Management of spatial data for visualization on mobile devices

Vector-based mapping is emerging as a preferred format in Location-based Services(LBS), because it can deliver an up-to-date and interactive map visualization. The Progressive Transmission(PT) technique has been developed to enable the e�cient transmission of vector data over the internet by delivering various incremental levels of detail(LoD). However, it is still challenging to apply this technique in a mobile context due to many inherent limitations of mobile devices, such as small screen size, slow processors and limited memory. Taking account of these limitations, PT has been extended by developing a framework of e�cient data management for the visualization of spatial data on mobile devices. A data generalization framework is proposed and implemented in a software application. This application can signi�cantly reduce the volume of data for transmission and enable quick access to a simpli�ed version of data while preserving appropriate visualization quality. Using volunteered geographic information as a case-study, the framework shows exibility in delivering up-to-date spatial information from dynamic data sources. Three models of PT are designed and implemented to transmit the additional LoD re�nements: a full scale PT as an inverse of generalisation, a viewdependent PT, and a heuristic optimised view-dependent PT. These models are evaluated with user trials and application examples. The heuristic optimised view-dependent PT has shown a signi�cant enhancement over the traditional PT in terms of bandwidth-saving and smoothness of transitions. A parallel data management strategy associated with three corresponding algorithms has been developed to handle LoD spatial data on mobile clients. This strategy enables the map rendering to be performed in parallel with a process which retrieves the data for the next map location the user will require. A viewdependent approach has been integrated to monitor the volume of each LoD for visible area. The demonstration of a exible rendering style shows its potential use in visualizing dynamic geoprocessed data. Future work may extend this to integrate topological constraints and semantic constraints for enhancing the vector map visualization.

Name dc:type.qualificationname

phd

Level dc:type.qualificationlevel

doctoral

Grantor dc:publisher.institution

National University of Ireland Maynooth

Year dc:date.issued

2012

Author dc:creator

• Ying, Fangli

• Computer Science

Language dc:language

en