磁异常信号奇异值分解的随机共振检测方法

Stochastic Resonance Detection Method for Magnetic Anomaly Signals Singular Value Decomposition

针对目前对水下目标磁异常信号低信噪比检测能力较差的问题,提出奇异值分解随机共振(SVD-RS)方法。该方法基于磁偶极子仿真水下目标磁偶极子信号,信号经过奇异值分解选择最优子信号,随后将子信号通过随机共振系统,并构建检验统计量,实现对水下目标磁异常信号的检测。仿真结果表明,通过奇异值分解的方法,能够将信号信噪比提高6 dB,使用奇异值分解随机共振检测方法,能够在输入信噪比为-12 dB时,对水下目标磁异常信号进行有效检测。

To solve the problem of detecting the underwater target magnetic anomaly signal from the strong geomagnetic field noise,a singular value decomposition-single resonance(SVD-RS)method was proposed.Firstly,the input signal was established under different noise intensity by the magnetic dipole model,and the most sub signal was chosen by SVD method.SVD method had ideal de-correlation characteristics and could extract useful signals and characteristic information under strong noise.The results showed that this method could improve the signal to noise ratio by about 6 dB.Then the sub-signal was passed through the stochastic resonance system,when the test statistic was constructed to detect the magnetic abnormal signal of the underwater target.The results showed that the SVD-RS detection method could effectively detect the magnetic anomaly of underwater targets.