Using Wave-Coefficients as Feature Vectors to Identify Aerospace Targets

Abstract

Abstract-An original target identification method using WaveCoefficients (WCs) as feature vector is proposed. The scattering fields of arbitrary shaped targets are expressed as a sum of spherical waves and the distinctive coefficients are exploited as the target feature. Decision rule based on correlation coefficient is established, and some analyses on the properties of the WCs are conducted. Numerical simulations of four targets are carried out and the recognition performances without and with noise are provided and discussed.

Using Polarization-Diverse Wave-Coefficients for Aerospace Target Recognition

“…The targets' sensitivity to azimuth angle is totally embodied in the WCs. Simulation results in [22], [23] show that the WCs' sensitivity to azimuth angle is directly proportional to the working frequency. Due to the properties ofˆ( ) n J  and (2) ( ) n H  , the dimension of the WCs is augmented with increasing of the operating frequency.…”

“…Thus, the operating frequency should be appropriately chosen to strike a balance between target information and sensitivity to the azimuth angle. A frequency band in resonance region is proposed in [23], in which the WCs are able to tolerate big aspect variation and contain enough target information at the same time. To take full advantage of different polarization channel WCs, a multiple polarization WCs method is proposed in this work to recognize aerospace targets.…”

“…Two-dimensional (2D) target identification with a feature vector termed wave-coefficients (WCs), was proposed in [22]. WC concept was extended to threedimensional (3D) targets in [23], where four aerospace targets were considered. Experimental results show that the WCs are with increased sensitivity to aspect angle at higher frequencies.…”

“…However, if the operating frequency is chosen too low, the targets are equivalent to a scattering point, and WCs based target recognition will lose effectiveness. An operating frequency range in resonance region was proposed in [23], in which the WCs can tolerate big aspect variation, bringing a decided advantage in a large number of target features. Because there are few target features for aerospace target which are able to provide reliable target recognition in resonance region.…”

“…Besides, the WCs method has merits such as flexible application, easy acquisition, etc. However, literature [23] only took one single polarization channel WCs into account. At some aspect angles, the WCs are not able to provide enough distinction, resulting in a comparatively poor recognition performance.…”