Infrasound signal classification represents a critical challenge that demands immediate attention. Feature extraction stands as the core concept for enhancing classification accuracy in infrasound signal processing. However, existing feature extraction methodologies fail to meet the requirements for long-distance detection scenarios. To address these limitations, this study proposes a novel classification framework based on the spatiotemporal characteristics of infrasound signals. The proposed framework incorporates advanced signal processing techniques, signal enhancement algorithms, and deep learning architectures to achieve precise classification of infrasound signals. This paper designs three sets of comparative experiments, and the results demonstrate that the proposed method achieves a classification accuracy rate of 83.9% on chemical explosion and seismic infrasound datasets, outperforming eight other comparative classification methods. This substantiates the efficacy of the proposed approach.
Infrasound signal classification remains a critical challenge in geophysical monitoring systems, where classification performance is fundamentally constrained by feature extraction efficacy. Existing two-dimensional feature extraction methods suffer from inadequate representation of spatiotemporal signal dynamics, leading to performance degradation in long-distance detection scenarios. To overcome these limitations, we present a novel classification framework that effectively captures spatiotemporal infrasound characteristics through Gramian Angular Field (GAF) transformation. The proposed method introduces an innovative encoding scheme that transforms one-dimensional infrasonic waveforms into two-dimensional GAF images while preserving crucial temporal dependencies. Building upon this representation, we develop an advanced hybrid deep learning architecture that integrates ConvLSTM networks to simultaneously extract and correlate spatial and spectral features. Extensive experimental validation on both chemical explosion and seismic infrasound datasets shows our approach achieves 92.4% classification accuracy, demonstrating consistent superiority over four state-of-the-art benchmark methods. These findings demonstrate the effectiveness of the proposed method.
23 February 2026 in Geophysical Journal International
SUMMARY: As a critical category of geophysical data, magnetic anomalies play vital roles in geological interpretation, resource exploration and target detection. For most applications involving magnetic anomaly data, the ideal data set should have uniformly distributed data points, high resolution and completeness without gaps. However, because of the environmental constraints and measurement limitations, magnetic anomaly data obtained from real-world measurements often fail to meet these requirements. Thus, interpolation techniques present effective and cost-efficient technical approaches for processing measured magnetic anomaly data to meet the aforementioned criteria. To our knowledge, current research on magnetic anomaly data interpolation has primarily focused on gridding methods for interpolating irregularly sampled data into gridded data and super-resolution interpolation methods aimed at enhancing spatial resolution. Meanwhile, studies on interpolation methods specifically designed to fill large-area data gaps remain relatively scarce. To address the challenge of reconstructing large-area missing magnetic anomaly data, we propose a data-driven method for magnetic anomaly data gap filling. First, based on the analysis of the characteristics of magnetic anomaly data, we construct an open-source magnetic anomaly interpolation data set (MAID) specifically designed for magnetic anomaly data interpolation tasks. Subsequently, we develop a magnetic anomaly data gap-filling generative adversarial network (MADGF-GAN) tailored for magnetic anomaly data gap filling. Upon sufficient training on the MAID training set, MADGF-GAN can directly fill gaps in given magnetic anomaly data. Finally, the effectiveness of MADGF-GAN is validated using four test samples from the MAID test set and Afghan aeromagnetic data. Compared with four existing interpolation methods, MADGF-GAN demonstrates considerable advantages in terms of interpolation accuracy, computational efficiency and practicality. This study demonstrates the potential of data-driven approaches in magnetic anomaly data processing, providing crucial technical support for related geoscientific applications.