Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Gis cience 2010 presentation

596 views

Published on

How to detect stops in GPS data? An exploratory spatial data analysis approach

Published in: Technology
  • Be the first to comment

Gis cience 2010 presentation

  1. 1. Uncovering patterns of movement suspension<br />D. Orellana, M. Wachowicz, H.J. De Knegt, A. Ligtenberg and A. K. Bregt<br />Centre for Geo-Information<br />Wageningen University<br />
  2. 2. Motivation<br />One of the main tasks on movement analysis is to detect the places where moving entities stop.<br />Suspension patterns represent the reduction of speed observed when moving entities stop.<br />
  3. 3. Limitations related with the approach<br />Predefined regions<br />Spatial thresholds<br />Speed thresholds<br />+ Temporal thresholds<br />
  4. 4. Limitations related with observation<br /><ul><li>There are no observations of speed=0
  5. 5. Slow movement may be undistinguishable from inaccurate observations.</li></li></ul><li>Our approach<br /><ul><li>For different kind of moving objects.
  6. 6. No need of spatial and temporal thresholds.
  7. 7. Scale-independent.</li></ul>Developing a method to detect movement suspension patterns for collectives of objects using the spatial association of movement vectors.<br />
  8. 8. Different representations of movement<br />Movement astrajectories<br />Movement as vectors<br />
  9. 9. Spatial association of movement vectors<br />A Local Indicator of Spatial Association (LISA) is used to find spatial clusters of low speed vectors. <br />(LISA) cluster map for under-five mortality rate in Nigeria:O. A. Uthman 2008.<br />
  10. 10. Methodology<br />
  11. 11. Implementation<br />Four datasets representing the movement of different kind of entities:<br />Children playing an urban mobile game.<br />Visitors walking in a recreational area.<br />Trucks delivering cement in a city.<br />Elephants moving in a Natural Park<br />
  12. 12. m<br />- 419 children in Amsterdam<br />- 61782 vectors<br />- 10 secondsfor 6 days<br />55 Clusters:<br /><ul><li>Checkpoints
  13. 13. GameEvents
  14. 14. Pedestriancrossings</li></ul>Urban Mobile Game<br />
  15. 15. - 372 pedestriansvisitingthepark<br />- 141826 vectors<br />- Variable time ratefor14 days<br />Dwingelderveld National Park<br />m<br /><br />152 Clusters:<br /><ul><li>Parkings
  16. 16. Attractions
  17. 17. Facilities
  18. 18. Cross - paths</li></li></ul><li>Trucks in Athens<br />[Theodoridis 2003] <br />- 50 Trucks delivering cement<br />- 111 419 vectors<br />- 30 seconds for 40 days<br />m<br /><br />z<br />252 Clusters:<br /><ul><li>Distributionpoints
  19. 19. Buildingprojects
  20. 20. Trafficjams?</li></li></ul><li>Elephants in South Africa<br />[De Knegt, 2003] <br />m<br /><br />z<br />Å<br />- 5 Elephants in South Africa<br />- 100 337 vectors<br />- 1 hourfor30 months<br />85 Clusters:<br /><ul><li>River banks
  21. 21. Water bodies
  22. 22. Fenced places</li></li></ul><li>Overviewof all experiments<br />Proportion of vectors classified as suspension and number of spatial clusters of movement suspension detected in each experiment.<br />Elephants<br />7%<br />85<br />Trucks<br />32%<br />252<br />Visitors<br />6%<br />152<br />Children <br />18%<br />51<br />
  23. 23. Evaluation of the results<br />In the Dwingelderveld National Park, over 90% of movement vectors were associated to relevant spatial features<br />False Positives: 9.1%<br />Unknown<br />Cross paths<br />Facilities<br />Attractions<br />Parking lots<br />Movement<br />Vectors<br />Vectors classified <br />as suspension<br />100%<br />80%<br />True Positives: 90.9%<br />60%<br />40%<br />20%<br />0%<br />
  24. 24. m  z Å <br />Given a movement dataset, the suspension patterns are spatial clusters of movement vectors that fulfil three conditions simultaneously: (a) having speed values below the average for a given data set; (b) having a positive local spatial association of these speed values; and (c) having a minimum statistical significance score of this association corresponding to an established confidence level.<br /><ul><li>Different kind of objects moving in different landscapes
  25. 25. Different spatial a temporal scales
  26. 26. Without thresholds</li></li></ul><li>Limitations<br /><ul><li>Different of moving entities together.
  27. 27. Large datasets.
  28. 28. Real time.</li></li></ul><li>Future work<br />Contextualisation and interpretation.<br />Temporal association of suspension patterns.<br />
  29. 29. Thank you!<br />Daniel Orellana<br />Centre for Geo-Information<br />Wageningen UR<br />daniel.orellana@wur.nl<br />http://wu.academia.edu/DanielOrellana<br />

×