基于半監(jiān)督學(xué)習(xí)的畸變雷達(dá)電磁信號快速識別研究

打開文本圖片集

中圖分類號：TB9；TP247.44 文獻(xiàn)標(biāo)志碼：A文章編號：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

雷達(dá)作為一種重要的無線電設(shè)備，在軍事、航空、氣象等領(lǐng)域具有廣泛的應(yīng)用。（剩余10795字）