Behind the scenes
Melvin the Mini Machine uses a smartphone and bespoke code and software to determine its location, write messages and recognize the people around him. Like the other parts of the machine, most of it is purpose-built and other parts are adaptations of (open source) software.
About the smartphone
About the smartphone
The phone we use is a HTC Desire running Android 2.3.3 which was given to us by
Blue Mango Interactive. We use an app custom built by Michael Schifferling that takes a picture and sends it to a webserver, where an image URL and GPS data are added to a MySQL database. Still with us?
This is where it gets technical
We used Visual Studio to build a program that processes and analyzes the photos taken by Melvin. For glyph recognition we worked with
GRATF from AForge.net, expanding the framework to work for an undefined number of glyphs, and replacing markers with words.
Meanwhile, on the website...
Once a picture is uploaded after it’s been processed, the data and the picture are published to
Facebook and
Twitter using the API from both platforms.
Melvin’s travels page runs the Google Maps API with a custom layout. And last but not least, the GPS data link uploaded by the phone to the markers on the map is a custom jQuery script.