Data-Driven Stability Analysis of Rock Slopes Based on “ArcGIS+3S+AI”

Read the full article See related articles

Listed in

This article is not in any list yet, why not save it to one of your lists.
Log in to save this article

Abstract

This study focuses on typical slopes in the Qingyuan Ma’anshan Mining Area,China,analyzing land use changes in the strongly deformed front edge of the slope by integrating high-resolution remote sensing imagery,artificial intelligence models,and ArcGIS technology,and comprehensively monitoring slope disasters at Tangjiao slope using 3S technology.Using the sponge city concept as an entry point for mining ecological restoration,combined with software such as ArcGIS to construct an SWMM model for designing sponge transformation schemes,simulating slopes under rainfall events of different return periods,and evaluating cumulative displacement and slope susceptibility in the Qingyuan Ma’anshan Mining Area using“3S+ArcGIS technology+artificial intelligence models.” Firstly,based on high-resolution remote sensing images of the Ma’anshan slope acquired from 2020 to 2025,3S technology was employed to obtain slope hazard-inducing environmental factors,and a detailed analysis of the geological conditions and slope distribution patterns in Ma’anshan was conducted.On this basis,object-oriented image segmentation and nearest neighbor classification methods were applied for classification,followed by an analysis of slope deformation and failure characteristics,as well as an investigation into the deformation and instability mechanisms of the Ma’anshan slope.Secondly,a combined slope displacement prediction model based on multivariate chaos theory and the Extreme Learning Machine(ELM)was proposed,with model validation performed using cumulative displacement data from the Ma’anshan slope in Qingyuan City as a case study.Finally,based on time-frequency spectrum deep learning combined with the DBSCAN clustering algorithm,the structural surface roughness and shear strength of the rock slope are determined.A convolutional neural network is introduced for deep learning to automatically extract features from the time-frequency spectrum and train the model,providing reference value and practical significance for estimating the shear strength parameters of engineering rock mass structural surfaces.A preliminary slope stability evaluation method based on improved D-S evidence theory selective integration is established to enhance the robustness and determinacy of slope stability evaluation results. Therefore,addressing issues such as the estimation of rock shear strength parameters,grouping of dominant rock mass structural surfaces,determination of structural surface roughness and shear strength,slope stability prediction,and reliability analysis for the slopes in Ma’anshan,Qingyuan City,this study introduces methods and technologies including ArcGIS,3S,D-S evidence theory,deep learning,and artificial intelligence.A data-driven rock slope stability analysis method is proposed to provide a methodological basis and valuable reference for slope disaster prevention and control.

Article activity feed