Utilizing deep learning for intelligent monitoring and early warning of slope disasters in public space design DOI Creative Commons
Ting Wang, Ying Wang

Frontiers in Environmental Science, Journal Year: 2025, Volume and Issue: 13

Published: May 13, 2025

Introduction The increasing frequency of slope disasters in urban and recreational public spaces, driven by climate change, presents significant risks to safety sustainable design. Conventional stability monitoring systems rely heavily on static models manual interventions, often lacking adaptability real-time predictive capacity. Earlier methods, including rule-based empirical approaches, use fixed thresholds assess risk factors such as soil moisture, angle, seismic activity. Although machine learning like decision trees support vector machines have improved predictions using historical data, their scalability remain constrained due the need for intensive feature engineering limited ability model complex nonlinear dynamics. Methods This study introduces a novel framework utilizing Deep Learning techniques enable intelligent, early warning disasters. Adaptive Spatial Design Model (ASDM) incorporates geospatial user behavior analytics, environmental sensing dynamically risk. It employs convolutional recurrent neural networks geo-hazard prediction, graph-theoretic optimization decision-making, adaptive spatial strategies enhance accuracy responsiveness changing environments. Results Experimental validation real-world datasets shows that proposed system effectively reduces false alarms improves response times 35% compared traditional methods. integration network-based prediction with planning enhances both precision timeliness disaster warnings. Discussion offers transformative, safe, functional approach management dynamic spaces. advances sustainability resilience optimizing design human-environment interactions. model's changes represents improvement mitigation strategies.

Language: Английский

Comprehensive analysis of sediment grain features and their engineering implications in the Yangtze River source area DOI Creative Commons
Zhijing Li, Xiaoxue Wang, Yujiao Liu

et al.

Frontiers in Soil Science, Journal Year: 2025, Volume and Issue: 5

Published: Feb. 25, 2025

Introduction The particle size characteristics of irregular sediments in the Yangtze River Source Area (YRSA) are pivotal for understanding mechanical properties sedimentary medium. Methods This study utilizes field sediment sampling, laser scanning, laboratory testing, and mathematical statistics to analyze morphological, geometric, mineralogical, accumulation particles region. Results Our findings indicate that YRSA have angular edges deviate from spherical shapes, exhibiting elongated flatter three-dimensional morphologies. In experiment, sliding plate method was used measure angle repose sediments, which found be 36.7° above water 35.9° below water. Both values higher than typical range non-plateau regions, indicating reduced mobility. composed fine-grained coarse-grained soils. distribution is primarily coarse sand (0.5-2.0 mm), fine gravel (2.0-5.0 medium (5.0-20.0 with a significant coarsening trend observed over past six years. mineral composition, dominated by quartz, feldspar, heavy minerals, stable, approximately 70% minerals having hardness ≥ 7 on Mohs scale. most abundant trace elements Ti, Mn, Ba, P, Sr, Zr, Cl. Discussion research reveals markedly different those natural sands necessitating reevaluation conventional theories engineering practices constructions this area. insights profound practically relevant, illuminating transport dynamics alpine river systems supporting sustainable regional development.

Language: Английский

Citations

0
Utilizing deep learning for intelligent monitoring and early warning of slope disasters in public space design DOI Creative Commons
Ting Wang, Ying Wang

Frontiers in Environmental Science, Journal Year: 2025, Volume and Issue: 13

Published: May 13, 2025

Introduction The increasing frequency of slope disasters in urban and recreational public spaces, driven by climate change, presents significant risks to safety sustainable design. Conventional stability monitoring systems rely heavily on static models manual interventions, often lacking adaptability real-time predictive capacity. Earlier methods, including rule-based empirical approaches, use fixed thresholds assess risk factors such as soil moisture, angle, seismic activity. Although machine learning like decision trees support vector machines have improved predictions using historical data, their scalability remain constrained due the need for intensive feature engineering limited ability model complex nonlinear dynamics. Methods This study introduces a novel framework utilizing Deep Learning techniques enable intelligent, early warning disasters. Adaptive Spatial Design Model (ASDM) incorporates geospatial user behavior analytics, environmental sensing dynamically risk. It employs convolutional recurrent neural networks geo-hazard prediction, graph-theoretic optimization decision-making, adaptive spatial strategies enhance accuracy responsiveness changing environments. Results Experimental validation real-world datasets shows that proposed system effectively reduces false alarms improves response times 35% compared traditional methods. integration network-based prediction with planning enhances both precision timeliness disaster warnings. Discussion offers transformative, safe, functional approach management dynamic spaces. advances sustainability resilience optimizing design human-environment interactions. model's changes represents improvement mitigation strategies.

Language: Английский

Citations

0