摘要
针对海洋环境噪声的特性非常复杂,而水下噪声监测的样本往往较少且不全面等问题,设计了适用于样本数量较少的海洋机械噪声检测系统。该系统基于MFCC特征参数提取将接收到的声音信号进行分类建立特征库,并用支持向量机的机器学习算法对陌生信号分类识别。仿真及试验数据分析的结果显示,在海洋环境噪声背景下该系统能较好地监测出机械噪声的存在。
In view of the complex characteristics of marine environmental noise,and the samples of underwater noise monitoring are often small and incomplete,a marine mechanical noise detection system suitable for small number of samples is designed.Based on MFCC feature parameter extraction,the received sound signals are classified and a feature library is established,and the unfamiliar signals are classified and recognized by SVM machine learning algorithm.The results of simulation and experimental data analysis show that the system can detect the existence of mechanical noise under the background of marine environmental noise.
作者
王利恒
赵智浩
WANG Li-heng;ZHAO Zhi-hao(School of Electrical and Information Engineering,Wuhan Institute of Technology,Wuhan 430205,China)
出处
《自动化与仪表》
2020年第12期54-58,共5页
Automation & Instrumentation
基金
武汉工程大学创新基金项目(CX2019157)。
关键词
海洋噪声监测
梅尔频率倒谱系数
支持向量机
marine noise monitoring
Mel frequency cepstrum coefficient(MFCC)
support vector machines(SVM)
