摘要
针对基于震动信号对地面目标识别虚警率高、识别概率低的问题,本文根据实测数据提取了信号的时域和频域特征作为神经网络的输入,提出了使用基于BP神经网络的震动信号目标识别算法。实验结果表明,该算法可有效地区分人和车辆目标,降低虚警率,对使用震动传感器进行监测具有重要意义。
Arming at the problem of high alarm rate and low recognition rate in vibration signal processing, the method based on BP neural network is proposed, using the time domain feature and the frequency domain feature as inputs of the network. The experimental results indicate that the BP neural network can distinguish the signal of car and human being effectively with low alarm rate, which is significant in monitoring based on vibration sensors.
引文
[1]刘保柱,苏彦华,张宏林.MATLAB 7.0从入门到精通(修订版)[M].北京:人民邮电出版社,2010.
[2]陈明,等.MATLAB神经网络原理与实例精解(附光盘)[M].北京:清华大学出版社,2014.
[3]毛志勇.市政道路路基质量控制检测方法的研究[D].山东大学硕士学位论文,2009.
[4]曲华,等.利用瑞利波频谱分析法无损检测混凝土结构[J],混凝土,2006(2).