2dview.jpg

Image Classification

When you want to map ground features (what is corn, what is wetland, what is forest), you can use color multi-spectral satellite images and airphotos to separate feature classes. Standard image classification models work on the spectral components of each ground cover in remotely sensed imagery. Some methods are completely automatic; others take advantage of your ability to identify known ground cover in sample locations.

navigate.jpgunsupervised classification

navigate.jpgclass / distance rasters

navigate.jpgmasking processing area

navigate.jpgFuzzy C Means

navigate.jpgmaximum likelihood

navigate.jpgminimum distance

navigate.jpgusing ground truth

navigate.jpgK Means

navigate.jpgco-occurrence analysis

navigate.jpgISODATA classification

navigate.jpgclassification dendrograms

navigate.jpgstepwise linear

View this Getting Started booklet: classify.pdf Image Classification (file date: )