基于EMD的阵列观测声信号预处理算法研究
|
摘要
在声阵列信号处理中,为了抑制高频噪声的干扰,提高观测声信号的信噪比,提出基于EMD的阵列观测声信息预处理算法。该算法首先基于EMD理论将观测信号分解为多个IMF函数分量,利用FFT变换得到IMF函数分量的频谱特性;其次在频域上结合目标声信号的特性,确定IMF函数分量提取阈值;然后根据FFT逆变换对得到重构信号的IMF时域信号完成阵列观测信号的重构;最后通过静态半实物仿真实验,结合均值、标准差及相对剩余声能三个统计参数,验证了该文算法不仅具有较好的信号处理稳定性,而且在声能上保留了原始信号91%以上的能量,为进一步研究声目标识别及定位提供了声能保障。
In order to suppress the high frequency noise interference and improve signal-to-noise ratio of the observed acoustic signal in acoustic array signal processing,the preprocessing algorithm of array measurement signal based on EMD is proposed. The measurement signal is decomposed into multiple IMF function components based on EMD,and then the spectral characteristics of IMF is obtained by FFT. Combining the characteristics of the target acoustic signal,the extraction threshold value of IMF function component is determined. The IMF time-domain signal of reconstruction signal is obtained according to FFT inverse transform to complete the reconstruction of array observed signals. In combination with mean value,standard deviation and relative surplus acoustic energy,the static semi-physical simulation is adopted to validate that the proposed algorithm not only has better signal processing stability,but also reserves more than 91% energy of the original acoustic energy signal.The algorithm provides a sound energy safeguard for further study of acoustic target identification and positioning.
In order to suppress the high frequency noise interference and improve signal-to-noise ratio of the observed acoustic signal in acoustic array signal processing,the preprocessing algorithm of array measurement signal based on EMD is proposed. The measurement signal is decomposed into multiple IMF function components based on EMD,and then the spectral characteristics of IMF is obtained by FFT. Combining the characteristics of the target acoustic signal,the extraction threshold value of IMF function component is determined. The IMF time-domain signal of reconstruction signal is obtained according to FFT inverse transform to complete the reconstruction of array observed signals. In combination with mean value,standard deviation and relative surplus acoustic energy,the static semi-physical simulation is adopted to validate that the proposed algorithm not only has better signal processing stability,but also reserves more than 91% energy of the original acoustic energy signal.The algorithm provides a sound energy safeguard for further study of acoustic target identification and positioning.
引文
[1]ZEKAVAT S A.Power-azimuth-spectrum modeling for antennaarray systems:a geometric-based approach antennas and propa-gation[J].Journal of IEEE 2003,51(12):3292-3294.
[2]SANKARANAYAAN A C,ZHENG Qinfen,CHELLAPPA Ra-ma.Vehicle tracking using acoustic and video sensors,ADA431619[R].US:[s.n.],2004:2-8.
[3]FULGHUM T L.The Jakes fading model for antenna arrays in-corporating azimuth spread vehicular technology[J].Journal ofIEEE,2002,51(5):967-977.
[4]BIRD John S,MULLINS Geoff K.Analysis of swath bathyme-try sonar accuracy[J].IEEE journal of oceanic engineering,2005,30(2):372-390.
[5]DE JESUS Jzlviec Uudn,CALVO Juan Jnsk Villacorta,LZ-QUIERDO-PUCNTE Albcrto.Surveillance system based on da-ta fusion from image acoustic array sensors[J].IEEE AES sys-tems magazine,2000,15(2):9-16.
[6]刘亚雷.三维运动声阵列对声目标跟踪理论研究[D].南京:南京理工大学,2013.
[7]WANG D,WANG L.Global motion parameters estimation usinga fast and robust algorithm[J].IEEE transactions on circuitsand systems for vide technology,1997,7(5):823-826.
[8]PORAT Boaz,FRIEDLANDER Benjamin.A frequency domainalgorithm for multi-frame detection and estimation of dim tar-gets[J].IEEE transactions on pattern analysis and machine in-telliqence,1990,12(4):398-401.
[9]BARNETT J.Statistical analysis of median subtraction filteringwith application to point target detection in inrfared back-ground[J].Proceedings of SPIE,1989,1050:10-18.
[10]宋玉来,卢奂采,金江明.单层传声器阵列信号空间重采样的声波分离方法[J].物理学报,2014,63(19):187-196.
[11]吕艳新,顾晓辉.多传声器小波多尺度信息融合滤波算法[J].仪器仪表学报,2012,33(4):788-794.
[12]ZHONG Xionghu,HOPGOOD J R.Particle filtering forTDOA based acoustic source tracking:Nonconcurrent multi-ple talkers[J].Signal processing,2014,96:382-394.
[13]KIM Arild Steen,MCCLELLAN J H,GREEN Ole,et al.Acoustic source tracking in long baseline microphone arrays[J].Applied acoustics,2015(87):38-45.
[14]TIAN Ye,CHEN Zhe,YIN Fuliang.Distributed Kalman filter-based speaker tracking in microphone array networks[J].Ap-plied acoustics,2015,89:71-77.
[15]ZHONG Xionghu,PREMKUMAR A B.Multiple widebandsource detection and tracking using a distributed acoustic vec-tor sensor array:A random finite set approach[J].Signal pro-cessing,2014,94:583-594.
[2]SANKARANAYAAN A C,ZHENG Qinfen,CHELLAPPA Ra-ma.Vehicle tracking using acoustic and video sensors,ADA431619[R].US:[s.n.],2004:2-8.
[3]FULGHUM T L.The Jakes fading model for antenna arrays in-corporating azimuth spread vehicular technology[J].Journal ofIEEE,2002,51(5):967-977.
[4]BIRD John S,MULLINS Geoff K.Analysis of swath bathyme-try sonar accuracy[J].IEEE journal of oceanic engineering,2005,30(2):372-390.
[5]DE JESUS Jzlviec Uudn,CALVO Juan Jnsk Villacorta,LZ-QUIERDO-PUCNTE Albcrto.Surveillance system based on da-ta fusion from image acoustic array sensors[J].IEEE AES sys-tems magazine,2000,15(2):9-16.
[6]刘亚雷.三维运动声阵列对声目标跟踪理论研究[D].南京:南京理工大学,2013.
[7]WANG D,WANG L.Global motion parameters estimation usinga fast and robust algorithm[J].IEEE transactions on circuitsand systems for vide technology,1997,7(5):823-826.
[8]PORAT Boaz,FRIEDLANDER Benjamin.A frequency domainalgorithm for multi-frame detection and estimation of dim tar-gets[J].IEEE transactions on pattern analysis and machine in-telliqence,1990,12(4):398-401.
[9]BARNETT J.Statistical analysis of median subtraction filteringwith application to point target detection in inrfared back-ground[J].Proceedings of SPIE,1989,1050:10-18.
[10]宋玉来,卢奂采,金江明.单层传声器阵列信号空间重采样的声波分离方法[J].物理学报,2014,63(19):187-196.
[11]吕艳新,顾晓辉.多传声器小波多尺度信息融合滤波算法[J].仪器仪表学报,2012,33(4):788-794.
[12]ZHONG Xionghu,HOPGOOD J R.Particle filtering forTDOA based acoustic source tracking:Nonconcurrent multi-ple talkers[J].Signal processing,2014,96:382-394.
[13]KIM Arild Steen,MCCLELLAN J H,GREEN Ole,et al.Acoustic source tracking in long baseline microphone arrays[J].Applied acoustics,2015(87):38-45.
[14]TIAN Ye,CHEN Zhe,YIN Fuliang.Distributed Kalman filter-based speaker tracking in microphone array networks[J].Ap-plied acoustics,2015,89:71-77.
[15]ZHONG Xionghu,PREMKUMAR A B.Multiple widebandsource detection and tracking using a distributed acoustic vec-tor sensor array:A random finite set approach[J].Signal pro-cessing,2014,94:583-594.