Existing noise removal processes for airborne electromagnetic (AEM) data generally consist of several steps, with each using a specific method to remove a specific type of noise. To improve the efficiency of AEM denoising and reduce the impact of the subjective judgment of the operators on the processing results, we have adopted a deep learning method based on a denoising autoencoder (DAE), which enables in one single processing step the removal of multisource noise. The most common noise sources in AEM data, including motion-induced noise, nearby or moderately distant sferics noise, power-line noise, and background electromagnetic noise, will be combined with a large number of simulation responses to build a training set. The data in the training set will be used to train the deep learning DAE neural network so that the neural network could fully learn the respective characteristics of the signal and noise and further effectively distinguish the AEM response signal (useful signal) from the above noise. The field data were processed using this method, and the processing results were compared with those obtained using traditional methods. The comparison test revealed that this method is helpful to reduce the influence of subjective factors on the quality of data results and compress the entire AEM data processing time.

We have developed a scheme for decoupling the induced polarization (IP) effect from time-domain electromagnetic (TDEM) data. This scheme is achieved by simultaneously sampling the resistivity and pseudochargeability in a Bayesian framework. The TDEM and IP responses are simulated separately with the sampled model parameters and then are stacked to fit the IP-affected TDEM data. Thus, the influence of the IP phenomenon is eliminated in the process of recovering the resistivity. To reduce the computational cost brought by the Bayesian sampling, we use a 2D parametrization instead of sampling the full 3D space and we use a linear perturbation approximation for calculating the IP response. The linearized inversion results are used as the initial model, and a multiple proposed points algorithm is used to accelerate the sampling. We validate the proposed method with synthetic and field examples showing that it restores accurate estimates of electrical structures from the TDEM data that are significantly affected by the IP phenomenon. Our method could advance the application of the TDEM method to the scenario in which the IP may affect the TDEM data and mask the underlying geologic targets.