SVM – A Method for Classification of Pests on Palms

Abstract

Currently SVM (Support Vector machine) is used in Object detection, Content based image retrieval, Text recognition, Biometrics, Speech recognition etc. In this article we describe an application of SVM to classify two different sounds originated from two different species using linear discriminate function. Among the various varieties of disease causing pests present on palms RPW (Red Palm Weevil), RB (Rhinoceros Beetle), leaf eating caterpillar and nematodes, we take special interest in classifying the first two species. The most economic pest is the RPW whose infestation is irreversible and is of most concern. In this work, we deal with only two pests the RPW and RB. The result of classification helps in early detection of presence of the particular species. During field inspection, sound captured by a digital voice recorder can be stored as a sound file for analysis which can be later used for SVM technique. This method gives high classification accuracy.