Infrasound signal processing

• Main Author: Yap, Kai En.
• Other Authors: School of Electrical and Electronic Engineering
• Format: Final Year Project
• Language: English
• Published: 2010
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Electronic systems::Signal processing
• Online Access: http://hdl.handle.net/10356/40359
• Institution: Nanyang Technological University

Infrasound is a low frequency acoustic phenomenon typically in the frequency range of 0.01 to 20 Hz. It has been used to monitor various man-made and natural events due to its inherent ability to propagate long distances. The detection and study of infrasound would greatly benefit society in a wide range of non-trivial applications. The purpose of this infrasound signal processing project is to establish a data acquisition system to capture and classify infrasound data. A detailed description of the equipment setup clarifies the methodology on how infrasound data is recorded. In addition, issues with the data collection process were identified and relevant measures were taken to overcome the problems. The infrasound data is preprocessed using techniques similar to speech processing, such as Mel-scale Frequency Cepstrum Coefficients (MFCC), to obtain a set of feature vectors which will be used to train and test the neural network. The benefit of this technique is that it is not affected by the record length, sampling frequency or the signal amplitude. A parallel neural network classifier bank is developed to classify infrasound events from six different classes of signals, where each module in the classification bank is a backpropagation neural network responsible for classifying one of the six events. For the six different infrasound events, the correct classification rate achieved is 92%.