基于半监督学习的畸变雷达电磁信号快速识别研究

打开文本图片集

中图分类号：TB9；TP247.44 文献标志码：A文章编号：1674-5124（2025）07-0147-07

Abstract： Once the radar signal is distorted， itcancause errors in the target recognition process，leading toa decrease in target recognition accuracy. In this context，conducting research on fast identification of distorted radar electromagnetic signals based onsemi supervised learning is of great practical significance.This study utilizes the short-time Fourier transform algorithm to perform time-frequency conversion on radar electromagnetic signals，obtaining time-frequency images and implementing preprocessing. Extract four texture features of time-frequency images through grayscale co-occurrence matrix. Representing four texture features as samples， input them into a semi supervised support vector machine in semi supervised learning to achieverapid recognitionof distorted radar electromagnetic signals.The results indicate thatthe intersection to union ratio is higher and the time is shorter， indicating that the studied method can complete more accurate distortion identification at a faster speed， proving the performance of the studied method.

Keywords： semi supervised learning; distortion; radar electromagnetic signal; quick identification methods

0 引言

雷达作为一种重要的无线电设备，在军事、航空、气象等领域具有广泛的应用。（剩余10795字）