Target identification optimization method based on high resolution one-dimensional range profiles

Abstract

The invention discloses a target identification optimization method based on high resolution one-dimensional range profiles. The target identification optimization method based on high resolution one-dimensional range profile includes the steps: determining a training sample set which includes N high resolution one-dimensional range profiles, and successively calculating the third order spectrum signature F of the N high resolution one-dimensional range profiles and the third order spectrum signature X of the N high resolution one-dimensional range profiles after dimension reduction processing; determining the low density area and the high density area; calculating the third order spectrum signature type of M1 high resolution one-dimensional range profiles after dimension reduction processing in the low density area, and the third order spectrum signature type of M2 high resolution one-dimensional range profiles after dimension reduction processing in the high density area; respectively calculating the classification sub hyperplane parameter of the low density area and the classification sub hyperplane parameter of the high density area; determining the test high resolution one-dimensional range profile, and calculating the third order spectrum signature Fz of the test high resolution one-dimensional range profile so as to obtain the third order spectrum signature z of the testhigh resolution one-dimensional range profile after dimension reduction processing; and finally determining that the third order spectrum signature z belongs to the target or belongs to interference.

Description

technical field [0001] The invention belongs to the technical field of radar target recognition, in particular to a target recognition optimization method based on a high-resolution one-dimensional range image, which is suitable for the identification of targets outside the database and the classification of targets in the database in radar target recognition, enabling radar target recognition The overall performance of the system is effectively improved, and it is also suitable for practical engineering applications. Background technique [0002] In the radar target recognition environment, there are often a large number of interfering objects, which seriously affect the overall performance of the radar target recognition system; Effective classification of objects in the library; the single-class classifier can realize the identification of objects outside the library, and can effectively describe the feature space of the training data, such as the support vector domain de...

AI Technical Summary

Problems solved by technology

[0003] The attitude sensitivity of the high-resolution one-dimensional range image of the target causes the uneven distribution of the target feature space. When using the radar target recognition method of the support vector model, the uneven distribution of the feature space will lead to a large error in the classification boundary

Images