Noise Robust Broadband Full Polarization Target Recognition Method Based on Polarization Decomposition

Abstract

The invention belongs to the technical field of radar target recognition, and particularly relates to a steady-noise complete-polarization broadband target recognition method based on polarization decomposition. The method includes the following steps that each set of complete-polarization channel distance image signals used for training serves as a corresponding training sample, and an average distance image and classification features of each corresponding training sample are obtained; an average distance image and classification features of each corresponding test sample are obtained; the average distance image of each corresponding test sample is aligned to the average distance image of each training sample; fusion features of the corresponding test sample of each training sample are obtained; classification and recognition are carried out on the corresponding test samples according to the fusion features of the corresponding test sample of each training sample and the classification features of each corresponding training sample.

Description

technical field [0001] The invention belongs to the technical field of radar target recognition, in particular to a noise robust broadband full-polarization target recognition method based on polarization decomposition, which can be used to classify different types of aircraft or vehicle targets. Background technique [0002] Broadband radar target recognition can use the high-resolution radar echo signal of the target to realize the judgment of the target type. High-resolution radar target echo can reflect more detailed information of the target, and has unique advantages that low-resolution radar does not have. The high-resolution range image provides the geometric structure information of the target along the distance direction, which is easy to obtain and process, and has a good application in target recognition. Under the single-polarization system, the polarization mode of the radar transmitting and receiving electromagnetic waves is unchanged. Under the irradiation ...

AI Technical Summary

Problems solved by technology

[0003] The above-mentioned methods based on high-resolution range images, target full polarization information, and the combination of the two are highly sensitive to noise

But in practice, due to the constant thermal noise of the receiver, the farther the target is from the radar, the lower the echo signal-to-noise ratio

In general applications, the range image data used in the training phase can be obtained by cooperative experiments or directly generated by electromagnetic simulation programs, and its signal-to-noise ratio is high; while the range image data used in the test phase is often obtained under non-cooperative conditions, High signal-to-noise ratio cannot be guaranteed, and the characteristics of the target data extraction will change

At this time, a model mismatch occurs between training and testing, resulting in a sharp deterioration in recognition performance as the signal-to-noise ratio of test samples decreases, resulting in a significant drop in recognition rate

Images