Module 4 Challenge: Supervised Classification

The link posted here is to a map created using ERDAS IMAGINE 2010 software as part of the University of West Florida On-line GIS Certification program class, Photo Interpretation and Remote Sensing (GIS4035/L). The link requires Internet Explorer to open and display properly.

Germantown Maryland

The image is a Supervised Maximum Likelihood classification from a satellite image of Germantown, Maryland classified for 14 land use types identified by spectral signatures identified using the ERDAS IMAGINE software. The process involved creating a unique spectral signature for each land use type by first creating an area of interest polygon around a specific known feature on the map in an attempt to capture pixel values that are unique to the feature and thus to the type of land use. Once the distinct signatures were created the software could reclassify the entire image using the signatures to show each of the land use type.

In theory the task was not too complicated. However, knowing when a specific signature covered enough pixels in the image was not an easy task, particularly after the assignment was changed to require that each land use contain pixel values in a certain range. In the end I got 12 of 14 within the proper ranges and was only off on the other two by a relatively small amount of pixels. However, after reclassifying signatures multiple times and regenerating the classified map 16 times, I feel I gave a tremendous effort to accomplish a task that just did not seem worthwhile as each successive classification once a certain point had been reached seemed to have a negligible effect on the image as a whole. In that respect I'm satisfied with the map regardless off the possible mark downs for these two classes that were only slightly out of the challenge parameters.