非高斯背景噪声下的微弱磁异常信号检测算法

摘要

As for the poorer effect of traditional OBF decomposition algorithm in non-Gaussian noise, a new detection algorithm was proposed using a band-pass filter combined with OBF decomposition. Firstly the Parks-McClellan optimal FIR filter was designed based on the frequency characteristic of magnetic anomaly signal. The raw signal was filtered by the band-pass filter so that non-Gaussian noise could fall into Gaussian distribution approximately and the information of magnetic anomaly sig nal was retained to the largest extent. Then the filtered signal was decomposed by OBF to pick up characteristic energy signal, which would be tested with threshold. The new algorithm was verified by the simulation data and actual measured data. The results show that the improved algorithm will in crease SNR and the detection ability of weak magnetic anomaly signal in non-Gaussian noise, which introduces the traditional OBF decomposition algorithm in non-Gaussian noise.%针对传统的OBF分解算法在非高斯噪声下检测性能较差的问题,提出了一种带通滤波结合OBF分解的磁异常信号检测算法.首先,根据磁异常信号的频域特征,设计了Parks-McClellan最优FIR滤波器.通过对舍噪信号的带通滤波,实现对非高斯噪声的近似高斯化,同时最大程度地保留磁异常信号的信息.然后,对滤波后的信号进行OBF分解,提取能量特征信号进行门限检测.最后,分别采用仿真数据和实测数据对算法进行了验证.结果表明:改进后的算法有效地提高了信噪比,增强了非高斯噪声下微弱磁异常信号的检测能力,解决了OBF在非高斯下检测性能差的问题.