Comparative Study of SVM and Naïve Bayes for Mangrove Detection Using Satellite Image

Abstract

Mangroves are a kind of plant which assumes an extremely fundamental job for security of our biological system. We presented the better approach for mangrove discovery by utilizing the help vector machine (SVM) and Naïve Bayes both are going under managed AI, and this calculation is utilized to group the image. The high-goals satellite information from Google earth is procured from an alternate locale of Mumbai, Maharashtra district, for recognition of mangroves. This exploration paper utilized two unique calculations, for example, Naïve Bayes classifier and Support Vector Machine for the discovery of perusing highlights from satellite images, and there are two calculations which are actualized utilizing the Matlab recreation tool stash. Support Vector Machine and Naïve Bayes are a directed grouping strategy applied on satellite image. In the wake of applying the calculations on the picture satellite, the precision of classifiers is determined utilizing perplexity grid and kappa coefficient. The execution of both methods of Support vector machine and Naive Bayes generate the detected area of mangrove in Mumbai, Maharashtra region. Exactness of Naïve Bayes saw as 99% with kappa value 0.9831, and the precision of help vector machine saw as 97% with a kappa estimation of 0.9631. The precision figuring utilizing disarray lattice and kappa coefficient shows that the Naïve Bayes classifiers is superior to help vector machine for the discovery of mangroves utilizing satellite picture.