Three-dimensional Simulation Of Particle Movement In Debris Flow Based On GIS And GAN

Today,artificial intelligence is widely studied as a popular subject of interest,owing to its possible application in the field of transportation.As a creative learning method,generative deep learning has a certain significance for the simulation of geo-hazard entities.This paper will discuss and study the natural flow of geohazard particles in a three-dimensional space.Based on the summary of the existing research and the idea of particle simulation and the particle motion data embedded with geographic information,this paper proposes a method of integrating GIS(Geographic Information System)and GAN(Generative Adversarial Networks).This paper presents a model of particle motion data processing which contains geographic information.Firstly,this paper proposes a method to obtain a small number of debris flow path point sets,determines the research object as the natural flow process of mud flow and through the research on the movement characteristics of mud flow,the main movement rules of debris flow particles are obtained.The confluence point of mud flow area is obtained by using the flow and flow direction analysis in hydrological analysis,and then the flow path point of debris flow is determined according to the surrounding terrain environment and elevation value.In addition,this paper introduces a deep learning engineering feature which is suitable for particle motion research.The drift value is obtained by the difference operation between the flow path point and the determined starting point of debris flow,and the drift value is used as the training data set of the neural network.The model builder in Arc GIS Pro software was further used to achieve the integration of each workflow,forming a reusable analysis toolkit.Then,through the introduction of GAN,the distribution function of mud flow path point set is studied,and the motion law of particles is mined,so as to obtain more sample data to be able to meet the sample size demand for simulation.The possible three-dimensional spatial positions of the mud flow in the flow process are expressed in the form of a large number.