[GIS] Extract/Split raster dataset by attribute (OBJECTID)

Question

I have a raster dataset with an attribute table generated from a series of image classification (unsupervised) and processing tools (majority filter + boundary clean). The attribute table shows pixels of the classified image, grouped by classification i.e. the "OBJECTID" field.

I want to extract the portion of the image by "OBJECTID" to divide the classified image into separate rasters and ultimately, separate feature classes. e.g. take all pixels with OBJECTID = 2 and copy into new raster or convert to feature class

This has proved somewhat impossible for me so hopefully there's something simple that I'm missing?

Answer 1

One way to do this is by using the Spatial Analyst --> Extraction -->Extract by Attributes Tool (see ESRI documentation here)

The tool interface is fairly straightforward. For example in the Where Clause option you can build a query with the query builder or just enter “OBJECTID” = 2 to select a subset where OBJECTID = 2

In arcpy it is similar:

outRaster = arcpy.sa.ExtractByAttribute(inRas, “OBJECTID = 2”)
outRaster.save(r’C:\test.img’)

Additionally if you do not want to have to do the above for each OBJECTID and want to go directly to a feature class, you may want to consider using the Conversion --> From Raster --> Raster to Polygon Tool (see ESRI documentation here)

Use the OBJECTID as the Field option to the tool.

At times this will result in a messy polygon feature class with many different polygon’s having the same OBJECTID to clean this up. Run the Data Management --> Generalization --> Dissolve Tool (see ESRI documentation here)

Select OBJECTID as the Dissolve Field and check the Create multipart features option checkbox.