The completeness of landslide inventories and the selection of evaluation models significantly impact the accuracy of landslide susceptibility assessments. Conventional field geological survey methods and single remote-sensing technology struggle to reliably identify landslides under complex environmental conditions. Moreover, prevalent landslide susceptibility evaluation models are often plagued by issues such as subjectivity and overfitting. Therefore, we investigated the uncertainty in susceptibility modeling from the aspects of landslide inventory quality and model selection. The study focused on Luquan County in Yunnan Province, China. Leveraging multisource remote-sensing technologies, particularly emphasizing optical remote sensing and InSAR time-series deformation detection, the existing historical landslide inventory was refined and updated. This updated inventory was subsequently used to serve as samples. Nine evaluation indicators, encompassing factors such as distance to faults and tributaries, lithology, distance to roads, elevation, slope, terrain undulation, distance to the main streams, and average annual precipitation, were selected on the basis of the collation and organization of regional geological data. The information value and two coupled machine-learning models were formulated to evaluate landslide susceptibility. The evaluation results indicate that the two coupled models are more appropriate for susceptibility modeling than the single information value (IV) model, with the random forest model optimized by genetic algorithm in Group I2 exhibiting higher predictive accuracy (AUC = 0.796). Furthermore, comparative evaluation results reveal that, under equivalent model conditions, the incorporation of a remote-sensing landslide inventory significantly enhances the accuracy of landslide susceptibility assessment results. This study not only investigates the impact of landslide inventories and models on susceptibility outcomes but also validates the feasibility and scientific validity of employing multisource remote-sensing technologies in landslide susceptibility assessment.

Landslide susceptibility assessment (LSA) heavily depends on the completeness of landslide inventories and the interpretability of predictive models. Conventional inventories, based solely on historical records, often fail to identify newly occurring or slow-moving landslides, leading to biased susceptibility estimates. To address this limitation, this study proposes a dynamic LSA framework that integrates multi-source remote sensing data, Extreme Gradient Boosting (XGBoost) modeling, and Shapley Additive Explanations (SHAP), with a case study in Yongsheng County, Yunnan Province, China. This study jointly uses multi-temporal optical remote sensing imagery and Sentinel-1 InSAR (Interferometric Synthetic Aperture Radar) deformation data to update the landslide inventory. Compared with the historical inventory containing 334 landslide points, the updated inventory incorporates an additional 140 deformation-related landslide hazard points. XGBoost models were developed using conditioning factors selected through multicollinearity analysis to evaluate the influence of inventory completeness on model performance. Results show that the model based on the updated inventory achieves a significant improvement in predictive accuracy. SHAP-based interpretation reveals that distance to roads and maximum deformation rate are the dominant factors controlling landslide occurrence, reflecting the combined effects of human activities and dynamic ground deformation. The resulting susceptibility map shows that the Area Under the Curve (AUC) value for susceptibility zoning of the updated sample increases from 0.857 to 0.928, with high and very high susceptibility zones occupying 8.28% of the study area. Overall, the proposed framework improves both the accuracy and interpretability of LSA and demonstrates the effectiveness of multi-source remote sensing data for dynamic landslide hazard assessment in mountainous regions.

In high-resolution remote sensing images, there are areas with weak textures such as large building roofs, which occupy a large number of pixels in the image. These areas pose a challenge for traditional semantic segmentation networks to obtain ideal results. Common strategies like downsampling, patch cropping, and cascade models often sacrifice fine details or global context, resulting in limited accuracy. To address these issues, a novel semantic segmentation framework has been designed specifically for large-format high-resolution remote sensing images by aggregating global and local features in this paper. The framework consists of two branches: one branch deals with low-resolution downsampled images to capture global features, while the other branch focuses on cropped patches to extract high-resolution local details. Also, this paper introduces a feature aggregation module based on the Transformer structure, which effectively aggregates global and local information. Additionally, to save GPU memory usage, a novel three-step training method has been developed. Extensive experiments on two public datasets demonstrate the effectiveness of the proposed approach, with an IoU of 90.83% on the AIDS dataset and 90.30% on the WBDS dataset, surpassing state-of-the-art methods such as DANet, DeepLab v3+, U-Net, ViT, TransUNet, CMTFNet, and UANet.