Module 5- Photo Interpretation and Remote Sensing- Supervised & Unsupervised Classification

This week we focused on learning how to  perform supervised and unsupervised land use classification in ERDAS. As always I feel like I learned a lot. After practicing the basic we were tasked to create our own land use classification map. Upon examining my final map I feel like my road area is likely overestimated. It seems like some of the regular "urban" area overlaps with the road pixels based on my examination of the histograms of all the spectral signatures I added, but I was unable to find a layer that they were distinct enough on to provide a more accurate map. 

Supervised classification of current land use in Germantown, MD. Image depicts 8 different land use categories as well as area in hectares for each one. An inset map shows the distance file, with areas of likely misclassification displaying as brighter and areas of likelier accuracy showing darker.