Denoising algorithm of ground-airborne time-domain electromagnetic method based on Variational Bayesian-based adaptive Kalman filter (VBAKF)

Abstract

The ground-airborne time-domain electromagnetic (GATEM) method, as a new type of electromagnetic method, offers detection efficiency and large investigation depth. It is an efficient electromagnetic detection technique for geological exploration and groundwater exploration. The field data are disturbed by motion noise and mixed noises, such as white noise and electromagnetic noises, which will reduce the data quality and accuracy of data interpretation. The characteristics of GATEM responses are analysed, and a denoising algorithm based on variational Bayesian-based adaptive Kalman filter (VBAKF) is proposed. Considering the attenuation characteristics of the GATEM response, the exponentially fitted data input into the VBAKF. Then the VBAKF with inverse Wishart priors is used to denoise GATEM field data, and the unknown process noise covariance matrix (PNCM) and measurement noise covariance matrix (MNCM) are estimated by variational Bayesian. The VBAKF algorithm is tested on the responses with different noises. It is verified on a layered geological model and compared with the adaptive Kalman filter (AKF) algorithm and the combined wavelet filter (CWF) algorithm. It is also tested on synthetic data of an anomalous model. Through synthetic data, the VBAKF algorithm can effectively suppress noise of GATEM data. Finally, the algorithm is applied to field data in Leshan, Jilin province, China. It effectively improves the quality of field data.