Rejectable Radar HRRP Target Recognition Method Based on CNN

Abstract

A CNN-based refusal radar HRRP target recognition method, the implementation steps are: (1) Obtain HRRP time-frequency domain characteristic data of radar high-resolution range image; (2) Select training sample set and test sample set; (3) Construct the convolutional neural network; (4) set the adjustable cost function of the convolutional neural network; (5) train the convolutional neural network; (6) obtain the output of the convolutional neural network; (7) judge whether the reconstruction error is greater than If it is the threshold value, the object will be rejected; otherwise, the recognition result will be obtained. The invention introduces a multi-layer convolutional neural network to extract high-dimensional features in the time-frequency domain features of radar HRRP data, which can effectively solve the problem of low target recognition accuracy caused by limited target feature information in the prior art, and at the same time The external target has an adjustable rejection ability, and it has better target recognition performance than the general method.

Description

technical field [0001] The present invention belongs to the technical field of radar, and further relates to a method for recognizing radar high-resolution range profile HRRP (High-Resolution Range Profile) targets based on convolutional neural network CNN (Convolutional Neural Network) in the technical field of radar target recognition . The invention can reject and judge the target outside the radar database for the radar high-resolution range image data, and is used for the subsequent target recognition of the target in the database. Background technique [0002] Radar high-resolution range profile (HRRP) contains rich radar target structure features, and has the advantages of easy acquisition, storage and processing. It is very valuable for radar target recognition and classification, and has become a research hotspot in the field of radar automatic target recognition. Convolutional neural network (CNN) is a deep learning method, which avoids the complex feature extract...

AI Technical Summary

Problems solved by technology

The disadvantage of this method is that only a shallow neural network is used to analyze the time-domain information of radar HRRP data, and the method involved cannot fully utilize the time-frequency domain information of the data to extract high-dimensional features, which limits the use of The amount of feature information for target recognition

The disadvantage of this method is that it does not make full use of the extracted high-dimensional features to realize the rejection before target recognition, and it lacks the ability to reject when there are abnormal targets outside the library in the radar high-resolution range image sample. Effective rejection performance will reduce the accuracy of target recognition

Images