1.

Can anyone confirm this same problem - or recommend a work around?

Thank you

[DanPatterson_Retired]

do you have a screen grab to illustrate this?
did you try filtering the grid a bit before processing using one of the options in the help files? "Simplify your data by filtering it. This can be done with tools such as Majority Filter, Aggregate, Shrink, and Focal Statistics".   The choice would depend upon the nature of your input data.

[FelixStracke]

Hi,

I am running into the same problem. I run the region group toll on whole North African Countrys and it runs without failure. But if the country is big (e.g. Algeria, Egypt) it creates the stipes without data. On smaller countrys it runs quite well (e.g. Morocco, Tunisia). Has anybody found a solution - I am getting crazy ...

Best regards

Felix

[AntonioNatale]

I have the same problem but with data not so big (a second level of italian's administrative unit). One horizontal strip is created in the output raster. I confirm this problem and I'm not able yet to find some solution on the web. I tryed also to repeat the Region Group tool changing some parameters and environment of processing (ArcMap, ArcCatalog e Arcpy) but without positive effects.
I attach one image of the result!
I think the best way to create the region raster is to create it out from a geodatabase. If you create a simple GRID you will not have any kind of mistake.
IT'S A BUG

[FelixStracke]

Hi,

thanks for your reply. I figured out that it is seems to be a size problem. The bug only appears when the dataset reaches specific size, e.g. Tunisia works quite well, Algeria was a mess.... Thus, my solution was to write a script that divides big image into smaller parts and than mosaic these parts together after region group.

Best regards

Felix

[AmyKeister]

I have the same problem.  I am re-running a model that I built in 9.3 and the region group results with the same bands of errors that are described above.  It ran fine on the same raster in 9.3 (the only difference is that I'm doing 8 neighbor instead of 4 neighbor this time).  Dividing the big image into smaller pieces won�??t work for me because I know I have at least one very large clump that would be split in the process.  Has ERSI created a fix for this problem yet?

[AnaGonzalez]

Hi,

I am doing research on Tree Canopy Cover and I am using the Supervised Classification option. I have been trying to run the Region Group tool but I always get a raster with a horizontal line on the center of the image. Is there someone who knows how to create the Region Group raster without the line?

[MichaelSutherland]

I have run into the same problem while trying to use the tool.  It appears as if the data of the original raster is maintained but simply shifted horizontally. 

I have tried running from within a file gdb, per team04's post, to no avail.  Striping persists.

[ChrisSnyder]

If you create a simple GRID you will not have any kind of mistake.

I think team04 meant that by using good ole' grid format as an input/output format (not a FGDB raster), the problem was avoided.

Just an FYI too, but generally, reading writting grid format in SA tools is way faster than FGDB raster format. If you use FGDB format stuff, behind the scenes it is formats to grid, runs the SA tool algorithm, and then, writes the grid output back to FGDB raster fromat. There is a lot of time that is potentially lost in all that reformating...

[PavanYadav]

Hi All,

We are currently investigating this potential bug. Once, our testing is done and we have a bug ID, I will contact you back. This issue might be related to two aspects:
1. Large input raster in terms of number of rows and columns; and/or
2. Unique value range and complexity of unique pixel values distribution.

Comments on this forum indicate that if the spatial extents (coverage) of the input data is fairly large, the output will have horizontal lines. So maybe the second condition is not related to this issue.

I will greatly appreciate if you could contact Esri Support and share your test data with us. Testing with different datasets will help us narrow down the problem.