The use of seismic signals brings new challenges for the effective detection and classification of vehicles due to their complex nature. The main challenge in the efficient classification of vehicle is to construct a feature vector that optimally represents the main features of the seismic signatures of the vehicles. This paper proposes a new feature extraction algorithm which extracts features from time domain, frequency domain, time-frequency domain and power spectral density (PSD) of the seismic signatures of the vehicles. This robust and effective algorithm has guaranteed performance when it is deployed for real-time applications. Extensive performance evaluation is done under different conditions to derive optimal configuration of different parameters. Classification results obtained from varied feature set has improved classification accuracy above 95%.
AUTOMATED VEHICLE DETECTION AND CLASSIFICATION USING SEISMIC SIGNAL PROCESSING
2014-03-29
IJITR; Vol 2, No 2 (2014): February - March 2014; 850-853
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
DDC: | 629 |
Moving source localization using seismic signal processing
Online Contents | 2015
|NTRS | 1969
|Automated Accident Detection in Intersections via Digital Audio Signal Processing
Online Contents | 2003
|