One of the common post-processing steps applied to satellite images is classification. In most cases, classification is performed to reveal the patterns in land cover in the area of interest. Since a satellite image is made up of pixels, image classification aims at (semi)-automatically categorizing pixels into different land covers. In agricultural applications, the purpose of classification is commonly to reveal the spatial distribution of crops.
Two types of technique can be adopted in categorizing pixels into classes (Lillesand et al., 2004). They are:
- supervised and
- unsupervised approaches.
The difference between these types lies in how a classifier (classification algorithm) is instructed to assign categories to pixels.