摘要
“黑广播”即非法调频广播。目前,各级各地行政执法部门不断向非法设置使用调频广播“亮剑”,取得了良好的效果。但是,在非法调频广播的判断上以人工判断为主,在机器学习上有所探讨,且主要以语音转文字后的机器学习为主,在利用调频广播的信号特征上进行机器判断的研究较少。因此,引入贝叶斯判别方法,以构建的6个信号特征为维度,分析构建的合法音乐广播、合法新闻广播以及非法调频广播,并用某一非法调频广播进行验证判断,获得了较高的准确率。
The illegal FM broadcasting has been notorious for a long time,and the administrative law enforcement departments at all levels have been using FM broadcasting illegally,which has achieved good results.However,the judgment of illegal FM broadcasting is mainly based on manual judgment and machine learning,which is mainly based on machine learning after voice to text,and less on machine judgment by using the signal characteristics of FM broadcasting.In this paper,Bayesian discriminant method is introduced to analyze the legal music broadcast,legal news broadcast and illegal FM broadcast based on the six signal features.Finally,a certain illegal FM broadcast is used to verify the judgment,and a high accuracy is achieved.
作者
林秋海
龙华
LIN Qiuhai;LONG Hua(Kunming University of Science and Technology,Kunming Yunnan 650100,China)
出处
《通信技术》
2021年第3期654-657,共4页
Communications Technology
基金
国家自然科学基金(No.61761025)。
关键词
非法调频广播
贝叶斯判别
信号特征
信号识别
illegal FM broadcasting
Bayesian discrimination
signal characteristics
signal recognition
