Aimed at the problem that magnetotelluric signals are frequently influenced by strong noise interference during acquisition, a new magnteotelluric sounding data strong interference separation method based on mathematical morphology filtering was proposed. Firstly, the noise reduction capability was checked through simulated signal with the common strong interference and then the structure elements size selection program was analysed according to different noise types. Finally, the morphological filtering method was applied to the measured magnetotelluric signals noise reduction process. Experiments using nonlinear conjugate gradient method for inversion were conducted to check the improvement of magnetotelluric signals quality. The results indicate that the proposed method can effectively eliminate large scale disturbance and baseline drift of magnetotelluric signals. The method can better keep the original features of magnetotelluric signals, and the quality of magnteotelluric sounding data is improved. Because the principle of the method is simple and the parallel computing speed is fast, it has good application value and is suitable for the strong interference separation in ore district.%针对矿集区大地电磁信号采集过程中常引入强噪声干扰等问题,采用数学形态滤波对大地电磁强干扰分离方法进行研究.在仿真信号中加入常见的强干扰来检验形态滤波的降噪能力,根据噪声类型选取不同结构元素尺寸及大小,并将形态滤波应用于实测大地电磁数据的降噪处理.采用非线性共轭梯度法进行反演,考查形态滤波对提高大地电磁测量数据质量的改善情况.研究结果表明:数学形态滤波能有效消除大地电磁强干扰中的大尺度干扰和基线漂移现象,重构信号基本保留原始大地电磁信号特征,改善大地电磁测深数据质量.由于该方法原理简单、并行运算速度快,具有较好的应用价值,适合于矿集区海量大地电磁强干扰分离.
展开▼
