Discriminant Analysis as a Tool for Detecting the Acoustic Signals of Termites Coptotermes Curvignathus (Isoptera: Rhinotermitidae)

Abstract

Various methods for termite detection have been developed, one of which is purely based on their acoustic signals. However, this method has a weakness, as it is difficult to separate the signals generated by the termites from noise in the environment. A combination of the feature extraction of the acoustic signals and a classification model is expected to overcome this weakness. In this investigation, we inserted 220 subterranean termites Coptotermes curvignathus into pine wood for feeding activity and observed their acoustic signals. In addition, three acoustic features (shortterm energy, entropy and zero moment power) were proposed to recognize the termites' acoustic signals. Subsequently, these features were analyzed and combined with discriminant analysis to produce a robust classification model. According to the numerical results, the integrated discriminant analysis and the acoustic feature in our termite detection system has an accuracy of 83.75%.