Research On OSM Spatial Data Redundancy Identification And Cleaning

At present,OpenStreetMap(OSM)is the open source geographic information sharing platform with the largest number of contributors and the richest geographical information in the world.It has the advantages of rich data,open source sharing,high current situation and fast update cycle,but lacks unified data editing standards.Moreover,both professional users and non-professional users can contribute version data through geographical knowledge background or activity rules,resulting in problems such as different quality,data redundancy and incomplete spatial data in OSM,which greatly affect the data application.Aiming at the problem of spatial data redundancy,this paper focuses on clean up the redundancy of point and area data of OSM,then analyzes and discusses the causes of the redundancy in OSM spatial data.For OSM area objects,a building area objects redundancy recognition and cleaning model based on hierarchical rules is proposed.For the OSM point objects,a hierarchical redundancy recognition and cleaning model of interest points is raised,and the validity of the redundancy cleaning model is experimentally studied with the real historical OSM data.The research contents of this paper are as follows:(1)Theoretical basis of redundant data detection and cleaning.Firstly,the causes of OSM spatial data redundancy are explored from three aspects:producers,spatial data maintenance and update,as well as the relationship between state and version.Secondly,the relevant theoretical basis of redundancy detection is summarized from the perspective of similarity,and then the relevant theoretical basis of data cleaning and redundancy data cleaning is systematically summarized,at the same time two methods of redundancy data cleaning are introduced.(2)An OSM building area objects redundancy recognition and cleaning model based on hierarchical rules is put forward.Firstly,the source of building area objects redundancy is investigated.Secondly,under the topological relationship and area overlap ratio,the redundant data is identified,and the redundant data is screened out.Then the attributes,geometry and topology of building elements are considered comprehensively.Nine indexes including attribute information integrity,mean area,threshold area,area difference,regular angle ratio,irregular angle ratio,irregular angle difference,slope coincidence value and comprehensive value were selected to construct corresponding hierarchical rules for different types of building redundancy,and the experimental results proved the rationality and effectiveness of the model.(3)A hierarchical OSM objects redundancy recognition and cleanup model is brought out.Using the concept of Euclidean distance and fuzzy reasoning,the redundancy of points of interest is divided into two types:exact redundancy and fuzzy redundancy.Then the number of versions,editing time,user reputation and other indicators are applied to carry out redundancy cleaning for the two types of data.The exact redundancy is moved by the cleanup method,and the fuzzy redundancy is cleaned by the reservation method.The experimental results show that the model is feasible.