Classification of raster data

Classification refers to classifying data to different categories; in the case of remote-sensing literature, this refers to classifying different land cover types, generally. This can be done in two ways: supervised and unsupervised. Supervised classification refers to a situation where we create a training area and generate a signature area from the training area, and then use that to classify a raster. In unsupervised classification, the user determines into how many clusters they want to divide the data beforehand, and then the image is classified into that number of clusters; finally, the user then identifies the relevant land classes for the clusters.