Sensors

The waveform stacking location method achieves microseismic source localization by computing characteristic functions (CFs) and stacking multi-channel data, without phase picking. It has been widely applied in geotechnical engineering. However, the low signal-to-noise ratio (SNR) caused by weak event energy and ambient noise often degrades localization accuracy. To enhance the localization precision and stability under low SNR conditions, this study employs the Stockwell transform (S-transform) to convert noisy time-domain data into the time–frequency domain. By analyzing the energy distribution of microseismic signal and noise in the time–frequency domain, frequency and time coefficients are introduced to enhance the energy of microseismic signal. Event location is achieved through the computation of CFs and multiple-cross-correlation stacking. Comparison of the location results when computing the CFs by the new method, the short-term average to long-term average ratio (STA/LTA) method, and the envelope (Env) method under varying noise levels demonstrates the superior noise resistance and improved localization accuracy of the new method. Finally, the effectiveness of the new method is validated using real seismic data collected from a coal mine.

Settlement and deformation of multi-purpose utility tunnels (MUTs) are critical factors affecting their structural integrity and service life; however, effective identification methods remain limited. This study proposes a comprehensive approach integrating Brillouin frequency domain analysis (BOFDA), fiber Bragg grating (FBG), and ground penetrating radar (GPR) technologies, which was successfully applied to an MUT comprising three tanks in Baiyin City, Gansu Province, China. BOFDA enables precise localization of settlement points, FBG-based dislocation meters facilitate posture recognition of the MUT, and GPR is employed for detailed analysis of settlement causes. The results indicate that MUT deformation primarily manifests as displacement at joint locations, supplemented by deformation of the tunnel structure itself. Rotation, even settlement, and uneven settlement were identified through three FBG-based dislocation meters installed on the top and side walls. The primary causes of MUT settlement include mudstone compression and collapse of loess.

Accurate segmentation of navigable waters and obstacles is critical for unmanned surface vessel navigation yet remains challenging in real aquatic environments characterized by complex water textures and blurred boundaries. Current models often struggle to simultaneously capture long-range contextual dependencies and fine spatial details, frequently leading to fragmented segmentation results. In order to resolve these issues, we present a novel segmentation model based on the CoAtNet architecture. Our framework employs an enhanced convolutional attention encoder, where a Fused Mobile Inverted Bottleneck Convolution (Fused-MBConv) module refines boundary features while a Convolutional Block Attention Module (CBAM) enhances feature awareness. The model incorporates a Bi-level Former (BiFormer) to enable collaborative modeling of global and local features, complemented by a Multi-scale Attention Aggregation (MSAA) module that effectively captures contextual information across different scales. The decoder, based on U-Net, restores spatial resolution gradually through skip connections and upsampling. In our experiments, the model achieves 95.15% mIoU on a self-collected dataset and 98.48% on the public MaSTr1325 dataset, outperforming DeepLabV3+, SeaFormer, and WaSRNet. These results show the model’s ability to effectively interpret complex aquatic environments for autonomous navigation.