Best option/tools in ArcGIS Pro for analyzing aspects of surrounding environment?

Question

I would like to analyze what exists within a 1 km radius around an area on a map. Specifically I'd like to know the proportion of forest, water, urban elements and agricultural fields. I have the specific areas mapped out as polygons in ArcGIS Pro and I was initially thinking that I'd download orthophotos for each area, make a buffer zone of 1 km around the area and then just do some simple addition and calculate how many hectares of each there is.

But my first problem is that the orthophotos are massive files and take ages to put into ArcGis it seems (my laptop is definitely having to work for it). I mean it's doable but now I'm wondering if there are any other options for the analyses I want to do? I have previously done slope analysis of the areas with DEM's and the DEM's are smaller files that I've already got into ArcGIS so could I do what I want to do with a DEM or is it no good? Is there even any tool in ArcGIS that automatically can recognize how much is forest, water etc. on a map or is visual studying of the maps the only way?

I'm definitely no wizard with ArcGIS so I do want to make it as simple as possible. Maybe someone else has a better option to suggest?. Thank you!

D_McRitchie · Answer

It might be worth considering Deep Learning for this workflow as there are a range of pre-trained models on the Living Atlas that you can use for identifying specific landuse or objects.

For example of the following might help

Segment anything

Land Cover Classification Landsat

Land Cover Classification Sentinel 2

These can be computationally intensive to run so it could be worth splitting your data into grids and specifying the processing extent to one grid at a time.

Hope that helps,

David

Ed_ · Answer

Hi @m1unch no I don't think you can use DEMs as they only have elevation information. You can use supervised classification analysis to automatically determine forest, water etc. These two processes will however first require you to create training data.