Raster and feature selection

2055
7
03-01-2016 03:16 AM
BaffourAwuah
New Contributor III

Is there any simple and easier way to extract feature information from classified raster dataset that intersects a feature class, and also maintain the other feature attributes? I'm using ArcGis 10.4

0 Kudos
7 Replies
DanPatterson_Retired
MVP Emeritus

There is the whole overlay toolset in the spatial analyst

An overview of the Overlay tools—Help | ArcGIS for Desktop

and their equivalent in vector world.

An overview of the Extract toolset—Help | ArcGIS for Desktop and plus overlay tools...

So you want to extract vector information using raster? raster extent, the actual bounds?

The options are many, what have your tried so far?

You have posted this in GeoNet help... you should move this to Spatial Analyst

0 Kudos
BaffourAwuah
New Contributor III

Dan, I tried spatial analysis Extraction (Extract by Mask). It worked alright but as you
know, the output table shows only Values and Counts based only on the 6
classifications (6 rows and 3 columns). I therefore have to find a way to
assign every feature to its classified raster.

0 Kudos
XanderBakker
Esri Esteemed Contributor

It is still not entirely clear what you are after, but it sounds to me that you might want to use Zonal Statistics as Table—Help | ArcGIS for Desktop and join the resulting table back to the polygons.

0 Kudos
BaffourAwuah
New Contributor III

My problem is that, I have Snow dataset in Raster format (classified in a range of ‘50
days of > 25cm deep of snow to 365 days: 6 classification), and different
land area Feature Classes (forest areas, cultivated pastures, open areas, etc.)
in different Municipal areas. I now want to find out the statistics of what
'classified snow range' falls within what Feature Class in each Municipal area.
I guess its difficult to explain better, but I think my problem is put across.

0 Kudos
DanPatterson_Retired
MVP Emeritus

you are describing a Combine operation Combine—Help | ArcGIS for Desktop

as one option.

Combine lets you do just that... combine the classifications of one set of data with another bringing the attributes of both together.  Interpreting the results can be challenging, but it does work.

data 1 ... classes   A,B,C

data 2 ... classes  A,B

output is not in order, but in order of occurrence but you will get unique combinations of the above assigned to a new class

0 Kudos
JayantaPoddar
MVP Esteemed Contributor
0 Kudos
DanPatterson_Retired
MVP Emeritus

Also... Move this thread ...  to an appropriate spot  ... you will get more targeted responses if you put questions in their proper place.