Minghua Luo

Landslides pose severe threats to life and property, necessitating rapid and accurate identification for effective hazard assessment and emergency response. This study proposes an improved SegFormer model for precise landslide extraction from high-resolution remote sensing imagery. To address the limitations of existing segmentation methods in handling complex backgrounds and irregular boundaries, the proposed framework integrates several structural enhancements. These include a squeeze-and-excitation module to suppress background noise, an auxiliary edge-fusion branch to capture explicit boundary details, and an adaptive feature gating mechanism to refine feature representation. The model is trained using focal loss to mitigate class imbalance and employs a three-stage recognition process, culminating in post-processing with dense conditional random fields for boundary refinement. Herein, the experimental results on a dataset of 100 high-resolution satellite images demonstrate that this approach outperforms the classical U-Net architecture and traditional thresholding techniques. The model achieved an accuracy of 0.976, a precision of 0.918, a recall of 0.947, and an F1-score of 0.932. These findings confirm that the proposed method offers exceptional accuracy and robustness, providing an effective automated tool for large-scale landslide detection in complex terrain.

Understanding the short-term underwater settling behavior of classified tailings is important for interpreting deposit formation in ponded areas. One-dimensional settling-column tests were conducted for 36 h to examine how fine-particle content (<0.075 mm) and nominal slurry mass concentration were associated with soil-water interface evolution, visually defined settling phases, and cumulative interface displacement. The observations showed an early rapid-settling period, a post-interface settling/compression period, and a later low-rate compression period. After 36 h, the columns displayed a clear-water zone above a visually layered deposit. Across the six treatment means, 36 h cumulative settlement decreased as fine-particle content or nominal mass concentration increased. These are time-specific descriptive trends, not evidence of lower ultimate settlement: equilibrium was not demonstrated, only duplicate runs were performed, and the archived concentration-series records do not contain the run-specific initial volume, height, or component masses needed to isolate a concentration effect. No rheological, pore-pressure, microstructural, strength, or water-content measurements were made. The results therefore describe short-term behavior in a static laboratory column and should not be directly extrapolated to field-scale tailings facilities.