The transient electromagnetic method (TEM) is one of the geophysical methods that can quickly detect underground space information. It obtains the resistivity structure of underground medium by inversion technology. The traditional inversion methods can only give a unique solution conforming to TEM data, but cannot solve the multiplicity problem of solutions. TEM Bayesian inversion technology can provide uncertainty information to solve this problem. However, its long calculation time makes it difficult to apply to engineering detection with high real-time performance. To solve these problems, a TEM one-dimensional inversion mixture density network (TEMIMDNet) is proposed in this paper. This method combines Bayesian theory with deep learning methods. By inputting TEM data into the trained network, the statistical parameters of the posterior probability density of the corresponding geological model can be quickly obtained. Thus, the uncertainty information of the geological model can be obtained. This method overcomes the problem of low efficiency of traditional Bayesian inversion. The simulated experiment shows that the TEMIMDNet method can not only obtain the probability density function graph of the geological model but also the average relative error between the maximum a posteriori model and the corresponding TEM response, which is less than 0.02. The field experiment shows that the TEMIMDNet method can output the statistical parameters of 52 survey points in only 4 ms, and the imaging results are consistent with the spatially constrained inversion method and OCCAM.

The magnetotelluric (MT) inversion technology is crucial for quantitatively interpreting deep mineral resources, especially when combined with rock physics information, enhancing accuracy in assessing underground structural parameters and spatial distribution. However, the traditional fuzzy c-means (FCMs) clustering-constrained inversion method requires prior rock physics information for each geological unit, limiting their application scope. We propose a guided FCMs (GFCMs) clustering-constrained inversion method based on adaptive virtual rock physics information, referred to as XG-FCM-constrained MT inversion. This method breaks free from the constraints of traditional methods by not relying on prior rock physics information. In terms of extracting virtual rock physics information, we employ a local density clustering algorithm to dynamically extract resistivity model information from MT inversion iterations, automatically determining the number of clusters and cluster centers. Regarding the inversion strategy, we construct an integrated objective function that combines data fitting, smoothing constraint, and GFCM constraint, implementing a two-stage iterative solution strategy of “smoothing first, clustering second.” Model testing demonstrates that compared to traditional smoothing-constrained MT inversion, the XG-FCM-constrained method achieves a significant improvement in the resolution of resistivity model reconstruction, clearly delineating the boundaries of underground anomalies. Even in situations where rock physics information is insufficient or absent, this method can effectively reconstruct high-quality underground resistivity models, reducing the dependence on complete prior information. The application of actual field data further highlights the advantages of the XG-FCM-constrained MT inversion method, providing robust support for accurately delineating geological unit boundaries and precisely identifying potential ore deposit target areas.

Surface nuclear magnetic resonance (SNMR) is the only non-invasive geophysical method for direct water detection, and is widely used in groundwater exploration. However, the dead time arising from the coupling effect of the transmitting magnetic field on the receiving coil results in the loss of the high level early free induction decay (FID) signal. To solve this problem, we propose a decoupling method based on symmetrical transmitting coils. A mathematical analysis was conducted to describe the decoupling principle and detection theory based on the novel coil structure. Simulation experiments that performed comparisons with the traditional method showed the potential of the new method for the detection of shallow thin aquifers with short relaxation time signals. Further studies of the detection effects with different coil key structure parameters were conducted and a three-layer coil was developed. Laboratory tests showed that the dead time was shortened to 1 ms, which made it possible to acquire the FID signal earlier.

Transient electromagnetic (TEM) method is a geophysical technique suitable for efficient detection of urban underground space, which can be used for advance detection of road collapse and underground water inrush accidents. In actual urban geological detection, engineers hope to be able to quickly characterize the underground space in 3-D, so as to assess the potential underground subsidence area or water inrush area risk. However, due to the complexity of urban geology, it is difficult for conventional 3-D imaging techniques to take into account both imaging efficiency and imaging accuracy. To address this issue, this article proposes a fast, high-resolution TEM quasi-3-D imaging strategy suitable for urban geology. We convert the TEM data into pseudoseismic wavelet data to identify the geological correlation of the 3-D spatial survey area, and then impose adaptive spatial constraints on the 3-D imaging. Simulation and application case results show that, compared with conventional imaging techniques, our proposed new strategy can effectively improve imaging resolution while ensuring computational efficiency. The research results provide a feasible technical solution for rapid high-resolution 3-D detection of urban geology.

There are many conventional methods that have been applied in transient electromagnetic (TEM) random noise suppression such as stacking-averaging. But, when the TEM system works in urban areas with strong noise, these methods are not effective due to the extremely low signal-to-noise ratio (SNR). We propose a new method combining the minimum noise fraction (MNF) algorithm and deep learning. The MNF and the deep neural network (DNN) are used to extract the complex features of signals from the noisy signal data. After using MNF to improve the SNR of TEM to a certain extent, the convolutional neural network (CNN) and gated recurrent unit (GRU) were used to extract spatial and temporal features of the signal, and the training was guided by the double loss function. To verify the effectiveness of the method, we have done quantitative experiments on synthetic noise and field noise respectively. The experimental results show that our method achieves the most advanced performance.

Machine learning inversion (ML-inversion) has been widely used in the interpretation of geophysical data. However, because the transient electromagnetic (TEM) is not sensitive to the response of the small-size or large-depth abnormal body, there are some problems such as poor imaging accuracy and low-resolution when directly processing raw TEM data with ML method. To solve this problem, we propose to use the pseudo seismic wavelet (PSW) data obtained from TEM wave field transformation to perform TEM ML-inversion. Simulation and example application verify that compared with the TEM data, the PSW data has higher recognition sensitivity to the electrical changes of underground anomalies, and the ML-inversion based on the PSW data can significantly improve the TEM imaging accuracy. Our study demonstrates that the PSW data can be used to achieve TEM ML-inversion, which may provide a new way for the further combination of electromagnetic data processing and ML technology.

Road collapse and underground water inrush accidents pose a serious threat to urban safety and development. Transient electromagnetic (TEM) method is an effective geophysical method for detecting urban underground space. However, due to the complexity of urban geological environment, the conventional TEM detection methods are difficult to meet the needs of efficient and high-resolution detection of urban underground space. To solve this problem, an adaptive high-resolution (AHR) TEM detection method based on pseudoseismic wavelet transform technology is proposed in this study. The simulation results show that compared with the conventional technologies, the AHR detection method can more accurately reflect the resistivity distribution information in complex geological environment. The proposed method is implemented to detect an urban underground cavity, and the imaging results are consistent with the actual results. The research results will provide technical support for the early warning of road collapse and water inrush in urban industrial development.