23 May 2011 Robust discrimination of human footsteps using seismic signals
Author Affiliations +
This paper provides a statistical analysis method for detecting and discriminating different seismic activity sources such as humans, animals, and vehicles using their seismic signals. A five-step process is employed for this purpose: (1) a set of signals with known seismic activities are utilized to verify the algorithms; (2) for each data file, the vibration signal is segmented by a sliding-window and its noise is reduced; (3) a set of features is extracted from each window of the signal which captures its statistical and spectral properties. This set is formed as an array and is called a feature array; (4) a portion of the labeled feature arrays are utilized to train a classifier for discriminating different types of signals; and (5) the rest of the labeled feature arrays are employed to test the performance of the developed classifier. The results indicate that the classifier achieves probability of detection (pd) above 95% and false alarm rate (pfa) less than 1%.
© (2011) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Aram E. Faghfouri, Aram E. Faghfouri, Michael B. Frish, Michael B. Frish, } "Robust discrimination of human footsteps using seismic signals", Proc. SPIE 8046, Unattended Ground, Sea, and Air Sensor Technologies and Applications XIII, 80460D (23 May 2011); doi: 10.1117/12.882726; https://doi.org/10.1117/12.882726

Back to Top