Pirate Maps
Experiments with portable maps on the Raspberry Pi
Ian Turton, Astun Technology Ltd.
@ijturton
Meet Pip the Pirate
Pip has a problem...
He and his crew are very forgetful and very poor cartographers.
So they can't always find their treasure
Pip's Requirements
- Basemapping of the world's oceans and coastlines
- Ability to store locally generated routes and POI
- Ability to share maps and local data across multiple machines
Pip needs some maps
- Obviously
- But... He's a poor pirate
- So an Open Map
- He gets Open Sea Map and Open Street Map
So, Pip is happy?
No! His data plan is killing him:
Every tile comes over the satellite!
Also he can't share his maps with the crew.
Caching
Pip realises that most tiles don't change (very often) and he could update when he was in port to save roaming charges.
- So he needs to be able to cache tiles locally
- With periodic updates when he needs them.
Software Needs
Remember Pip is still broke (and he has the phone bill still to pay).
- FOSS solution needed
- Local Cache
- Ability to update as required
Pip after some searching chooses MapProxy.
Other Software needs
Pip also needs to be able to store his routes and other useful information (where he buried the treasure).
He picks GeoServer to provide a WFS-T service to QGIS with PostGIS to store his data.
Hardware Needs
At first Pip, thinks he will use his laptop to store the cache. But then other crew members can't access the maps when he has shut it down.
As Pip is (still) broke he looks for a cheap computer.
He could get...
- a server, but cost & power requirements too high
- a laptop, takes up space, screen draws power
- a SoC, promising low power and cheap.
Which System on a Chip?
Pip needs:
- Low Power
- Cheap
- Networking
- Data storage
- Ability to run required software
Raspberry Pi 3
- Cheap ~£35 or ~ €40 or US$45
- Powerful
- 1.2GHz 64-bit quad-core ARM Cortex-A53 CPU
- 1GB RAM (LPDDR2 SDRAM)
- On-board Wireless LAN - 2.4 GHz 802.11 b/g/n (BCM43438)
- 4 x USB 2.0 ports
- Linux
- Add a hard drive (a spare from an old laptop)
Install the Software
sudo apt-get install python-imaging python-yaml libproj0
sudo apt-get install build-essential libjpeg-dev zlib1g-dev
libfreetype6-dev
sudo -H pip install Pillow
sudo -H pip install MapProxy
sudo apt-get install oracle-java8-jdk unzip
wget http://sourceforge.net/projects/geoserver/files/GeoServer
/2.11.1/geoserver-2.11.1-war.zip
unzip geoserver-2.11.1-war.zip
sudo cp geoserver.war /var/lib/tomcat7/webappsAdd Apache
sudo apt-get install apache2
sudo apt-get install apache2-dev
sudo apt-get install python-dev
wget https://pypi.python.org/packages/c8/d5/...
/mod_wsgi-4.5.17.tar.gz
tar xvzf mod_wsgi-4.5.17
./configure
make
sudo make install- add mod_wsgi to run Mapproxy directly
- adjust mod_jk to forward tomcat through Apache.
Fill the Cache
- Mapproxy seeding is easy
- Seeding both maps to level 10 takes time
- A lot of files
Warning
If you get a disk full message when df -h shows you have lots of space left try df -ih.
It is very easy to run out of inodes default is 15M - you need at least 16M
Connect QGIS
Set up WFS-T
- in PostGIS create a new database
- create a new table (either PostGIS or QGIS)
- create store and layer in GeoServer
- create a user with write permission in GeoServer
- connect QGIS to GeoServer
- Turn on editing, add new features.
Further Hardware
For full resilience add a solar panel and battery pack.
Also can add a VGA or small video screen, USB keyboard & mouse.
For bonus tracking it is easy to add a GPS chip to the Raspberry Pi and track the boat directly into the PostGIS database.
Conclusions
A Raspberry Pi:
- is capable of handling caching maps via Mapproxy
- can run GeoServer (but not fast)
- WFS-T is acceptable
- WMS is just OK for simple data
- can share the cache between machines to reduce load on external network
- can share locally changing data between local machines
Who could use this?
- Obviously pirates
- Disaster response
- Field surveys
- Any one with a need for maps and a limited internet connection.