Generalisation and Multiple Representation

Generalisation quality and conflict detection

5704.1 - Cartographic Conflict Detection for Linear Features in Vector-based Maps for the Web Browser

Wednesday, July 5
4:10 PM - 4:30 PM
Location: Virginia B

Web cartographic applications are shifting from server-generated map renderings towards vector-based maps that are dynamically rendered in the Web browser. Many applications are focusing on encoding Open Street Map (OSM) data with the Mapbox Vector Tile Specification. However, the principles of tiled vector data can be applied to any kind of features, as demonstrated by the vector-based maps implemented in the GeoVITe (Geodata Versatile Information Transfer environment) portal.

GeoVITe provides services for viewing and downloading Swiss national geodata produced by the Federal Office of Topography swisstopo. In 2015/2016, in the frame of Geodata4SwissEDU, a collaborative project substantially co-financed by, GeoVITe changed the technology for viewing vector data from WMS-generated raster maps to vector maps. The vector tiles in GeoVITe are generated on-demand in a compressed GeoJSON format from the original topographic data stored in a PostgreSQL/PostGIS database using a generalization function and delivered to a tile cache using GZIP HTTP compression. In our recent research, we have noticed that the main advantages of using vector tiles such as design flexibility (vectors can be styled when requested), size (vector tiles have smaller file size than comparable raster tiles) and caching (the tiled file structure allows for efficient caching) remain valid even when replacing the Google Protocol Buffers encoding of vector tiles with the more GIS-friendly GeoJSON encoding.

The automatic generalization function needed for generating the vector tiles is generic, thus independent of the original scale of the vector data, due to the fact that the geoportal users visualize the data at different scales (irrespective of the precision of the original vectors). Therefore, we have found that we can achieve an acceptable dynamic generalization of linear features based on a negative exponentiation of the zoom level, resulting in an effective simplification factor valid for any vector dataset covering Switzerland. Additional rules (geometry thinning, followed by simplification and line smoothing) results in an acceptable automatic generalization of vector tiles for different scales.

In this context, cartographic conflicts between the different vector layers arise due to the reduction of geometric complexity. As a consequence, symbolized linear features (e.g. streets) become inconsistent with the background (e.g. land use) or other linear geometries (e.g. rivers). Moreover, considering the map design flexibility inherently linked with the use of vector tiles, in the future additional cartographic conflicts can be inadvertently introduced as soon as a user changes the styling of the map by increasing the width of linear features.

The first step in solving such cartographic conflicts is to detect them. Previous research has proven that detecting and solving cartographic conflicts improve the cartographic output of WMS-rendered maps. In this context, various algorithms (e.g. agent based modeling or exponential back-off algorithm) implementing a “Divide and Conquer” strategy or a “Recursive” strategy can be implemented on the server-side in order to detect and subsequently minimize conflict areas, resulting in overriding SLD rules that can be applied by the WMS server in near real-time (DOI: 10.1080/00087041.2015.1119468). Such conflict resolution algorithms also need to be adapted for vector-based maps.

A promising Web browser technology for porting conflict detection and resolution algorithms to vector-based maps is represented by HTML5 Web Workers. So far, tests made on a 2.6 GHz Intel Core i7 notebook with 512 Web Workers for an area around the city center of Zurich took over 30 seconds for detecting conflicts. The presentation will finalize with a solution for improving performance with the help of tiled auxiliary uniform grids generated together with the vector tiles. The uniform grid encodes and enumerates cartographic conflicts using a space-primary generalization technique by modeling the inter-layer map space in tiled data structures.

Ionut Iosifescu Enescu

Senior Research Associate
ETH Zurich

Ionut Iosifescu Enescu is a senior researcher at the Institute of Cartography and Geoinformation of ETH Zurich. He holds degrees in Geomatics and Computer Science, and obtained his Ph.D. degree in Cartography and Geoinformation from ETH Zurich in 2011. He is the technical project manager of ‘Geodata4SwissEDU’, a cooperation project for developing a Swiss national geoinformation service for education and research. He is a recognized expert in spatial data and software architectures for the geospatial domain.


Send Email for Ionut Iosifescu Enescu

Charalampos Gkonos

Research Associate
ETH Zurich

Charalampos Gkonos is a Research Associate at the Institute of Cartography and Geoinformation at ETH Zurich. He received his Diploma degree in Rural and Surveying Engineering at NTU Athens and his MSc degree in Geomatics at ETH Zurich (MSc Thesis grade: 6.00/6.00, excellent). His research interests include but are not limited to web-cartography, GIS, HCI, sustainable development and cartographic education. His current work focuses on the dissemination and use of geodata for research purposes within the ETH domain besides teaching and cartographic education. Part of his work has been published in the International Journal of Sustainable Energy Planning and Management, the Journal of Location Based Services and the ISPRS International Journal of Geo-Information.


Send Email for Charalampos Gkonos

Cristina M. Iosifescu Enescu

PhD Student
Institute of Cartography and Geoinformation, ETH Zurich

Cristina M. Iosifescu-Enescu is a PhD candidate at the Institute of Cartography and Geoinformation of ETH Zurich, Switzerland. She graduated with a Diploma in Geodesy at the Technical University of Civil Engineering Bucharest in 2003 and a MSc in Psychology at the University of Zurich in 2016. Her research interests focus on Dream Cartography, graphical user interfaces for Web Cartography and the Cognitive Theory of Dreams.


Send Email for Cristina Iosifescu Enescu

Lorenz Hurni

Prof. Dr.
ETH Zurich, Institute of Cartography and Geoinformation

Lorenz Hurni has been Associate Professor of Cartography and director of the Institute of Cartography at the ETH Zurich since November 1996 (Full Professor since October 2003). The emphasis of Hurni's research lies in cartographic data models and tools for the production of printed and multimedia maps. Another focus of research covers interactive, multidimensional multimedia map representations. Under his lead, the prizewinning multimedial "Atlas of Switzerland", commissioned by the Federal Council, as well as the "Swiss World Atlas", the official Swiss school atlas, commissioned by the Swiss Conference of Cantonal Ministers of Education, are being developed.


Send Email for Lorenz Hurni

Timofey Samsonov

Leading Researcher
Lomonosov Moscow State University, Faculty of Geography, Department of Cartography and Geoinformatics

Timofey Samsonov is a leading researcher at the Department of Cartography and Geoinformatics, Faculty of Geography, Lomonosov Moscow State University, Moscow, Russia. Timofey holds a PhD in Cartography (2010) from Lomonosov MSU. His interests include generalization of spatial data, multiscale mapping, terrain mapping and analysis, spatial analysis and automation in cartography. Timofey is active in two ICA Commissions: ICA Commission on Generalisation and Multiple Representation and ICA Commission on Mountain Cartography.


Send Email for Timofey Samsonov


5704.1 - Cartographic Conflict Detection for Linear Features in Vector-based Maps for the Web Browser

Attendees who have favorited this

Please enter your access key

The asset you are trying to access is locked. Please enter your access key to unlock.

Send Email for Cartographic Conflict Detection for Linear Features in Vector-based Maps for the Web Browser