State of the Map Asia (SotM-Asia) is the annual regional conference of OpenStreetMap (OSM) organized by OSM communities in Asia. First SotM-Asia was organized in Jakarta, Indonesia in 2015, and the second was organized in Manila, Philippines in 2016. This year’s conference, third in the series, was organized in Kathmandu, Nepal on September 23 – 24, 2017 at Park Village Resort, Budhanilkantha, Kathmandu, Nepal.
We brought nearly 200 Open Mapping enthusiasts from Asia and beyond to this year’s SotM-Asia. The event provided an opportunity to share knowledge and experience among mappers; expand their network; and generate ideas to expand map coverage and effective use of OSM data in Asian continent. We chose ‘from creation to use of OSM data’ as the theme of this year’s conference, emphasizing on the effective use of OSM data. We also brought together a government panel from four different countries in this year’s SotM-Asia. We believe this event will deepen the bond and enhance collaboration among OSM communities across Asia.
More information about the conference can be found on: http://stateofthemap.asia.
Enhancing Worker Digital Experience: A Hands-on Workshop for Partners
Eugene Lisovskiy, CEO, MAPS.ME | SotM Asia 2017 |The Future of OSM in Asia and Beyond
1. Eugene Lisovskiy
CEO of MAPS.ME
Artificial Intelligence, Neural Networks,
Augmented Reality and Robotics:
New Challenges for OSM.
2. ❖ Trends in imagery Industry
❖ Neural networks educated by OSM community.
❖ Dashcams and Navigation with Augmented Reality (AR).
❖ Commercial data automatic imports.
❖ Regular mobile maps users data (GPS traces & POIs).
❖ Video/radar/lidar data from semi/self driving cars/robots.
❖ Summary: The Future of OSM Mapping
Agenda
5. Satellite Imagery
Pros:
❖ Cheaper low-orbit satellite launch (thank to SpaceX).
❖ Higher resolution imagery (30cm/pixel already).
❖ No problems with governments.
Cons:
❖ 2 years to launch new satellite, still expensive (min $62M).
❖ 1 satellite captures the whole earth in 2 years.
❖ Not enough resolution to read road and street signs.
❖ Clouds, trees.
10. AI and Neural Networks
Pros:
❖ Well educated AI recognise images better than human.
❖ No supercomputer required. Community laptops is enough.
❖ Can work with standard imagery, lidar, radar data.
Cons:
❖ No cons, obviously.
❖ Hi-res source data required for better results.
❖ Still the humans need to verify and educate AI.
31. 1. Acquire new mappers:
➢ Mapathons, SOTM
➢ Working with schools and students
2. Engage existent mappers:
➢ Direct (email/push) marketing
➢ Gamification, loyalty programs & activities
3. Start Using AI:
➢ Hi-res satellite and aerial imagery
➢ Develop algorithms, Educate Neural Networks
4. Partner with Self/Semi driving cars providers
How To Engage More People Into Mapping?
35. More MAPS.ME Users = More Active OSM Mappers
Hardcore
OSM
Mappers
Casual POI
Mappers
New Mobile Maps
Users
Mobile mappers are 18 times more
active than regular mappers.
OpenStreetMap.org
❖ 4000k registered mappers
❖ 30k monthly active mappers
❖ 0,8% of active mappers
MASP.ME
❖ 176k registered mappers
❖ 25k monthly active mappers
❖ 14% of active mappers
36. The Future of OSM Mapping (2020+)
POIs details (contacts, services) along
streets and inside buildings.
Roads details (lines, limits,
surface), building details
(street name, building
numbers).
Buildings,
roads, rivers,
forest, terrain,
1. Satellite + AI + Humans education
Source: hi-res satellite/aerial/drone imagery open network.
Processing: automatic IA recognition + verification and neural
networks education by humans.
2. Cars + AI + Humans education
Source: imagery, lidar, radar data from self/semi driving cars.
Processing: automatic IA recognition + verification and
neural networks education by humans.
3. Humans + AI validation
Source: data partners, mobile edits + user images
Processing: collaborative human validation.