Research On Method For Digital Landform Classification Of Linear Engineering

Linear engineering refers to highway,railway,oil,urban comprehensive pipe network,transmission line and other comprehensive corridor projects.With the rapid development of China’s economy,the trend of continuous growth was promoted by all kinds of linear engineerings along with new infrastructures.Due to the particularity of linear engineering project,the wide areas were covered and the variety of landform units were crossed by construction long paths.The design of engineering,the disaster prevention and the landscape protection were affect a lot by the complex terrain environment.How to realize the digital terrain classification of linear engineering scientifically and efficiently is an important challenge for the fine construction of linear engineering.Under the background of digitization,the linear engineering terrain classification has some problems,such as single classification feature,vague knowledge of classification decision and great influence of subjective factors.The research of linear engineering landform classification was supported by high-accuracy data and advanced technical according to the development of digital elevation model(DEM)and digital terrain analysis(DTA).How to use the digital terrain analysis technology to mine the terrain information that contained in the linear engineering domain DEM needs to be further studied.Aim at the problems and defects of the classification in linear engineering terrain,the digital strip landform of linear engineering was described,extracted and classified based on the basis of digital terrain analysis theory,highway engineering,transmission line engineering and pipeline engineering are selected as representatives.A digital classification method applicable to linear engineering landform was proposed,the corresponding research results were achieved.The main research work includes:(1)A new landform classification decision scheme of highway strip landform was proposed according to the requirements of topographic classification in Design Specification for Highway Alignment(JTG D20-2017)and Technical Standard of Highway Engineering(JTG B01-2014).The decision scheme was made base on slope,relative elevation,relief degree.And the random forest algorithm(RF)in the field of artificial intelligence was introduced to improve the traditional RF algorithm according to the imbalance characteristics of the highway strip landform data,and an automatic classification method for highway strip landform data is proposed.The highway grid DEM was used as the data source for the classification experiment.The experimental results showed that the improved RF method could effectively realize the automatic classification of the highway strip landform,the complex data superposition and tedious grid calculation process in the overlay analysis method were avoided,the valley,plain,mountain,micro hill and heavy hill into five typical highway landform types were reasonably divided,and the classification results caused by the uneven distribution of the original sample data were improved.(2)Aiming at the problems of high intensity,low efficiency and strong subjectivity of traditional manual drawing and extraction of micro topography of transmission lines,according to the micro topography classification system in the operation specification for overhead transmission lines,a new method based on topographical position index(Topographical Position Index)was proposed,Based on the digital terrain analysis technology,the automatic extraction method of transmission line micro terrain was established.The grid DEM data and micro terrain disaster classification data of nine transmission lines in a power grid company from 2012 to 2018 were extracted and verified.The experimental results showed that this method can effectively extract five types of typical transmission line micro terrain,including pass terrain,mountain terrain,uplift terrain,canyon terrain and water vapor terrain,and reveal the correlation between micro terrain distribution and line disaster occurrence,which cpuld be used as reference provide technical support for power grid construction feasibility demonstration and fine line design.(3)The experience and achievements of predecessors on the landform of pipeline engineering were summarized,and the influence of landform on the construction of pipeline engineering was systematically analyzed by this paper,a set of pipeline engineering terrain classification method which meets the needs of engineering construction and has good adaptability according to the actual situation that the research has not formed a systematic and unified understanding of pipeline engineering landform classification was proposed,Digital landform classification of pipeline engineering was realized by layer superposition.On this basis,dynamic process of area change and transfer of each landscape type under different pipeline landform types were studied.The intensity of landscape change was analyzed and the spatial distribution law of landscape types was explored in pipeline engineering domain on different landform,the certain theoretical and practical guidance were provided for pipeline engineering design optimization,line maintenance management and ecological environment coordinated and sustainable development.In this paper,the influence of landform factors on linear engineering was systematically analyzed,and the classification of linear engineering terrain was systematically studied by combining machine learning algorithm,disaster prevention,landscape analysis and digital terrain analysis.A set of digital terrain classification methods for linear engineering based on grid DEM was proposed,which made up for and expands the existing methods and the ideas of engineering terrain classification.A scientific reference for linear engineering construction was provided by this study,which is an innovative exploration of combining digital terrain analysis technology with engineering landform analysis.