Object Detection and Classification in Multichannel Seismic Network

M. Zubair, O.E. Löpprich, S.E. Ghobadi, K. Hartmann, and M. Koch (Germany)


Seismic, Footstep, Auto-correlation, K-mean clustering, Periodicity


Object detection and classification is an important problem for the outdoor surveillance applications. This paper explains the strategy for object detection and classification in a seismic network. The main emphasis is on person detection and differentiates from other objects like dog in a multichannel seismic test field. Identify footsteps pattern is an important feature for this application. Auto correlation technique is used to define the periodicity of the footsteps in each channel and the final decision is based on the information collected from all the channels. Also the K-mean clustering is applied to detect and reduce the false alarms in the multichannel seismic data.

Important Links:

Go Back