The time-domain electromagnetic (TDEM) method is acknowledged for its simplicity in setup and non-intrusive detection capabilities, particularly within shallow subsurface detection methodologies. However, extant TDEM systems encounter constraints when detecting intricate topographies and hazardous zones. The rapid evolution in unmanned aerial vehicle (UAV) technology has engendered the inception of UAV-based time-domain electromagnetic systems, thereby augmenting detection efficiency while mitigating potential risks associated with human casualties. This study introduces the UAV-TDEM system designed explicitly for discerning shallow subsurface targets. The system comprises a UAV platform, a host system, and sensors that capture the electromagnetic response of the area while concurrently recording real-time positional data. This study also proposes a processing technique rooted in robust local mean decomposition (RLMD) and approximate entropy (ApEn) methodology to address noise within the original data. Initially, the RLMD decomposes the original data to extract residuals alongside multiple product functions (PFs). Subsequently, the residual is combined with various PFs to yield several cumulative sums, wherein the approximate entropy of these cumulative sums is computed, and the resulting output signals are filtered using a predetermined threshold. Ultimately, the YOLOv8 (You Only Look Once version 8) network is employed to extract anomalous regions. The proposed denoising method can process data within one second, and the trained YOLOv8 network achieves an accuracy rate of 99.0% in the test set. Empirical validation through multiple flight tests substantiates the efficiency of UAV-TDEM in detecting targets situated up to 1 m below the surface. Both simulated and measured data corroborate the proposed workflow’s effectiveness in mitigating noise and identifying targets.

The development of deep learning greatly promotes the progress of speaker verification (SV). Studies show that both convolutional neural networks (CNNs) and dilated time-delay neural networks (TDNNs) achieve advanced performance in text-independent SV, due to their ability to sufficiently extract the local feature and the temporal contextual information, respectively. Also, the combination of the above two has achieved better results. However, we found a serious gridding effect when we apply the 1D-Res2Net-based dilated TDNN proposed in ECAPA-TDNN for SV, which indicates discontinuity and local information losses of frame-level features. To achieve high-resolution process for speaker embedding, we improve the CNN–TDNN structure with proposed repeated multi-scale feature fusions. Through the proposed structure, we can effectively improve the channel utilization of TDNN and achieve higher performance under the same TDNN channel. And, unlike previous studies that have all converted CNN features to TDNN features directly, we also studied the latent space transformation between CNN and TDNN to achieve efficient conversion. Our best method obtains 0.72 EER and 0.0672 MinDCF on VoxCeleb-O test set, and the proposed method performs better in cross-domain SV without additional parameters and computational complexity.

Frequency-domain electromagnetic (FDEM) measurements for shallow subsurface exploration are frequently corrupted by noise, which masks weak secondary-field responses and degrades interpretation. We propose an electromagnetic diffusion CNN (EMDiC) for 1D multi-frequency FDEM denoising, where denoising is formulated as conditional diffusion-based generation. EMDiC combines an analytic frequency–spatial encoder, a Feature-wise Linear Modulation (FiLM)-conditioned convolutional hourglass backbone, and a physics-informed composite loss built on velocity loss to improve waveform reconstruction under severe noise. A reproducible synthetic dataset is constructed through layered-earth forward modeling with concentric Transmitter–Receiver (TX–RX) geometry, multiple target categories, and mixed noise waveforms. On synthetic benchmarks covering multiple noise levels and material types, EMDiC achieves the best overall performance in Root Mean Square Error (RMSE), Signal-to-Noise Ratio (SNR), and Normalized cross-correlation (NCC) among 1D U-Net, diffusion-based variants, and representative neural baselines, with the clearest gains under medium-to-strong noise and for targets with pronounced induction responses. Ablation experiments verify the complementary contributions of electromagnetic positional encoding (EMPE), FiLM conditioning, and the composite loss. Field data validation with a self-developed GEM-3 system further shows that EMDiC improves cross-frequency coherence and suppresses oscillations while preserving the main response characteristics.

Recent advances in speech restoration and bandwidth extension have significantly improved audio quality under controlled conditions. However, their performance often degrades drastically in real-world scenarios where severe distortions and bandwidth limitations coexist. These challenges become particularly critical under extreme conditions, such as low signal-to-noise ratios and heavily degraded spectral features, where conventional methods struggle to recover intelligible speech. Compound distortions severely disrupt the time-frequency structure of speech signals, undermining the robustness and perceptual quality of existing enhancement systems. Although generative approaches, such as generative adversarial networks, excel in producing perceptually plausible outputs, they remain highly sensitive to input instability, limiting their reliability in such conditions. To address these challenges, we propose a novel two-stage framework that leverages a robust pre-training strategy to jointly restore multiple speech distortions for quality improvement. The method integrates a Scale-Shift Linear Transformation, which effectively suppresses non-stationary noise while preserving essential speech components. Additionally, a GAN-based fine-tuning strategy is introduced to enhance perceptual realism and reconstruct coherent high-frequency details. Experimental results demonstrate that our model, trained solely on the VCTK corpus, achieves high-fidelity restoration across various distortion scenarios. Its strong generalization ability under unseen and challenging conditions highlights the practicality and robustness of the proposed method for real-world speech restoration applications.