Identifikasi Hama Ulat Kubis Menggunakan Transformasi Wavelet dengan Klasifikasi Probabilistic Neural Network

Abstract

This research proposes a method for moth pests identification using wavelet transformation and Probabilistic Neural Netwok (PNN). Wavelet transformation method is used to get image features. This study focused on 3 types of moth pests that Crocidolumia binotalis zeller, Spodoptera exigua (Hubner), and Spodoptera litura F. To get the best performance of classifier was analyzed by confussion matrix and k-fold cross validation with k = 5. Result for identification of moth pests diamondback in this study showed an average accurancy of 80.74%. This result shows that the wavelet transformation and classifier Probabilistic Neural Network can be applied for the identification of cabbage caterpillar pests