Font Size: a A A

Study On Navigation Safety Situation Analysis And Visualization Of Inland Vessel

Posted on:2017-10-25Degree:MasterType:Thesis
Country:ChinaCandidate:Q Y LiuFull Text:PDF
GTID:2381330566453018Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
Our country has abundant resources of inland waterway transport,water area is vast,but the most important navigable waters environment is complex,ship traffic is dense,the complex inland water environment increases the security risk of the inland navigation vessels objectively.In recent years,the ever-accelerating trend of large vessels and the bridge construction on the pace of inland waterway becomes more obvious and heavy;the number of the accident in complex navigable water environment was increased,such as the Oriental Star accident,Anhui Shenzhou-67 accident,etc.The risk of security is gradually emerging;the importance of pre-control management about inland shipping accidents for inland navigation vessels is increasingly prominent.Visualization is a way of helping people to explore,understand and analyze the data through interactive visual interface,in an intuitive way to present the image of complex or abstract information and make it to be understood quickly and easily.Therefore,building security situation assessment and prediction mechanism of navigation,and analyzing the post-visualization,is very urgent needed,it will help reducing the incidence of shipping accidents,reducing social and economic losses,and has important theoretical value and practical significance for water traffic safety management.Taking river(Yangtze)navigation safety situation as the study object,basing on the cleaning and processing of the ship collected data,extracting features based on the security situation made parallel coordinates,exporter's analysis and spectral clustering method for assessment and prediction security situation are proposed based on set of safety situation assessment of inland vessels navigating analysis and inland navigation safety situation prediction method based on DAE depth study to explore based on parallel coordinates visualization interaction model,as well as the map-based security situation visualization rendering.The main research work as follows:1)Extracting data collection of the inland navigation vessels and elements of the situationBased on AIS system data,channel information,weather information and other data of the inland waters ships navigation,with the actual situation of inland navigation of the ship,it summed up the relevant factors to form the corresponding factor set.In a related information data base of inland waters ships navigation,using parallel coordinates to render factors set data visually,doing interactive operations to extract a greater correlation factor to generate the initial set of feature security situation.Combining with exporters' score,using the spectral clustering algorithm to analyze the correlation between the elements for further,dividing data into clustering,forming inland navigation safety situation feature set.2)Inland navigation safety situation assessment methodAccording to the comprehensive inland navigation safety situation,using the entropy method to determine the elements' weight,this paper proposes to set the security situation assessment based method of navigation inland analysis;assessment of the security situation in the inland navigation of the ship,based on the feature set of data,conduct inland navigation safety posture validation;3)Inland navigation safety trend prediction methodPrediction for inland navigation safety trend prediction method is proposed based on noise from Coding(DAE)depth learning.Construction of inland navigation safety feature vectors situation and determine the predicted state using network coding noise from the feature vector to reconstruct the formation of a new feature vector using Softmax regression classifier to add a new feature vector label study to determine the predicted state.Situation in inland navigation safety feature data set based on experimental verification inland navigation safety situation prediction method based on deep learning DAE correctness.4)Inland navigation safety situation VisualizationThe use of parallel coordinates visualization method to extract inland navigation safety situation of the initial set of elements.Combined inland navigation safety situation assessment results with the initial set of data elements in parallel coordinates visualization presented to analyze the elements of focus,greater impact assessments of inland navigation safety elements of the situation.Using the map to render inland navigation safety situation assessment and prediction results visually.
Keywords/Search Tags:Inland Navigation, the Security Situation, Deep Learning, Parallel Coordinates, AIS Data
PDF Full Text Request
Related items