Maplat - Historical map viewer technology that guarantees nonlinear bijective conversion without distortion
1. Maplat - Historical map viewer
technology that guarantees nonlinear
bijective conversion without distortion
Kohei Otsuka – Code for History
1
2. Agenda
• About presenter
• What is Maplat?
• A method of coordinates transformation
• Line-to-line conversion
• Comparison with Stroly
• Application of Maplat
• Summary & Proposal
2
4. • Kohei Otsuka
• Has been working in the location based entertainment business for 15 years
• Mapion Inc. (Now One Compath Inc.) - Participated in the launch of the location game
called "Keitai Kunitori Gassen" and was in charge of both planning and development
directions as well as scenario creation and program implementation.
• ATR-Promotions Inc. (Now Stroly Inc.) - Participated in the launch of a smartphone app
using historical maps, and was in charge of both planning and development directions as
well as program implementation and patent development.
• My current main business is technical support for both map content and platform at HERE
Technologies.
• As an individual volunteer, I have been involved in open source and open data activities,
including the development of Maplat, which was announced this time, and am preparing
to launch Code for History.
4
7. Problems to use historical maps / pictorial
maps (1)
• Historical maps and pictorial maps contain regional attractions and are useful in
tourism and education, but...
• It is difficult for ordinary people with little historical knowledge to read historical
maps and compare them with the current cityscape.
• Conventional GIS technologies can overlap historical maps with current maps, but
they distort historical maps so badly that they lose their appeal.
7
8. Problems to use historical maps / pictorial
maps (2)
• An example of a historical map distorted by conventional GIS technology
8
9. Maplat made it possible to overlay historical
maps over accurate maps without distorting them 9
10. 10
Always align the historical map
with the accurate map at the
center of the display.
Roughly calculate the direction
and scale around the center
point to make it appear that
they overlap roughly as a whole.
12. A method of coordinates transformation (1)
Creating a triangulate network 12
Creating a triangulate network by
establishing GCPs between a
historical map and an accurate map.
13. A method of coordinates transformation (2)
Inside of the outer packaging of the GCPs 13
Historical Map Accurate Map
In the outer packaging of the GCPs,
coordinates are transformed by vector
calculation using triangulate network.
14. A method of coordinates transformation (3)
Outside of the outer packaging of the GCPs 14
Historical Map Accurate Map
Extend the triangulate network area by
defining the map vertices on the accurate
map side with our own equations.
15. 15
A method of coordinates transformation (4)
Outside of the map envelope
Historical Map Accurate Map
Define triangles between the vertices
of the map and the centroid of the
GCPs, and transform the coordinates
with the vector extension.
17. 17Problem of previous logic
Historical Map Accurate Map
Lines may not be converted to lines if
they cross the edges of the
triangulate network.
18. 18
Improvement method:
Adoption of the constrained edge (1)
Historical Map Accurate Map
The constrained edges are always taken
as the edges of a triangulate network.
Constrained edge
19. 19
Improvement method:
Adoption of the constrained edge (2)
Historical Map Accurate Map
If the constrained edges are poly lines,
auxiliary GCPs are inserted automatically.
Constrained edge
21. Stroly is similar historical map technology
established since 2010 21
22. Stroly’s way of coordinates transformation 22
Historical Map Accurate Map
Stroly also uses vector computation,
but uses nearby 3 GCPs instead of
triangulate network to select GCPs.
This destroys the topology of the map.
23. Comparison with Stroly (1)
Errors after bidirectional conversion 23
Conversion by Maplat
Conversion by Stroly
Average conversion error
after bidirectional conversion (pixels)
Maplat: 0.00000000018
Stroly: 11.09373539396
24. Comparison with Stroly (2)
Functional comparison matrix 24
Items Stroly Maplat
Ease of publication ✔:Can be published after editing ×:Manual deployment
Bijective conversion × ✔:Patent pending
Orient/Scale conversion ×:Bugs with mismatched scales ✔:Orient / scale are precisely converted
Line-to-line conversion × ✔
Map overlay ×:Toggle only, slow ✔:Always overlaying, briskly
Off-line operation × ✔:PWA capable
Network environment ×:Internet operation only ✔:Can run locally or on Intranet
HTML embedding ×:IFRAME Embedded Only ✔:DIV embedded, API controllable
Share function ×:Map sharing only ✔:Share maps and viewpoints
Mobile readiness ×:Only their own apps ✔:iOS and Android libraries available
POI features ×:Show info only ✔:Move or jump to any link
Open source × ✔
27. Educational Application 27
Using features that are more familiar than conventional GIS, it has been
adopted for educational fieldworks, such as walking through the history
of the city, and have shown effective educational effects.
Win 3 awards
from Japan MILT!
29. Summary
• By using the same method of coordinate transformation as a conventional GIS and
using real-time scaling and orientation around a central point, Maplat is able to
superimpose the historical map with other maps without distorting it or losing its
appeal.
• A similar approach is called Stroly, but because it is not based on conventional GIS
technology, Maplat has an advantage in the ability to convert lines to lines, as well
as in the ability to bijective conversion by maintaining topology.
• Because of these features, Maplat has shown significant educational effects in
the field work of history education.
29
30. Proposal
• The GCP information required to operate Maplat is also important when processing
historical maps using conventional GIS technology. Therefore, I have given a CC
license to the configuration information of each map including Maplat's GCPs and
have provided them as Open Data.
• Pushing the envelop, I would like to propose a subset of Maplat's map configuration
file as a metadata distribution format that include the GCPs of the historical map.
• The details will be published on the Github wiki at a later date.
30