Translator Disclaimer
18 September 2003 Vehicle counting and classification algorithms for unattended ground sensors
Author Affiliations +
Abstract
Unattended ground sensor technology used for battlefield awareness and other wide area surveillance applications requires state-of-the-art algorithms to address the unprecedented challenges faced in detecting, classifying and tracking military combat vehicles. The performance of traditional acoustic sensor systems often degrades unacceptably against the dynamic and highly mobile multiple target environments in which today's forces must operate. In the present work, a target counting algorithm has been developed to solve problems attributed to unstable tracking performance by resolving track loss deficiencies inherent to closely spaced target environments. The algorithm provides a way to discriminate between vehicles as they pass through an acoustic "trip-line" formed by a sensor in a predetermined field of view (FOV). The proposed approach is realized through an adaptive beamforming algorithm that achieves enhanced directivity in a principal look direction by significantly reducing the effects of interferers outside the precise bearing of the steering direction. The classification algorithm described herein facilitates a minimal representation for features extracted from harmonically related structures characteristic to acoustic emissions from ground vehicles found in battlefield environments. The reduced feature space representation exploits an ordering for the principal narrowband components found in a vehicle's engine noise and has proven effective in solving fundamental problems associated with discriminating between vehicles in variable SNR environments. The performance of the algorithms is demonstrated using signature data collected during various acoustic sensor field test experiments.
© (2003) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Myron E. Hohil, Jeffrey R. Heberley, Jay Chang, and Anthony Rotolo "Vehicle counting and classification algorithms for unattended ground sensors", Proc. SPIE 5090, Unattended Ground Sensor Technologies and Applications V, (18 September 2003); https://doi.org/10.1117/12.488127
PROCEEDINGS
12 PAGES


SHARE
Advertisement
Advertisement
Back to Top