Raster Resolution (Extraction by Mask)

[Anonymous User]

Original User: gismare

Hi,

I am having trouble retaining raster resolution when extracting by mask. Before: [ATTACH=CONFIG]16058[/ATTACH] After: [ATTACH=CONFIG]16059[/ATTACH]

I originally had issues with the resolution around the boundaries of the mask, but changed cell size in Environments. Now, the resolution of the different classes are terrible. Any help regarding this
would be much appreciated!

Thanks!

[Anonymous User]

Original User: MBoucher21

In the images, the boundaries of the raster follow the mask quite well. Is there something happening in the classification of raster values? Are they going integer on you?

[DeanMorgan]

Hi Mark,

Thank you for your message.

It should be noted that the raster was created through kriging prior to extraction, although I'm not sure how this would have any bearing on the resolution following extraction. With regards to the classifications, they do
not appear to have changed, with the original values/classes staying intact. However, clearly the resolution of the class boundaries have changed, bemusing! Or perhaps something trivial I am overlooking?!

[Anonymous User]

Original User: MBoucher21

If you post your raster data in a zip file, I'll look at it.

[LucasDanzinger]

Just a thought, but if you open the Extract by Mask tool, click the Environments button, then click Raster Analysis, does it look like the output cell size is set to something different? Perhaps it is picking up the cell size from the mask feature?

[Anonymous User]

Original User: gismare

The raster data prior to extraction? Would this be the file extensions .ovr/.aux/.tfw?

.................................

I originally set the output size to 'same as layer __________' (raster prior to extraction) which yielded a cell size (x,y) of 0.0304, and gave the poor image (second image) attached.

I then tried again with 'As Specified Below' and chose a rather arbitrary value, nonetheless smaller, cell size of 0.002. This tighten up the boundaries fairly nicely (third image), but clearly there's still something wrong

[ATTACH=CONFIG]16137[/ATTACH]

[JeffreySwain]

What were the resampling methods used to display each raster?  If the cell sizes are the same, then there must be something else creating that look. I would consider the resampling method being used to display the raster under the layer properties.  Also if the issue is only in the pyramids, consider altering the resampling method used to create the pyramids.  Perhaps switching from Nearest Neighbor to Bilinear will make the displays match.

[Anonymous User]

Original User: gismare

Okay, so I created a raster through kriging again, left default settings in Environments, and then changed resampling to 'Bilinear Interpolation' and yes the classes were nicely blended: [ATTACH=CONFIG]16195[/ATTACH]

But of course, the boundaries need tightening up. So I extract again, changing output cell size to a better resolution, and fingers crossed, changing resampling as before will blend then classes, but no: [ATTACH=CONFIG]16196[/ATTACH]

Closer......:confused:

[DeanMorgan]

Could anyone please suggest how to retain the resolution of a raster when using the Extraction by Mask tool???

Thanks!

[DeanMorgan]

Anyone?? There must be a way around this problem, for what seems to be so fundamental. Any help would be greatly appreciated....