topic Re: NoData not really NoData? in ArcGIS Spatial Analyst Questions

NoData not really NoData?

Rogierde_Jong — Wed, 02 Feb 2011 13:50:17 GMT

Hello,

In ArcMap 9.3, I have multiple raster datasets which apparently have NoData pixels (that's what Identify shows). But the cells do not behave like NoData, e.g.

* the NoData color doesn't work, instead the pixel has the 'lowest' color in the color range;
* checking for NoData values using Con(IsNull([in_grid])) yields only 0, so no pixels are recognized as NoData
* when applying a focal mean function (3x3), all 9 pixels get the same "NoData" characteristics as mentioned above, although the other 8 pixels do have (floating point) values.

What might be the problem and how can I filter/replace these "NoData" pixels?

Best regards,
Rogier de Jong

Re: NoData not really NoData?

DonovanCameron — Wed, 02 Feb 2011 18:06:55 GMT

would this article help? I refer to it often when I can't remember exactly how to fill those notorious NoData strips inherent in Geobase DEM's that don't edge match completely.

Seams between mosaic'd DEMS (Patching gaps) - Quantdec

A quick way to visually check what your NoData values is in the Source tab of that rasters properties window:[ATTACH]4552[/ATTACH]

Then navigate to the Display tab and place that checkmark as seen here and enter the RGB values that you seen in the Source tab earlier:[ATTACH]4553[/ATTACH]

related forum (1 of many...)

Re: NoData not really NoData?

Rogierde_Jong — Thu, 03 Feb 2011 14:36:34 GMT

Thank you Donovan! I have not figured out yet what went wrong exactly, but in the end it worked. What's strange to me is that the Identify cursor shows NoData for certain pixels, while the test IsNull(in_grid) does not show NoData-pixels (I get a raster with only 0). I looked up the NoData value in the properties source tab (as you mentioned) and it appeared to be the lowest possible signed 32-bit value. Filtering using Con(in_grid < - 30000, 0.00, in_grid) solved the problem.

Re: NoData not really NoData?

AlfredWagtendonk — Thu, 09 Jun 2016 12:02:29 GMT

This is a recurring question probably related to the file format (but the file format is not indicated below) and the application of a compression type for which I give you my recurring answer:

I had a similar problem and first tried to convert my data as well. But, than noticed my Geotiff had a compression applied ('PACKBITS'). After exporting my dataset (right click layer =>Export Data) to an uncompressed Tiff (Compression type: NONE), the problem was solved: NoData cells behave again like expected and you can carry out again common NoData operations. So have a look at your layer properties to see if your Tiff has been compressed as well and if so uncompress it.

Good luck and best wishes,

Alfred Wagtendonk