Abnormal changes in groundwater levels are one of the key indicators of seismic precursors. Before an earthquake occurs, the groundwater level often shows varying degrees of abnormality. These anomalies typically manifest as a sudden rise or fall in groundwater levels and will last for a period of time. On Jul. 12, 2020, at 6:38 AM, a 5.1 magnitude earthquake occurred in Guye District, Tangshan City, Hebei Province, China (39.78°N, 118.44°E). This paper uses this earthquake as a case study to analyze the groundwater level data from two observation wells, Zhaogezhuang well in and Yutian Ji 03 well. To accurately identify seismic precursor anomalies, the groundwater level data is divided into seismically Active (SA) and seismically inactive (non-SA) periods, forming the basis for dataset segmentation. This paper proposes a TCN-LSTM-Attention model that combines the advantages of effective feature extraction with TCN and capturing complex temporal dependencies with LSTM. Experiments show that the designed model has strong abilities in predicting groundwater levels and identifying earthquake precursor anomalies. To enhance the accuracy of anomaly detection, this study employs an Exponentially Weighted Moving Average (EWMA) control chart to precisely pinpoint the onset of anomalies. Through validation with earthquake in Jianshui County, Yunnan Province, China, the model effectively identified groundwater level anomalies under different geological conditions, confirming its generalization and practicality. Finally, this article conducted cross validation on the designed model, which has improved its reliability in practical applications. This study has certain scientific innovation and practical value in earthquake precursor analysis, providing new technical support and analysis methods for the development of earthquake warning technology and disaster prevention and reduction work.
A landslide is one of the most common geological disasters, which is associated with great destructive power and harm. In recent years, semantic segmentation models have been applied to landslide recognition research and have made some achievements. However, the current method still has issues, overlooking small targets like fine cracks, missegmenting boundaries, and struggling to differentiate spectral signatures such as those of different rock types in landslide-prone areas. In this paper, a landslide detection model based on the DeepLabv3+ framework, DeepLabv3+-ResNet101-ECA, is proposed. The backbone feature extraction network of DeepLabv3+ is replaced with ResNet101 to enhance the feature extraction ability of the model for small objects. The ECA attention mechanism is integrated into the model to improve the accuracy of the object segmentation and improve the detection accuracy. Taking the landslide in Bijie City, Guizhou Province, as the research object, compared with the original DeepLabv3+ model, the precision of DeepLabv3+-ResNet101-ECA is increased by 1.17%, the recall rate is increased by 2%, the F1 score is increased by 0.96%, and the MIou is increased by 2.36%. Finally, transfer learning is used to verify the generalization ability of the model. The results show that the improved model has a better detection effect on landslides.
Pulse signals refer to electromagnetic waveforms with short duration and high peak energy in the time domain. Spatial electromagnetic pulse interference signals can be caused by various factors such as lightning, arc discharge, solar disturbances, and electromagnetic disturbances in space. Pulse disturbance signals appear as instantaneous, high-energy vertical-line pulse trains (VLPTs) on the spectrogram. This paper uses computer vision techniques and unsupervised clustering algorithms to process and analyze VLPT on very-low-frequency (VLF) waveform spectrograms collected by the China Seismo-Electromagnetic Satellite (CSES) electric field detector. First, the waveform data are transformed into time–frequency spectrograms with a duration of 8 s using the short-time Fourier transform. Then, the spectrograms are subjected to grayscale transformation, vertical line feature extraction, and binarization preprocessing. In the third step, the preprocessed data are dimensionally reduced and fed into an unsupervised K-means++ clustering model to achieve automatic recognition and labeling of VLPTs. By recognizing and studying VLPT, not only can interference be recognized, but the temporal and spatial locations of these interferences can also be determined. This lays the foundation for identifying VLPT sources and gaining deeper insights into the generation, propagation, and characteristics of electromagnetic radiation.
The spaceborne Electric Field Detector (EFD) is one of the payloads of the China Seismo-Electromagnetic Satellite (CSES-01), which can measure electric field data at near-Earth orbit for investigating fundamental scientific topics such as the dynamics of the top-side ionosphere, lithosphere–atmosphere–ionosphere coupling, and electromagnetic field emissions possibly associated with earthquake occurrence. The Extremely Low-Frequency (ELF) waveform shows anomalous step variations, and this work proposes an automatic detection algorithm to identify steps and analyze their characteristics using a convolutional neural network. The experimental results show that the developed detection method is effective, and the identification performance reaches over 90% in terms of both accuracy and area under the curve index. We also analyze the rate of the occurrence of steps in the three components of the electric field. Finally, we discuss the stability of the statistical results on steps and their relevance to the probe’s function. The research results provide a guideline for improving the quality of EFD data, and further applications in monitoring the low-Earth electromagnetic environment.
With the rapid growth of population, more diverse crowd activities, and the rapid development of socialization process, group scenes are becoming more common, so the demand for modeling, analyzing, and understanding group behavior data in video is increasing. Compared with the previous work on video content analysis, factors such as the increasing number of people in the group video and the more complex scene make the analysis of group behavior in video face great challenges. Therefore, a group behavior pattern recognition algorithm based on spatio-temporal graph convolutional network is proposed in this paper, aiming at group density analysis and group behavior recognition in the video. A crowd detection and location method based on density map regression-guided classification was designed. Finally, a crowd behavior analysis method based on density grade division was designed to complete crowd density analysis and video group behavior detection. In addition, this paper also proposes to extract spatio-temporal features of crowd posture and density by using the double-flow spatio-temporal map network model, so as to effectively capture the differentiated movement information among different groups. Experimental results on public datasets show that the proposed method has high accuracy and can effectively predict group behavior.