Value Added Data Analysis with ArcGIS Pro

Introduction

    This week, I worked through the ArcGIS tutorial: Calculate Impervious Surfaces from Spectral Imagery. Impervious surfaces are impenetrable by water and can have detrimental environmental effects. One way to identify such surfaces is to segment and then classify aerial imagery within ArcGIS. 

Methodology

Step 1) Segmenting the Imagery

    To determine what is impervious vs pervious, sections of the aerial imagery must be identified as such. First I extracted the spectral bands using the “Extract Bands” Raster function. This classifies segments of the imagery based on color which simplifies further classifications.

Figure 8.1: Initial segmentation of the aerial imagery

Step 1) Classifying the Imagery

    I then used the “Classification Tools” to perform a supervised classification using training samples that I gathered myself. The training samples consisted of polygons drawn on known areas which helped the algorithms to identify like areas. Once the processing was completed I used the “Reclassifier” to touch up some errors that were misclassified.

Figure 8.2: Training samples used to classify imagery

Figure 8.3: Final classification of imagery

Conclusion

    Segmenting and classifying imagery is useful for digesting aerial images beyond just finding impervious surfaces. It can be used to identify pretty much anything that can be picked up in a band with the appropriate sensor. I have seen this used in everything from finding water in a landscape or determining the progress of plant growth. This can be used for agricultural purposes, lands planning, and many more applications.