Electromagnetic target classification method based on knowledge vector embedding

Abstract

The invention discloses an electromagnetic target classification method based on knowledge vector embedding, and the method comprises the steps: building a graph structure of an electromagnetic target through the data of known electromagnetic target information, and carrying out the embedded vector representation of a graph node corresponding to each electromagnetic target type based on a graph neural network; the method comprises the following steps: acquiring an electromagnetic target signal, performing short-time Fourier transform on electromagnetic target data to obtain time-frequency data of the electromagnetic target data, and preprocessing the time-frequency data to serve as a sample for training a convolutional neural network; constructing a convolutional neural network, training the convolutional neural network based on a result represented by the embedded vector of the graph node corresponding to the electromagnetic target category, and finally obtaining a reference vector for subsequently classifying and identifying the acquired electromagnetic target signal; and the acquired electromagnetic target signal is classified and identified by using the obtained reference vector. The method is high in applicability, the category relationship knowledge is combined into network training, and the defect that a traditional classification network is only suitable for recognizing categories appearing in a training set is overcome.

Description

technical field [0001] The invention relates to the fields of communication and artificial intelligence, and in particular to an electromagnetic target classification method using knowledge vector embedding. Background technique [0002] In the knowledge graph, the knowledge of the graph structure mainly includes node knowledge and relational knowledge. A node can represent a known category, and the relationship is the degree of connection between each category, which can intuitively reflect the closeness of the relationship between nodes. After embedding the graph structure into a vector form, the vector representation of each node can be Reflect knowledge of relationships between nodes. In recent years, due to the development of the graph neural network, it can effectively use the attributes and mutual relationships of the nodes in the graph structure, and rely on their mutual relationships for modeling, and embed the nodes as vectors to provide a computable data format f...

AI Technical Summary

Problems solved by technology

The present invention converts the hard decision of the traditional classification neural network represented by the convolutional neural network into a soft decision for calculating the similarity between vectors, thereby solving the hard decision when the current sample embedding results cannot reflect the relationship between the categories to which the sample belongs and the classification decision. issue of judgment

Images