Comparison bayes and knn classifier in satellite images

Apply

Comparison bayes and knn classifier in satellite images

Journal Title: Journal of Science and today’s world - Year 2014, Vol 3, Issue 1

Abstract

Remote sensing provides important coverage, mapping and image classification of landcover features, such as vegetation, soil, water and forests. Therefore, The choice of type of classifier can have a big impact on the accuracy of the classification. We are experimenting with two supervised classifications. The classifiers chosen are some of the common classifiers used in most practical applications. K-nearest neighbor is a very simple classifer. To classify a new test point one simply finds the k closest training data points in the predictor space. One then classifies to the class, which corresponds to the largest number of these training points; And the second classifier that was chosen is Bayes classifier, as it is one of the simplest non-parametric classifiers that are given high classification accuracy. A Bayes classifier is a simple probabilistic classifier based on applying bayes theorem (from Bayesian statistics) with strong (naive) independence assumptions. A more descriptive term for the underlying probability model would be "independent feature model". Here we compare two different classification methods and their performances. It is found that bayes classifier performed the best in our classification.

Authors and Affiliations

I. Elyasi, R. A. Sadeghzadeh

Keywords

remote sensing image classification bayes classifier k-nearest neighbor (knn)

EP ID EP29409
DOI -
Views 364
Downloads 5