Fusing Multi-feature will benefit radar target classification with much more belief. However, since radar target attributes such as high range resolution profiles, waveforms, frequency spectra, time-frequency spectra, higher order statistics, polarization spectra and flight path are of different dimensions, it is hard to make decision by fusing multi-feature directly. Fortunately, rough set makes decision by examining the fitness of each condition attribute separately, while neural network is powerful for dealing with nonlinear problems. With radial projection of target dimension, cruising velocity and height as condition attribute, a multi-feature rough neural network fusion classifier is presented. Simulation of the proposed classifier based on an information system with 25 targets belonging to 6 classes shows accuracy not less than 93%, while attributes are subjoined with typical radar errors. © Springer-Verlag Berlin Heidelberg 2005.
CITATION STYLE
Shi, Y., Ji, H., & Gao, X. (2005). A radar target multi-feature fusion classifier based on rough neural network. In Lecture Notes in Computer Science (Vol. 3497, pp. 375–380). Springer Verlag. https://doi.org/10.1007/11427445_61
Mendeley helps you to discover research relevant for your work.