Lab #8: Calculating Impervious Surface Area

Introduction

Lab 8 goes over image classification, training samples, and segmentation. The tutorial I used to create these maps can be found here. This tutorial is very useful and a great step by step process. Classification helps distinguish what the surface of the map really is, rather than the physical appearance it has. This map helps people see a large landmass, using polygons and colors to display simple information.

Methodology

This first image is the image we started with. we used the data the tutorial gave us to alter this image.

Figure 1: Tutorial Base Map

The first step was to use the raster function "extract bands" that simplifies the image with colors that makes it easier to see objects. I used the bands in 4,1,3 order with the "best match" that helps match the surfaces to a band.

Figure 2: Extract Bands Tool

Figure 3: Extract Bands Image

The next step was to use the "Classification Wizard" tool to create a segmented layer. This classification uses a process called segmentation to group neighboring pixels based on the similarity of their spectral characteristics.

Figure 4: Classification Wizard Tool

Figure 5: Classification Wizard Methods

The classification method we used was supervised with an object-based classification type and the default schema.

In the next step, we created training samples and classified the image using the "Training Samples Manager" page of the Classification Wizard tool. We added new classes naming them impervious or pervious.  Then added a name, value, and color. After we created the two main classes we made subclasses for the roofs, roads, driveways, bare earth, grass, water, and shadows.

Figure 6: Parent Class 

Figure 7: Subclasses

Figure 8: Subclasses 

Results 

Finally, we classified the image using the "Support Vector Machine" tool which can handle larger images and is less susceptible to discrepancies. This created a new layer using the training samples I created earlier.

Figure 9: Surface Classification Result

This map isn't perfect it has some flaws with the coloring and locations of each sample. This is where you would take the time to go into the map and clean up each object in the map to make it a clean and presentable map.

After the map was processed I finally went in and used the "Reclassify with a Region" tool.

Figure 10: Reclassify Within a Region

I then remapped classes, confirmed the current class to any, and changed the new class to pervious.

Figure 11: Remap Classes

The tutorial asked me to draw this polygon using the remap class.

Figure 12: Remap Class Polygon

The map was then automatically reclassified as a pervious surface

Figure 13: Reclassified as a Pervious Surface

In this last final step, I created a Final Classified Dataset, ran it, and clicked finish. I also created a map of the extracted bands map in ArcGIS Pro.

Figure 14: Final Classified Dataset

Figure 15: Extracted Bands ArcGIS Pro Map

Conclusion

The results from this tutorial helped me understand how great of a tool classification is. This tool helps create a simplified map that people can use in construction, real estate, or agriculture. This skillset I learned will come in handy if I have to create maps for a future employer, and I am happy that I got to take the time to sit down and learn this fundamental tool.