Image Classification

Problem

The managers of the Black Water Wildlife refuge would like information regarding land cover in order to develop their management plan.

Analysis Procedures

Methods

First I change the image symbology from true-color to color-infrared to faciiliate identifcation of the 6 landcover types represented in the image: Forest, Cultivated Field, Barren Area, Developed/Impervious, Wetland and Water. I then created some training samples, saving them as a shapefile, and then I ran the Supervised Classification using those training samples as reference. After reviewing the result, I then exported it to a .tiff file and using the Build Raster Attribute table tool I calculated the area of each class. I then reran the entire process selecting some additional training sample polygons on areas that were poorly classified in the initial procedure.

Results

Application & Reflection

In this assignment, I learned how to classify landcover from raster images. I can imagine how useful this would be for the city of Raleigh’s planning office because they have this “infill” procedure in their permitting process where your structure cannot be more than a certain ratio of your property. By using an aerial image before and after the build you could classify it by structure and not structure and the ratio calculation could be performed easily to determine which properties meet compliance on that.