A review on classification of satellite image using Artificial Neural Network (ANN)

Abstract

Artificial Neural Networks (ANNs) have been useful for decades to the development of image classification algorithms applied to several different fields. Image classification is the major component of the remote sensing to extract some of the important spatially variable parameters, such as land cover and land use (LCLU). The aim of this study is to investigate the capability of Artificial Neural Network system (ANNs) for classifying the satellite images using different algorithm which are back-propagation algorithm and K-means algorithm with different approaches. ANN's classifier is compared with two classification techniques of conventional classifier which are Maximum Likelihood (ML) and unsupervised (ISODATA). Neural network classification is based on the training data set and it the proper classification. ML and ISODATA classifiers are broadly used in many remote sensing applications. Overall classification accuracy and Kappa Coefficient were calculated to get the comparison of the performance the image classification. The optimal performance would be identified by validating the classification results with ground truth data. The accurate classification can produce the correct LU/LC map that can be used fir variety.