Streaming Progressive Transmission Of Vectordata For Web Visualization

With the popularity of the Internet, WebGIS is a general trend to publish, share and applygeographical information.WebGIS realizes the data exchange between client and server in the form of Webservices. There are two general data formats for data transmission in Web services including raster andvector. The raster map tiles convert the spatial data to the image format and display in the browserdirectly.But the map images accessed by client have already been processedand cannot be edited andanalyzed as original data.Vector data provided by Web services is mainly used for data query and editingand suitable fordata exchange withsmall amount of data in Internet.When the client needs to deal withlarge-volum vector data, it is difficult to meet this demand for existing WebGIS.This paper aims tosolvethe problemsof large-volum vector data transmission in WebGIS, focusing on improving vector dataorganization modeand using streaming progressive transmission mechanism.Streaming progressive transmission transmits the outline information of data firstly, and thenincreases the details step by step gradually to achieve the demand of data. In this way to shorten theresponse time and improve the system performance.The research is conducted in the following steps. First,by analyzing the basic requirements of the vector data streaming transmission, we study and design aservice framework.The second step is to determine the sequence of progressive transmission according tothe importance of vector elements.The third step is toorganize the vector data structures that ensure eachbatch of elements can be transmitted and processed as an independent uniton the client. The fourth step isto transmit the different batches of data to client usingreal-time transport protocol.Finally,build a prototypesystem to verify the feasibility and availability of this research. The main research results in this paper are as follows:(1) The general service framework is designed in object orientedstreaming progressive transmission.There are three key aspects for this framework, which is selection strategy for spatial features on server,data organization and mechanism suitable for progressive transmission in the network layer, and cache andreconstruction of vector data on client. It covers the key technologies ofspatial feature measure model formap information, vector features selection strategy, vector data coding standards, streaming mediatransport protocol and service,network transmission quality controland vector datavisualization.(2) The measurement modelsfor quantity of information of vector features are defined based oninformation entropy. Weanalyzed the measure methods of geographic information based on informationentropy, and built the measurement models of element geometry factor, spatial distribution factor andthematic attribute factor. These factors are the basis of space elements selection strategy and display order.At the same time, we realized the gradual selection algorithm for point, polyline and polygon elements.These algorithms are composite effects of geometry factor and spatial distribution factor. The spatialfeatures with more contribution to the map can be selected firstly to ensure the data with large informationarriving at the client firstly.(3) The vector data storage mechanism is designed for streaming progressive transmission. Wedesigned a vector data structure which is independent block storage.This independent block structuremeans that a streaming block unit can display independently, not depend on other transport unit.Thisstructure can support the data structure of point, polyline, polygonand other basic geometricand abide bythe OpenGIS standard encoding specification. On this basis, a vector data file converter is developed toconvert an ordinaryvector file to a streaming vector data file with independent block structure. Thisstructure is the key for realizing streaming transmission of vector data in the Internet. (4) The streaming progressive transmission method is based on real-time transport protocol andreal-time transport control protocol. Referencing to the multimedia model, we design the packing andunpacking algorithms aimed to vector data. These algorithms realize the process of sending and receivingvector data, achieving the target of several reciprocating transmission. Then by the key technology ofguaranteed delivery reliability and error recoveryto achieve the application target of vector data multipletransmissions.(5) We develop a prototype system of streaming progressive transmission for vector datavisualization.The system includes vector data streaming organization on server, streaming service onnetwork layer, and progressive display of vector data on client. We test this system from the amount of datatransmission, information transmission rate, response time and the display effect on the client using1:100million Chinese basic geographic databases. The results show that when the vector elements is selected inaccordance to theselection models,the amount of information transfer rate reached to80%of the totalamount basically. However the number of transmitted geometric objects is only10%of total.Transfer timeof streaming model is reduced by more than40%than Web Feature Service. And the response time of thissystem is within5seconds.The WFS is easier to lose response yet in the case of large amount of data.The results of the experimental datademonstratethe technical feasibility and usability of thisresearch.It is illustrated the advantage of using the streaming progressive transmission method for vectordata visualization in the Web. This paper enriches the technical system of the WebGIS forlarge-volumevector data and has great significance to develop the application potential of WebGIS.