GSGuanqun Sheng

Guanqun Sheng

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Guanqun Sheng has a total of 34 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Impact of Dry–Wet Cycles on the Strength Characteristics of Phosphogypsum-Based Composite Stabilized Laterite Soils

    by Kai Huang, Ying Sun, Bin Xu, Xianzeng Shi, Guanqun Sheng, Chengkai Fan

    Laterite soil (LS) is widely distributed and commonly utilized in engineering construction. However, its natural bonding degrades under cyclic dry–wet (D-W) conditions, resulting in volume instability and strength deterioration. To mitigate this problem, phosphogypsum (PG), as a sulfate-bearing industrial by-product, is often used in combination with cement or other stabilizing agents to improve the engineering properties of LS. Nevertheless, the durability and degradation mechanisms of PG-cement (PGC) stabilized LS under D-W cycles remain insufficiently understood. This study systematically investigates the durability of PGC-stabilized LS subjected to D-W cycles through unconfined compressive strength (UCS), scanning electron microscopy (SEM), and mercury intrusion porosimetry (MIP) tests. The results reveal that the UCS of PGC stabilized LS initially decreased after the first three cycles owing to internal loosening, then increased between the third and fifth cycles. Beyond ten cycles, a significant decline in UCS was observed, resulting from the completion of PG-cement hydration and ongoing structural degradation. SEM and MIP tests demonstrated that the degradation mechanism of stabilized soil arises from complex interactions among water, materials, chemical processes, and mechanical forces. The hydrophilicity of clay minerals influenced pore shrinkage and strength loss, with microstructural damage ultimately leading to macroscopic strength reduction. The findings provide both theoretical insights and practical guidance for the long-term performance assessment and green utilization of PGC in geotechnical engineering.

  • DCAF-Net: An Intelligent Denoising Method for Microseismic Signals Based on the Combination of Time and Wavelet Domains

    by Guanqun Sheng, Junyi Pang, Wei Chen, Xingong Tang

    Microseismic data are typically characterized by a low signal-to-noise ratio (SNR), which makes their processing challenging. Therefore, increasing the SNR value of microseismic data by attenuating noise is a critical step in data processing. Traditional denoising methods mostly rely on prior knowledge and have low efficiency in large-scale data scenarios. Recently, intelligent denoising methods have become a focal point in the field of microseismic denoising. However, the existing deep learning methods commonly focus on either the time or frequency domain, thus failing to adequately leverage the time-frequency information in the process of microseismic signal separation from noise. To address this shortcoming, this study proposes a dual cross-attention fusion network named DCAF-Net, which is designed by combining the time and wavelet domains. The proposed DCAF-Net replaces the pooling layers used for downsampling with a discrete wavelet transform, capturing frequency information from the decomposition coefficients and enhancing a microseismic signal’s ability to represent details across multiple scales. In addition, a spatial-transformer attention block based on the Transformer encoding structure is designed to improve the feature extraction capacity of microseismic signals prior to the skip connections. Further, an adaptive attention fusion block is introduced to the proposed network to enhance the fusion of microseismic data at different scales, thus effectively achieving signal-to-noise separation. The proposed model is verified and compared with the representative methods, such as bandpass filtering, wavelet transform, the DnCNN model, and the U-Net model. The results indicate that at a low SNR value, the proposed DCAF-Net model can improve the SNR by an average of 52.46% compared to the U-Net model. Meanwhile, the proposed approach is compared with two representative frequency-domain-based deep learning denoising methods, namely the DeepDenoiser and DeepSeg methods. The results of extensive experiments conducted on both synthetic and field microseismic data demonstrate that the proposed DCAF-Net model can enhance the amplitude preservation and noise suppression capabilities of microseismic signals compared to the existing models, ensuring optimal performance.


  • An Intelligent First-Arrival Picking Method of Microseismic Signals Based on the Small Sample Expansion

    by Wenxuan Huang, Guanqun Sheng, Xingong Tang, Kai Ma, Jingyi Lu, Hang Sun

    An accurate and efficient first arrival picking method of microseismic signals is essential for improving the microseismic recording processing performance. However, weak seismic magnitude and low signal-to-noise ratio (SNR) of microseismic signals make highly accurate and efficient microseismic signal processing challenging when traditional signal picking methods are used. Therefore, to use microseismic signal features effectively and improve both the accuracy and the efficiency of the first arrival picking of microseismic signals, this paper proposes a first arrival picking method of microseismic signals based on deep learning. In addition, a generative adversarial network with latent space and adaptive layer-instance normalization for microseismic signals generation (G-LA-MSG) is used to generate a large number of effective microseismic signals under unsupervised conditions to expand the microseismic data having a limited number of samples. On the basis of the sample expansion, an AND-OR grammar pyramid scene parsing network (AOG-PSPNet) is designed to select the first arrivals of microseismic signals. By adopting the AND-OR grammar block (AOGBlock) structure, the network can fully extract the sequence data features, achieve good SNR separation in a low SNR environment, and improve the accuracy of first arrivals of low SNR microseismic signals. The results of the arrival picking experiments conducted under the condition of a low SNR for synthetic microseismic records and real microseismic records show that the proposed method can achieve a superior performance compared to the commonly used deep learning-based methods for first arrival picking of microseismic signals.


  • Open Access

    Semi-Picking: A semi-supervised arrival time picking for microseismic monitoring based on the TransUGA network combined with SimMatch

    by Qianfang Wang, Guanqun Sheng, Xingong Tang, Kai Xie

    SUMMARY: An accurate and efficient method for picking the first arrival of microseismic signals is crucial for processing microseismic monitoring data. However, the weak magnitude and low signal-to-noise ratio (SNR) of these signals make picking arrivals challenging. Recent advancements in deep learning-based methods for picking the first arrivals of microseismic signals have effectively addressed the inefficiencies and inaccuracies of traditional methods. Nevertheless, these methods often require many training samples, and the substantial size and labelling effort significantly hinder the development of deep learning-based first-arrival picking methods. This study introduces Semi-Picking: a semi-supervised method for picking the first arrival of microseismic signals, utilizing the TransUGA network and SimMatch. This approach automatically labels microseismic signals following sample augmentation by establishing a semi-supervised learning framework, significantly reducing the time required for sample labelling. Initially, the TransUNet model is enhanced by incorporating the Self-Supervised Predictive Convolutional Attention Block (SSPCAB) module to create a Deep-TransUNet architecture, which more effectively separates signal from noise in microseismic signals with low SNR and improves the accuracy of first-arrival picking. Subsequently, the data sets for this study are compiled from microseismic traces collected from field monitoring records. Finite-difference forward modelling is applied to the microseismic data to train the network, and hyperparameter tuning is performed to optimize the UGATIT and Deep-TransUNet architecture. The outcomes of the arrival-picking experiments, conducted under conditions of low SNR using both synthetic and real microseismic records, demonstrated that Semi-Picking offers robust resistance to incorrect labels. This resilience stems from the synergistic use of the semi-supervised learning framework and self-attention mechanisms. The proposed method demonstrates superiority over the TransUNet, the SSPCAB-TransUNet, the UNet++ and the traditional short-term average/long-term average method, respectively, with the picking error rate of the Semi-Picking Net being less than 0.1 s. The proposed method outperforms the commonly used deep learning-based approaches for picking the first arrivals of microseismic signals, exhibiting superior performance.


  • Open Access

    EnNuSegNet: Enhancing Weakly Supervised Nucleus Segmentation through Feature Preservation and Edge Refinement

    by Xiaohui Chen, Qisheng Ruan, Lingjun Chen, Guanqun Sheng, Peng Chen

    Nucleus segmentation plays a crucial role in tissue pathology image analysis. Despite significant progress in cell nucleus image segmentation algorithms based on fully supervised learning, the large number and small size of cell nuclei pose a considerable challenge in terms of the substantial workload required for label annotation. This difficulty in acquiring datasets has become exceptionally challenging. This paper proposes a novel weakly supervised nucleus segmentation method that only requires point annotations of the nuclei. The technique is an encoder–decoder network which enhances the weakly supervised nucleus segmentation performance (EnNuSegNet). Firstly, we introduce the Feature Preservation Module (FPM) in both encoder and decoder, which preserves more low-level features from the shallow layers of the network during the early stages of training while enhancing the network’s expressive capability. Secondly, we incorporate a Scale-Aware Module (SAM) in the bottleneck part of the network to improve the model’s perception of cell nuclei at different scales. Lastly, we propose a training strategy for nucleus edge regression (NER), which guides the model to optimize the segmented edges during training, effectively compensating for the loss of nucleus edge information and achieving higher-quality nucleus segmentation. Experimental results on two publicly available datasets demonstrate that our proposed method outperforms state-of-the-art approaches, with improvements of 2.02%, 1.41%, and 1.59% in terms of F1 score, Dice coefficient, and Average Jaccard Index (AJI), respectively. This indicates the effectiveness of our method in improving segmentation performance.


  • Open Access

    DANC-Net: Dual-Attention and Negative Constraint Network for Point Cloud Classification

    by Hang Sun, Yuanyue Zhang, Jinmei Shi, Shuifa Sun, Guanqun Sheng, Yirong Wu

    Convolutional neural networks, as a branch of deep neural networks, have been widely used in multidimensional signal processing, especially in point cloud signal processing. Nevertheless, in point cloud signal processing, most point cloud classification networks currently do not consider local feature correlation. In addition, they only adopt ground-truth as positive information to guide the training of networks while ignoring negative information. Therefore, this paper proposes a network model to classify point cloud signals based on feature correlation and negative constraint, DANC-Net (dual-attention and negative constraint on point cloud classification). In the DANC-Net, the dual-attention mechanism is utilized to strengthen the interaction between local features of point cloud signal from both channel and space, thereby improving the expression ability of extracted features. Moreover, during the training of the DANC-Net, the negative constraint loss function ensures that the features in the same categories are close and those in the different categories are far away from each other in the representation space, so as to improve the feature extraction capability of the network. Experiments demonstrate that the DANC-Net achieves better classification performance than the existing point cloud classification algorithms on synthetic datasets ModelNet10 and ModelNet40 and real-scene dataset ScanObjectNN. The code is released at https://github.com/sunhang1986/DANC-Net.

  • Arrival-time picking of microseismic events based on MSNet

    by Guanqun Sheng, Shuangyu Yang, Xin-Gong Tang, Xiaolong Guo

    Arrival-time picking of microseismic events is a critical procedure in microseismic data processing. However, because field monitoring data contain many microseismic events with low signal-to-noise ratios (S/Ns), traditional arrival-time picking methods based on the instantaneous characteristics of seismic signals cannot meet the picking accuracy and efficiency requirements of microseismic monitoring owing to the large volume of monitoring data. Conversely, methods based on deep neural networks can significantly improve arrival-time picking accuracy and efficiency in low-S/N environments. Therefore, we have adopted a deep convolutional network that combines the U-Net and DenseNet approaches to pick arrival times automatically. This novel network called MSNet not only retains the spatial information of any input signal or profile based on the U-Net, but also extracts and integrates more essential features of events and nonevents through dense blocks, thereby further improving the picking accuracy and efficiency. An effective workflow is developed to verify the superiority of our method. First, we describe the structure of MSNet and the workflow of our picking method. Then, data sets are constructed using variable microseismic traces from field microseismic monitoring records and from the finite-difference forward modeling of microseismic data to train the network. Subsequently, hyperparameter tuning is conducted to optimize the MSNet. Finally, we test the MSNet using modeled signals with different S/Ns and field microseismic data from different monitoring areas. By comparing the picking results of our method with the results of U-Net and short-term average and long-term average methods, the effectiveness of our method is verified. The arrival-picking results of synthetic data and microseismic field data indicate that our network has increased adaptability and can achieve high accuracy for picking the arrival time of microseismic events.


  • Open Access

    Fast Angle Estimation and Sensor Self-Calibration in Bistatic MIMO Radar With Gain-Phase Errors and Spatially Colored Noise

    by Guanqun Sheng, Han Wang, Fang-Qing Wen, Xinhai Wang

    Multiple-input multiple-output (MIMO) radars are essential in many Internet-of-Things (IoT) applications. Gain-phase error calibration is an important branch in MIMO radar. Existing calibration algorithms are only suitable for white Gaussian noise scenario, however, spatially colored noise is more practical in engineering implementation. In this paper, we stress the problem of direction finding and sensor self-calibration in a bistatic MIMO radar in the co-exist of unknown gain-phase error and spatially colored noise, and a covariance tensor-based parallel factor analysis (PARAFAC) estimator is proposed. To suppress the spatially colored noise and exploit the multi-dimensional structure of the array measurement, a covariance tensor is firstly established and the temporal cross-correlation operation is followed. Then the PARAFAC decomposition is carried out to obtain the factor matrices. Thereafter, automatically paired direction estimation is achieved via least squares. Finally, the element-wise multiply/divide technique and the Lagrange multipliers are adopted to obtain the gain-phase error vectors. The algorithm is analyzed in terms of identifiability and Cramer-Rao bound (CRB). Numerical simulations results show the effectiveness and improvement of the proposed estimator.