强干扰地区CSAMT数据信息熵与有理函数滤波的处理方法

A processing method of csamt data with strong electromagnetc interferences by using information entropy and rational fuction filtering

CSAMT在强干扰地区采集的数据,尽管在测量时采取了多种多样压制干扰的措施,但强干扰余音仍是影响CSAMT数据质量的主要因素.如何在数据处理时进一步削减干扰,是提高CSAMT反演数据质量与反演解释效果的重要一环.本文主要针对矿山强电磁干扰的特点,从美国Zonge公司的GDP仪器为选频测量数据出发,提出了一种基于信息熵去噪与有理函数滤波相结合的数据处理方法.首先采用CSAMT数据误差熵处理,逐次剔除大的干扰数据,直到得到满意的信息熵,从而提取实际的有用信号数据.然后对经过信息熵处理后的测深曲线,进行有理函数滤波处理.再次剔除干扰大的跳点,得到较圆滑的测线纯异常数据.通过数字模型验证,其方法正确,精度较平均处理结果高.经在强干扰矿山实测数据处理表明,该处理方法压制干扰效果明显,可以达到提高信噪比,改善实测数据质量的目的.

Though adapting many measures for suppressing interferences during field data collection, lingering of strong EM interferences remains a major factor affecting data quality. An important link for improving data inversion and interpretation is to further suppress interferences at the step of data pre-processing. With features of strong EM interferences in mines considered, a processing method is suggested in the paper, based on noise removing by using information entropy combined with rational function filtering, of CSAMT data collected with the GDP meter made by Zong Company, USA. Those CSAMT data having strong inferences should first be removed successively by using average information entropy until satisfied entropy is obtained and so with useful information achieved. The sounding curves based on the processed data are then filtered by using rational function, and with tears rejected. Finally, smooth curves, called ＂pure anomalies＂, are obtained in this way. The method mentioned above has been tested with digital models. The test results show that the accuracy of the method is higher than that of averaging. After performing processing of measured data collected in mines having strong EM interferences, it is believed that the suggested method is rather effective for improving S/N ratio and quality of measured data.