GIS 4930: Special Topics; Project 2: MTR Analyze

This week we were tasked with classifying areas of Mountaintop Removal (MTR) and NonMTR on four landsat images for our group's study site.  There are four groups to sign up for in which I decided to become a part of Group 1. The fact that there are four landsat images and four members in my group, we all took responsibility for classifying one image.  

The first step was to create a single raster dataset of seven landsat bands that pertained to my image with the Composite Bands tool in ArcMap.  Next, the Extract by Mask tool was used to create a raster of portion of the composite landsat that fell within the study area.  At this point, the image was ready to be classified into two categories: MTR and NonMTR.  This step was accomplished by using a new software known as ERDAS.  Within ERDAS, the Unsupervised Classification tool was selected, creating 50 classes from the masked image.  Following this, the MTR and NonMTR pixels were selected and titled as such in the attribute table.  I then saved the layer and added it into ArcMap.  The raster was then reclassified with all areas other than the MTR areas as No Data. This was then converted into a polygon.  My results from this assignment are shown below. 

River banks, roads, and flat mountainous regions have similar spectral reference characteristics as MTR sites, so a portion of the red area is not accurate. This will be corrected next week during the Report Week assignment.