Semantic segmentation, a fundamental task in remote sensing, plays a crucial role in urban planning, land monitoring, and road vehicle detection. However, compared to conventional images, multispectral remote sensing images present significant challenges due to large-scale variations, multiple bands, and complex details. These challenges manifest in three major issues: low cross-scale object segmentation accuracy, confusion between band information, and difficulties in balancing local and global information. Recognizing that traditional remote sensing indices, such as the Normalized Difference Vegetation Index and the water body index, reveal unique semantic information in specific bands, this paper proposes a feature-decoupling-based pseudo-Siamese semantic segmentation architecture. To evaluate the effectiveness and robustness of the proposed algorithm, comparative experiments were conducted on the Suichang Spatial Remote Sensing Dataset and the Potsdam-S Aerial Remote Sensing Dataset. The results demonstrate that the proposed algorithm outperforms all comparison methods, with average accuracy improvements of 80.719% and 77.856% on the Suichang and Potsdam datasets, respectively.
Remote sensing image change detection, being a pixel-level dense prediction task, requires both high speed and high accuracy. The redundancy within the models and detection errors, particularly missed detections, generally affect accuracy and merit further research. Moreover, the former also leads to a reduction in speed. To guarantee the efficiency of change detection, encompassing both speed and accuracy, a VMamba-based Multi-scale Feature Guiding Fusion Network (VMMCD) is proposed. This network is capable of promptly modeling global relationships and realizing multi-scale feature interaction. Specifically, the Mamba backbone is adopted to replace the commonly used CNN and Transformer backbones. By leveraging VMamba’s global modeling ability with linear computational complexity, the computational resources needed for extracting global features are reduced. Secondly, considering the characteristics of the VMamba model, a compact and efficient lightweight network architecture is devised. The aim is to reduce the model’s redundancy, thereby avoiding the extraction or introduction of interfering and redundant information. As a result, the speed and accuracy of the model are both enhanced. Finally, the Multi-scale Feature Guiding Fusion (MFGF) module is developed, which strengthens the global modeling ability of VMamba. Additionally, it enriches the interaction among multi-scale features to address the common issue of missed detections in changed areas. The proposed network achieves competitive results on three publicly available datasets—SYSU-CD, WHU-CD, and S2Looking—and surpasses the current state-of-the-art (SOTA) methods on the SYSU-CD dataset, with an of 83.35% and of 71.45%. Moreover, for inputs of size, it is more than three times faster than the current SOTA VMamba-based change detection model. This outstanding achievement demonstrates the effectiveness of our proposed approach.
Traditional studies on Transient Electromagnetic Method (TEM) typically assume that the subsurface medium is non-magnetic. However, in regions with igneous rock accumulations or where the subsurface is rich in ferromagnetic minerals, neglecting the magnetic properties of the underground medium may lead to erroneous interpretations for TEM data. This paper conducts a 3-D TEM forward modeling and inversion study considering the non-uniformity cases of magnetic permeability. 3-D TEM forward modeling employs an edge-based finite element method using unstructured grids and a second-order implicit backward Euler method, achieving a modeling algorithm that simultaneously considers non-uniform models of magnetic permeability and resis-tivity. The accuracy of the modeling algorithm is verified by comparing it with the analytical solution of a homogeneous half-space model and the solution of a 1-D TEM forward modeling algorithm. 3-D TEM inversion employs the L-BFGS algorithm and synthetic examples considering non-uniform magnetic permeability are presented. The inversion results show good recovery for the resistivity and magnetic permeability models. Comparisons with the inversion results that neglect the non-uniformity of magnetic permeability validate the importance of considering the variation of permeability in 3-D TEM forward modeling and inversion.
Remote sensing image change detection, being a pixel-level dense prediction task, requires both high speed and high accuracy. The redundancy within the models and detection errors, particularly missed detection, generally affect accuracy and merit further research. Moreover, the former also leads to a reduction in speed. To guarantee the efficiency of change detection, encompassing both speed and accuracy, a VMamba-based Multi-scale Feature Guiding Fusion Network (VMMCD) is proposed. This network is capable of promptly modeling global relationships and realizing multi-scale feature interaction. Specifically, Mamba backbone is adopted to replace the commonly used CNN and Transformer backbones. By leveraging VMamba’s global modeling ability with linear computational complexity, the computational resources needed for extracting global features are reduced. Secondly, considering the characteristics of the VMamba model, a compact and efficient lightweight network architecture was devised. The aim is to reduce the model’s redundancy, thereby avoiding the extraction or introduction of interfering and redundant information. As a result, the speed and accuracy of the model are both enhanced. Finally, the Multi-scale Feature Guiding Fusion (MFGF) module is developed, which strengthen the global modeling ability of VMamba. Additionally, it enriches the interaction among multi-scale features to address the common issue of missed detection in changed areas. The proposed network has achieved competitive results on three datasets, and remarkably surpasses the current state-of-the-art performance on SYSU-CD, with F1 of 83.35% and IoU of 71.45%. Moreover, for inputs of 256×256 size, it is more than three times faster than the current state-of-the-art Mamba-based change detection model. This outstanding achievement demonstrates the effectiveness of our proposed approach.
The Jiao-Liao-Ji belt is one of three major Paleoproterozoic orogenic belts within the North China Craton, yet its tectonic evolution remains controversial. To address this, we investigated the crustal structure using receiver function analysis of 38 broadband seismic stations from profile NCISP-10, which traverses the Liaodong region in the northeastern North China Craton. Our results reveal that: (1) the Moho depth beneath the profile is 32–34 km and the average crustal Vp/Vs ratio is ∼1.75, consistent with regional values across Liaodong. (2) the Tanlu fault zone comprises two branches in the study area: the Yilan-Yitong fault, which penetrates the Moho with a vertical offset of 2–3 km, indicating that it is a crustal scale normal fault, whereas the Dunhua-Mishan fault shows no significant Moho offset. (3) the boundary between the Jiao-Liao-Ji belt and Longgang block is identified at station 09 (near Benxi County, Liaoning Province), manifesting as a south-dipping normal fault with the Longgang block to the northwest and the Jiao-Liao-Ji belt to the southeast. The interface between the Archean basement and the overlying sedimentary rocks is ∼5 km deep beneath the Taizihe depression (with extensive surface exposure of Archean basement rocks to the northwest in the Longgang block), contrasting with that of ∼13 km deep beneath the Jiao-Liao-Ji belt, suggesting formation through early Paleoproterozoic extensional rifting. Our findings illuminate the deep crustal structure of the Jiao-Liao-Ji belt, with the normal fault between the Longgang block and Jiao-Liao-Ji belt, which supports a rift-and-collision tectonic model for the Jiao-Liao-Ji belt.
Multimodal knowledge extraction has become increasingly important for applications that require compact and informative representations of text-image data, yet existing models often remain computationally expensive for practical use. To address this issue, we propose MKE, a lightweight framework for multimodal knowledge extraction based on knowledge distillation and efficient multimodal fusion. Specifically, MKE transfers linguistic knowledge from a large BART teacher model to a compact StuBART backbone and combines Flash-Attention-based cross-modal interaction with a lightweight nonlinear transformation module. Experiments on the MSMO English news dataset show that MKE achieves competitive multimodal summarization performance with 270M parameters and a computational cost of 155.63 GMACs (311.27 GFLOPs) per sample. We further evaluate inference efficiency under a controlled RTX 4090 setup, and we clarify that deployment on smartphones, IoT systems, and wearable devices remains a promising direction for future work rather than a scenario directly validated in this study.