Performance Analysis of the Scenario-Based Construction Method for Real Target ISAR Recognition

Abstract

Due to the difficulty in estimating the 2D image plane of the inverse synthetic aperture radar (ISAR) image, we recently proposed a new paradigm to construct the training database based on the flight scenario. However, because the flight condition for the training and the test data was identical, much more study is required for this method to be applied to the real ISAR scenario. This paper presents a study on the factor that can affect the applicability of scenario-based method to the real target ISAR recognition. Simulation results using five scatterer models show that accurate measurement of flight direction and aspect angle variation are required and enough bandwidth larger than 200 MHz should be guaranteed for the successful classification.