Detecting Rice Weevils in Stored Grains using MFCC and CNN

Abstract

Sitophilus oryzae or commonly known as Rice Weevil is a pest that infests stored rice grains. The feeding and movement of this pest within the grain produce sounds that are audible to a high-performance microphone sensor. This study presents a method for early detection of Rice Weevils based on the frequency of the sound they produce. Sound features were extracted using MFCC and utilized as input features to train a Convolutional Neural Network (CNN). The classifier was evaluated using 2800 acoustic samples containing positive and negative datasets. The experimental results show that the method can be an alternative approach for the early detection of the presence of rice weevils in stored rice grains.