Artificial target structure type inversion method based on bistatic polarization characteristic vector

Abstract

The invention provides an artificial target structure type inversion method based on a bistatic polarization characteristic vector, and the method comprises the following steps: 1, building a CAD model library of a template part; 2, acquiring electromagnetic scattering simulation data of the template component; 3, constructing a bistatic polarization characteristic vector; 4, training a neural network; 5, scattering center detection is carried out on the ISAR image of the artificial target; and 6, predicting the component structure type of the scattering center by using a neural network. The target component structure inversion method is applied to the field of polarization radar target identification, and solves the problems of low robustness and low identification degree of a traditional coherent polarization decomposition inversion target component structure type. Through simulation verification, the method provided by the invention is higher in accuracy on the premise of not increasing the operation complexity and the calculation time, and has very high engineering application value.

Description

【Technical field】 [0001] The invention relates to a method for inverting the structure type of an artificial target by using an ISAR image, which belongs to the field of polarized radar target recognition, and specifically relates to the artificial target ISAR image acquired, and then the neural network prediction and detection completed by using bistatic polarization feature vector training The structure type of the target part corresponding to each scattered center is obtained, so as to realize the method of cognition of the overall structure type of the artificial target. 【Background technique】 [0002] Different target structures have different polarization effects. Therefore, polarimetric radar can invert the target structure type, thus providing rich information for target recognition and cognition. The artificial target structure is stable, and the scattering intensity changes slowly with time and space; if the resolution of the polarimetric radar is sufficient, there...

AI Technical Summary

Problems solved by technology

However, the target model adopted by the current coherent polarization decomposition method is relatively simple, which can only distinguish the extreme cases of odd and even scattering and dipole target models. At the same time, the coherent polarization decomposition inversion method often obtains the structure type that the actual target itself does not exist, which creates difficulties for the cognition of the target

[0003] Aiming at the problem that the current polarization decomposition method is not robust enough to invert artificial target structure types, and the recognition degree of typical structure types is not high, this invention breaks the traditional research ideas of polarization decomposition, and based on typical component structure types such as dihedral angle, , trihedral angle, dipole, helix, plate and cylinder quasi-monostatic scattering data, a 5-dimensional bistatic polarization eigenvector is proposed to extract and characterize the polarization scattering characteristics of the target, and the classification model is obtained by neural network training , so as to realize the effective classification and identification of the structure types of different parts of the scattering center

Images