Tuesday, October 8, 2013

Module 10 Lab: Supervised Classification


In module 10 we were tasked with adding unique values to all pixels in a raster image. Supervised classification involves collected sets of pixels to define spectral signatures. I then had to evaluate the accuracy and use them to classify the entire image. Once I got the signatures recoded into eight specific classes I was able to notice some spectral confusion with in the roads signature. The area was far too large for what was actually in the given image. 


No comments: