Day1-1440-High precision mapping with streetlevel imagery

Friday 2:40pm, Main hall
Oindrila Gupta

http://2017.stateofthemap.org/2017/high-precision-mapping-with-streetlevel-imagery/

Transcript

1. High Precision Mapping with Street-level imagery
2. Current state of the map:
3. 116.8million highways Source: Jennings Anderson
4. Source: Jennings Anderson 28.6% (33million) roads have names
5. • 28.6% have names • 9.6% are oneways • 6.4% have speed-limits • 4.2% have access tags
6. 242million buildings Source: Jennings Anderson
7. • 3.6% have building levels • 1.6% have names
8. 11.6million amenities 47% have names
9. How do we improve this?
10. Most common sources for remote mapping:
11. Satellite and aerial imagery
12. Third party data
13. GPS Traces
14. Street-level imagery
15. What is Street-level imagery?
16. Images having: • latitude, longitude • altitude (optionally) • bearing
17. Most common open-licensed, street-level imagery providers
18. Editing OpenStreetMap using street-level imagery
19. has plugins for: • Mapillary • OpenStreetCam
20. Download plugins: ‘Preferences’ -> ‘Plugins’
21. Open plugins from ‘Imagery’
24. has support for Mapillary
25. In ‘Map Data’ turn on ‘Photo Overlay’ and ‘Traffic Sign Overlay’
27. Mapping common features from street-level imagery
28. Traffic Signs
29. detects traffic signs from images
40. Building Details
42. How we use street-level imagery at Mapbox
43. Mapping Navigational assets
44. osm-navigation map: mapbox.github.io/osm-navigation-map/
45. Verifying user feedbacks
46. Detecting harmful/incorrect edits
47. Future Potential
48. Source: mapbox.github.io/osm-analysis-collab/editor-density Mappers are everywhere
49. Current Mapillary coverage
50. Email: oindrila@mapbox.com Twitter: Oindrila_G
51. Questions?