Unsupervised an Supervised Classification

 

    For lab 5, I learned the difference between unsupervised and supervised classification in ERDAS Imagine. In exercise 1, I completed an unsupervised classification. I identified individual pixels and assigned them to a feature (trees, roads, shadows, etc.). I did this by using the attribute table to change the colors of the pixels to colors that are easier to identify for each feature. In exercise 2, I created an AOI image to collect spectral signatures. First, I used the inquire cursor to identify different coordinates and create polygons to highlight them. I used the signature editor to name them properly. I then used region growing properties to create points of interest, and used the inquire (legacy) tool to type in exact coordinates. In exercise 3, I used histograms to evaluate signatures. 

    I enjoyed this lab because I was able to label specific coordinates with specific features, such as land type/land use. I had trouble at the end of the lab for the deliverable, because I could bring the map into ArcGIS to create the final layout because my file was a .sig file. I was running out of time to submit the lab, so I took screenshots of the map and signature editor, and typed all of the information I needed into a pdf document. In the future, I plan to read up more on how to bring .sig and AOI data over to ArcGIS.