Generalisation and Multiple Representation

Generalisation quality and conflict detection

5704.3 - Detecting changes in Openstreetmap building footprints with combined shape analysis

Wednesday, July 5
4:50 PM - 5:10 PM
Location: Virginia B

1. Background
Currently volunteered geo-data such as Openstreetmap (OSM) has been collected at an unprecedented speed and updated every minute. Besides the known problems in data quality (Haklay 2010, Neis et al. 2012), crowd-sourced geo-data can still be used as timely sources for change analysis and data revision. On the other hand, authoritative geographical data such as those maintained at national mapping agencies are usually updated much slower. For example, topographic data in Kadaster, the Netherlands has been revised every two years since 2010 (Stoter 2009), possibly due to its full coverage update mechanism. Those more commercial oriented data providers usually have a quarterly update cycle for their navigation and POI data. This can lead to a situation that, though claimed with better quality, authoritative or commercial data in many countries may always be outdated by its volunteered counterparts, at least at locations such as urban regions (Zielstra & Zipf 2010, Fan et al. 2014). It is therefore a reasonable choice to combine the strength of both sides (i.e. volunteered and authoritative data). For example, by employing continuous revision methods, Ordnance Survey, UK has shortened the updating cycle of their OS MasterMap Topography Layer to six weeks.

This paper presents techniques that compare building outlines in two data sets, and discover data discrepancies that may be caused by physical changes to the objects rather than by other causes. In particular, we compare OSM buildings with authoritative topographic data at 1:10k, and detect plausible changes for future updating processes.

2. Methods
Changes to building objects can be mainly identified by comparing the similarity in positions and shapes. However, many reasons may explain the discrepancy between two data sets. First, specifications used in data collection can be different. For example, cadastral application would collect a building outline along with its annexes such as entries and garages. These annex structures are not included in topographic datasets, where only outlines of building roofs are recorded. Second, the datasets to be compared can be maintained at different scales. This can further complicate the change detection process because shapes at a smaller scale can be simplified, aggregated and displaced. To minimize such an effect we choose to compare OSM with topographic data at 1:10k (TOP10NL), where the latter is outdated compared with the former data.

To detect changes in OSM as compared to TOP10NL, we used a rule-based technique which combines evidence from different similarity measures. In addition to similarity measures in building size and orientation, we measure the shape similarity between buildings using an adapted turning function. Our adaptation to the classic turning function is discussed in detail in this paper. To summarize, the turning function is used in two different ways: (a) as a global similarity measure and (b) as local partial similarity detector. These shape descriptors are useful to identify if shapes on both data sets are globally similar and dissimilar, or if they have changed only in part of their shapes.

3. Results and conclusions
Experimental results suggest that changes to building objects happen even for well developed areas, and that the change rate is not very high (less than 36% in industrial areas). By visual inspection, we find that the precision of change detection is satisfactory (over 91%) in our test data. Types of changes detected include: new objects constructed, objects destructed, objects expanded and contracted (partial changes), and objects unchanged. The extracted information can be used for continuous updating of multi-scale databases and can be used in urban studies to understand the urbanization process.

Xiang Zhang

Associate Professor
Wuhan University

Xiang Zhang received his Ph.D. degree from the Faculty of Geo-Information Science and Earth Observation (ITC), University of Twente, the Netherlands. He is now an associate professor in the Department of Cartography and Geographical Information Science, Wuhan University. His research interests include automated cartography and volunteered geographical information, its processing, quality assessment and integration.


Send Email for Xiang Zhang

Xiongfeng Yan

School of Resource and Environmental Science, Wuhan University

Xiongfeng Yan received the BS and MS degree in cartography from Wuhan University, China, and is pursuing his PhD in Wuhan University. His main research interests include data update and cartography generalization. Has published 3 papers in refereed conferences and journals.


Send Email for Xiongfeng Yan

Weijun Yin

Wuhan University

Weijun Yin is a MSc student in the Department of Cartography and Geographical Information Science, Wuhan University. He is currently working the subject of integration and updating of spatial data.


Send Email for Weijun Yin

Tinghua Ai

Wuhan University

Tinghua Ai received the PhD in 2000 in the topic of generalization from Wuhan University, China. He is a full professor and head of department of cartography and geography information engineering at school of resource and environment sciences, Wuhan University. His main research interests include map generalization, visualization and spatial cognition. He has published more than 100 papers in refereed conferences and journals.


Send Email for Tinghua Ai

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


5704.3 - Detecting changes in Openstreetmap building footprints with combined shape analysis

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 Detecting changes in Openstreetmap building footprints with combined shape analysis