Using arcpy with aRcGeo

Introduction

This short vignette is meant to demonstrate the utility of working with arcpy functions in the R environment. The following will be a series of processing functions applied to a raster file.

Loading arcpy

In order to showcase some examples of arcpy functionality in R, we will first need to import the python. This is where we will need to use the aRcGeo::init_arcpy() function to import the module.

library(aRcGeo)
#> The functions in this library require the user to have `arcgisbinding` loaded.
#>                       Please load `library(arcgisbinding)` before running any aRcGeo functions.

init_arcpy(conda_env = "arcgispro-py3-DeepLearning")

Using ArcGIS Pro’s Stretch and Segment functions

Importing Original Raster

We can import our original raster by using a file.path function. In this case, the imported file is in a .tif format.

Buchanan_Postal_Route_tif <- file.path(trim_path, "Postal_Routes_Selection.tif")

An RGB render of this image looks like the following:

Buchanan Postal Routes Original

Stretched Raster

We would now like to use the stretch function in the arcpy module to process the image. The following code applies a Percentage Clipping stretch with a minimum percentage value of 0.05 and a maximum percentage value of 85. This will greatly reduce the background noise of the image.

Stretched_Buchanan <- arcpy$ia$Stretch(Buchanan_Postal_Route_tif, 
                                       stretch_type ="PercentClip", 
                                       min_percent = 0.05, 
                                       max_percent = 85)
Stretched_Buchanan$save("Stretched Buchanan")

Our processed image now looks like the following:

Buchanan Postal Routes Stretched

Segmented Raster

Once the stretched image process is complete, the raster will also go through a segmentation process which we can call in the following function.

Segmented_Buchanan <- arcpy$sa$SegmentMeanShift(in_raster = Stretched_Buchanan, 
                          spectral_detail = 5,
                          spatial_detail = 15)
Segmented_Buchanan$save("Segmented Buchanan")

The applied segmentation function from the arcpy module has inputs of detail on a scale from 0-20. The following image is the output of a high spatial and low spectral segmented image.

Buchanan Postal Routes Segmented

Discussion

As the examples demonstrate, arcpy functionality is possible in R using R syntax to execute Python functions in R code chunks rather than having to call scripts. R is also able to recognize ESRI spatial objects, making conversions and compatibility with other R-spatial packages a possibility.