Time domain features in combination with a support vector machine classifier for constructing the termite detection system

1st International Conference on Food Security and Sustainable Agriculture in The Tropics

农业科学;经济学

Nanda, M.A.^1 ; Seminar, K.B.^1 ; Nandika, D.^2 ; Maddu, A.^3

Department of Mechanical and Bio-system Engineering, Bogor Agricultural University, West Java, Bogor

16680, Indonesia^1

Department of Forest Products, Bogor Agricultural University, West Java, Bogor

16680, Indonesia^2

Department of Physics, Bogor Agricultural University, West Java, Bogor

16680, Indonesia^3

关键词: Acoustic features;    Acoustic signals;    Numerical results;    Support vector machine classifiers;    SVM classifiers;    Termite attack;    Termite detection;    Time domain features;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/157/1/012037/pdf
DOI  :  10.1088/1755-1315/157/1/012037

学科分类：农业科学（综合）

来源: IOP

【 摘 要 】

Over the last decade, it has been broadly reported that wooden buildings have been massively degraded due to termite attacks. The termite detection system based on the acoustic signal has been proposed to overcome such termite attacks. In this study, we investigate the implementation of the support vector machine (SVM) at the termite detection system. In this work, the pine wood as the medium for termite infestation was divided into two groups, i.e., the wood infested by termites Coptotermes curvignathus ('infested') and the normal wood ('uninfested'). The acoustic signal from each group was analyzed to produce the acoustic features, i.e., energy (E) and entropy (H). Subsequently, the acoustic features were included to construct the SVM Classifier. According to the numerical results, the SVM classifier achieved an accuracy of 93.21 ± 2.58%.