论文部分内容阅读
着重研究在复杂的城市背景噪声环境下,对入侵人员脚步信号的提取与检测.通过对脚步信号及城市背景噪声的频谱特征分析,提出一种基于震动信号的人员脚步信号提取方法,该方法通过小波包分解(WPD)结合局域判别基(LDB)划分信号特征频带,有效地分离了城市中的过往车辆对入侵人员脚步信号的干扰并做出准确提取,仿真与实验结果证明,与常用的脚步识别算法相比,该算法鲁棒性,普适性及抗干扰性更好,符合实际工程应用需求.
This paper focuses on the extraction and detection of footstep signals of intruders under the background of complex urban background noise.Based on the analysis of footstep signals and the spectral characteristics of urban background noise, this paper proposes a method of footstep signal extraction based on vibration signals, Wavelet Packet Decomposition (WPD) combined with Local Discriminant Base (LDB) divides the signal characteristic band to effectively separate the past vehicles in the city from interfering footstep signals and make accurate extraction. Simulation and experimental results show that, Compared with the footstep recognition algorithm, this algorithm has better robustness, universality and anti-interference ability, and meets the needs of practical engineering applications.