基于模板的瞬时RFI特征识别算法初探  

Preliminary Study of Transient RFI Feature Recognition Algorithm Based on Template

作　　者：张亚州 张海龙[1,2,3,4] 张萌 王杰[1,4] 冶鑫晨 王万琼[1] 李嘉 杜旭 Zhang Yazhou;Zhang Hailong;Zhang Meng;Wang Jie;Ye Xinchen;Wang Wanqiong;Li Jia;Du Xu(Xinjiang Astronomical Observatory,Chinese Academy of Sciences,Urumqi 830011,China;University of Chinese Academy of Sciences,Beijing 100049,China;Key Laboratory of Radio Astronomy,Chinese Academy of Sciences,Nanjing 210033,China;National Astronomical Data Center,Beijing 100101,China)

机构地区：[1]中国科学院新疆天文台,新疆乌鲁木齐830011 [2]中国科学院大学,北京100049 [3]中国科学院射电天文重点实验室,江苏南京210033 [4]国家天文科学数据中心,北京100101

出　　处：《天文研究与技术》2022年第5期479-486,共8页Astronomical Research & Technology

基　　金：国家自然科学基金(11873082,11803080,12003062);国家重点研发计划(2018YFA0404704);中国科学院青年创新促进会;国家天文科学数据中心;中国科学院天文台站设备更新及重大仪器设备运行专项资助。

摘　　要：瞬时射频干扰(Radio Frequency Interference,RFI)信号是最难以识别并消除的一类干扰。针对射电天文观测过程中遇到瞬时射频干扰问题,提出一种基于射频干扰信号峰值的特征提取方法,利用该方法生成射频干扰信号的特征模板,并基于动态时间规整(Dynamic Time Warping,DTW)算法和打分策略设计特征识别算法,实现了射频干扰信号识别以及分类功能,利用ARRL官网(http://www.arrl.org/sounds-of-rfi)提供的射频干扰数据完成了特征模板的交叉验证对比。测试结果显示,射频干扰特征模板与同类别射频干扰相似度高,表明该算法能够对射频干扰进行识别和分类。该算法为瞬时射频干扰的特征提取以及识别提供了一种新思路。Transient radio frequency interference(RFI) signals are one of the most difficult types of interference to identify and eliminate. This paper proposes a feature extraction method based on RFI signal peaks for the problem of transient RFI encountered during radio astronomical observations. The method is used to generate a feature template for RFI signals, and a feature recognition algorithm is designed based on the Dynamic Time Warping(DTW) algorithm and scoring strategy to achieve RFI signal recognition and classification. The cross-validation comparison of the feature templates is completed using the RFI data provided by the ARRL website. The test results show that the RFI feature templates are highly similar to the same category of RFI, indicating that the algorithm is capable of identifying and classifying RFI, which provides a new idea for the feature extraction and identification of transient RFI.

关 键 词：射频干扰 周期计算 特征提取 特征识别 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]