Lab 10 - Supervised Image Classification

GIS 4035

 

 

The map above is of Germantown, Md.  In this weeks project we were to work with Supervised Image Classification.  I worked to create spectral signatures and AOI features within EDRAS.  I was then able to produce a classified image from the satellite image that identified key areas of Land Use so they could be monitored moving forward.  This exercise also reinforced the ability to identify spectral confusion by identifying features that might have similar characteristics, but when viewed in different bands became more unique to identify.

 

Bringing all these elements together I was able to produce the above map, and show where there may be discrepancies in the classification with the distance inset provided.

 

Overall I enjoyed this lab, and it allowed me to gain a better knowledge of how images can assist in land classification through automated processes and tools in the EDRAS program.



Lab 9: Unsupervised Classification

GIS 4035

 

 

In Lab 9 we let the computer do some of the work.  Unsupervised Classification was my introduction to automated classification of an image based off of different spatial and spectral resolutions.  There was a little manual work to be done, but overall the program assisted greatly.  I was able to use both ArcGIS as well as ERDAS to reclassify and code images simply based off of the pixel colors and grouping like pixels together. 

 

In ArcMap I utilized tools like Maximum Likelihood Classification tool, and the Iso Cluster tool, while in ERDAS I utilized tools such as Unsupervised Classification found in the Raster tab, Classification group.

 

Ultimately using the tools provided I was able to generate the map above which shows each Class by name, and from this I was able to calculate the total Permeable and Impermeable surfaces.