基于模糊神经网络的播音信号识别技术研究

何艳萍

摘  要： 播音信号识别对于提高播音质量具有十分重要的意义，当前播音信号识别方法存在误识率高，播音信号识别效率低等不足，为了获得更优的播音信号识别结果，设计基于模糊神经网络的播音信号识别技术。首先分析当前播音信号识别技术的研究进展，指出各种播音信号识别技术存在的不足;然后采用空间变换和奇异值分解算法提取播音信号识别特征，并采用模糊神经网络建立播音信号识别分类器;最后在Matlab 2017平台上进行播音信号识别仿真实验，结果表明，模糊神经网络获得了理想的播音信号识别，播音信号识别正确率要高于当前其他播音信号识别技术，减少了播音信号的误识率，缩短了播音信号识别的时间，提升了播音信号识别速度，具有较高的实际应用价值。

关键词： 播音信号; 信号识别; 模糊神经网络; 奇异值分解算法; 识别特征; 识别速度

中图分类号： TN911.73?34; TP391                    文献标识码： A                 文章编号： 1004?373X（2020）03?0054?04

Research on broadcast signal recognition technology based on fuzzy neural network

HE Yanping

（School of Literature and Journalism， Baoji University of Arts and Sciences， Baoji 721013， China）

Abstract： Broadcasting signal recognition is of great significance to improve the quality of broadcasting. At present， there are some shortcomings in the methods of broadcasting signal recognition， such as high misrecognition rate and low recognition efficiency. In order to obtain better results of broadcasting signal recognition， a broadcasting signal recognition technology based on fuzzy neural network is designed. Firstly， the research progress of the current broadcasting signal recognition technologies is analyzed， and the shortcomings of various broadcasting signal recognition technologies are pointed out. Then， the spatial transformation and singular value decomposition algorithm are used to extract the recognition features of broadcasting signals， and the fuzzy neural network is used to establish the recognition classifier of broadcasting signals. Finally， the simulation experiment of the broadcasting signals is performed on the platform of Matlab 2017. The simulation results show that the broadcast signal recognition technology based on fuzzy neural network achieves ideal recognition of broadcasting signals， and its correct recognition rate of broadcasting signals is higher than that of other current recognition technologies. It reduces the error recognition rate of broadcasting signals， shortens the recognition time of broadcasting signals， and improves the recognition speed of broadcasting signals， which proves that it has a high practical application value.

Keywords： broadcast signal; signal recognition; fuzzy neural network; singular value decomposition algorithm; recognition feature; recognition speed

0  引  言

隨着我国经济的不断发展，人们收听播音节目的手段和方式越来越多，如手机、电视、电脑等，它们对播音信号的质量要求越来越高，因此播音信号识别技术的研究引起了人们的高度重视和关注[1?3]。

多年來，许多学者对播音信号识别问题进行了深入的研究，提出了许多播音信号识别技术[4?7]。传统方法主要包括：改进子空间匹配追踪导波的信号识别技术、基于同源性检验的信号识别技术、基于对信号功率谱拟合的信号识别方法等[8?10]，它们有自己的应用领域，但是它们具有比较明显的缺陷，如通用性差、不易实现，误识率比较高。……