Image classification

Abstract

Image classification refers to the procedure of assigning each pixel in an image to a particular land use/cover class of interest. The purpose of this chapter is to perform supervised classification using single date and multidate Landsat 5 TM imagery and machine learning methods. Five machine learning methods such as k-Nearest Neighbors (KNN), Artificial Neural Networks (ANN), single Decision Trees (DT), Support Vector Machines (SVM) and Random Forests (RF) machine learning classifiers will be used for image classification. The tutorial exercises show that multidate Landsat 5 imagery and the RF method provide relatively good results.