In my remote sensing class, I transform a Landsat satellite image of the Roanoke area into a land cover map. I used three classification methods, cluster unsupervised classification, maximum likelihood supervised classification, and minimum distance supervised classification. I split the land cover into five categories, urban, agriculture, forest, water, and soil. To measure if my supervised method were accurate I ran a signature mean plot and kappa statistics using random point sampling. The outcome is my supervised process was accurate except for the soil was hard to separate between urban.
Roanoke Land Cover Classification
Tools used
Erdas Imagine
Plug-ins used
iso cluster unsupervised classificationmaximum likelihood classificationminimum distance classification
tags
image classification
More by Jason Boatjri
You might also like
Join the community!
We're a place where geospatial professionals showcase their works and discover opportunities.