Radar target RCS identification method and device based on complex neural network

Abstract

The invention relates to a radar target RCS identification method and device based on a complex neural network. The complex RCS data of a target is identified and classified through a complex neural network, complex convolution is carried out on the complex RCS data through a complex convolution layer, then a Complex Relu activation function is used for activating the complex RCS data, and then maximum pooling operation is carried out; a first real number convolution layer performs real number convolution on the complex features, then activates the complex features by using a Relu activation function, and then performs maximum pooling operation; and a second real number convolution layer performs real number convolution on the maximum real number feature, activates the maximum real number feature by using a Relu activation function, then performs average pooling operation, and finally outputs a classification feature vector. According to the method, deep learning recognition of the complex RCS data with richer target feature information is realized, and the target recognition accuracy can be effectively improved.

Description

technical field [0001] The invention relates to the technical field of radar target recognition, in particular to a radar target RCS recognition method and device based on a complex neural network. Background technique [0002] Since World War II radar was used to detect military air targets, military radar with all-weather and long-distance detection functions has been a research hotspot for researchers from various countries for decades as the "clairvoyance" of the battlefield. [0003] As the main feature of radar detection, target RCS (radar cross section) is often used for military detection, tracking and identification. The original RCS data is complex data. Compared with the single real-part amplitude RCS information, the complex-number RCS has richer representation capabilities. Not only the real-part amplitude of the complex-number RCS can express a lot of target size information, but its imaginary part phase also contains rich If all the target distance informatio...

AI Technical Summary

Problems solved by technology

[0004] The technical problem to be solved by the present invention is to solve the problem that the neural network cannot be used to identify and apply complex RCS data in the past, aiming at the defects in the prior art, a radar target RCS identification method and device based on the complex neural network are provided

Images