TY - GEN
T1 - Radar hedeflerinin siniflandirilmasinda saçici merkezlerin kestirimi için yeni bir karma yaklaşim
AU - Gültekin, Özgür
AU - Günel, Tayfun
AU - Erer, Işin
PY - 2006
Y1 - 2006
N2 - Radar images, range profiles and scattering centers are used as feature parameters in radar target classification applications. Scattering center parameters, when used as feature parameters, enable an efficient compression of feature space compared to classical target classification methods based on radar images and range profiles. A method used for the estimation of scattering centers via cancellation of side lobes is the CLEAN algorithm. In this work, model based Prony, MUSIC, ESPRIT and evolutionary based CLEAN methods are applied for the estimation of scattering centers. A hybrid method is proposed which improves the convergence of evolutionary based CLEAN. Scattering centers which are estimated by aforementioned methods are classified using correlation based matching score method, Bayes classifier and artificial neural networks. Classification is accomplished using simulated data of four different aircraft models created by the point target model at different frequency bands and aspect angles.
AB - Radar images, range profiles and scattering centers are used as feature parameters in radar target classification applications. Scattering center parameters, when used as feature parameters, enable an efficient compression of feature space compared to classical target classification methods based on radar images and range profiles. A method used for the estimation of scattering centers via cancellation of side lobes is the CLEAN algorithm. In this work, model based Prony, MUSIC, ESPRIT and evolutionary based CLEAN methods are applied for the estimation of scattering centers. A hybrid method is proposed which improves the convergence of evolutionary based CLEAN. Scattering centers which are estimated by aforementioned methods are classified using correlation based matching score method, Bayes classifier and artificial neural networks. Classification is accomplished using simulated data of four different aircraft models created by the point target model at different frequency bands and aspect angles.
UR - http://www.scopus.com/inward/record.url?scp=34247160400&partnerID=8YFLogxK
U2 - 10.1109/SIU.2006.1659770
DO - 10.1109/SIU.2006.1659770
M3 - Konferans katkısı
AN - SCOPUS:34247160400
SN - 1424402395
SN - 9781424402397
T3 - 2006 IEEE 14th Signal Processing and Communications Applications Conference
BT - 2006 IEEE 14th Signal Processing and Communications Applications Conference
T2 - 2006 IEEE 14th Signal Processing and Communications Applications
Y2 - 17 April 2006 through 19 April 2006
ER -