Geographic Information Systems Stack Exchange is a question and answer site for cartographers, geographers and GIS professionals. It only takes a minute to sign up.
Sign up to join this community
I am trying to calculate how much of the electric feeders exist underneath tree covers. I am using NAIP infrared satellite imagery and have a shp file for a feeder circuit. Like so:
I need to find out how much of my green lines actually exist underneath the vegetation cover, and do segmentation of the line shape file based on that information. I have looked at deep learning capabilities within ArcGIS.
I am wondering if there is a simpler way of doing this using ArcGIS?
A simple approach would be to use NAIP imagery to calculate NDVI with the red and NIR (near infrared) bands and set a threshold over which you consider the electrical feeders to fall in a vegetated area. The formula for NDVI is NDVI = (NIR — VIS)/(NIR + VIS) where VIS is the red NAIP band. You can calculate this in ArcGIS using raster math.
However, this doesn't tell you whether the vegetation is tree vs. grass or shrub. Even if you were to successfully identify trees in the imagery using a neural network approach, you wouldn't know whether the feeders are over or under the trees. There would also likely be confusion between trees and shrubs which would also complicated the result. You might consider whether LIDAR data is available to determine vegetation height, which would help with these two concerns.
Once you have identified locations where trees exist (using either simple NDVI or an indicator of height), you'll need to intersect these to identify lines that intersect with the trees. To do this, you can use Raster to Polygon to convert your tree data set from raster to polygon and then Intersect the tree polygons with the feeder network to determine feeders that intersect with trees.
