SVM based target classification using RCS feature vectors

Travis D. Bufler; Ram M. Narayanan; Traian Dogaru

doi:10.1117/12.2176759

21 May 2015 SVM based target classification using RCS feature vectors

Travis D. Bufler, Ram M. Narayanan, Traian Dogaru

Proceedings Volume 9461, Radar Sensor Technology XIX; and Active and Passive Signatures VI; 94610I (2015) https://doi.org/10.1117/12.2176759
Event: SPIE Defense + Security, 2015, Baltimore, MD, United States

Abstract

This paper investigates the application of SVM (Support Vector Machines) for the classification of stationary human targets and indoor clutter via spectral features. Applying Finite Difference Time Domain (FDTD) techniques allows us to examine the radar cross section (RCS) of humans and indoor clutter objects by utilizing different types of computer models. FDTD allows for the spectral characteristics to be acquired over a wide range of frequencies, polarizations, aspect angles, and materials. The acquired target and clutter RCS spectral characteristics are then investigated in terms of their potential for target classification using SVMs. Based upon variables such as frequency and polarization, a SVM classifier can be trained to classify unknown targets as a human or clutter. Furthermore, the application of feature selection is applied to the spectral characteristics to determine the SVM classification accuracy of a reduced dataset. Classification accuracies of nearly 90% are achieved using radial and polynomial kernels.

Citation Download Citation

Travis D. Bufler, Ram M. Narayanan, and Traian Dogaru "SVM based target classification using RCS feature vectors", Proc. SPIE 9461, Radar Sensor Technology XIX; and Active and Passive Signatures VI, 94610I (21 May 2015); https://doi.org/10.1117/12.2176759

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available