AMS Hidden Amsterdam

This project supports the Hidden Amsterdam Festival, a 3-day festival, to be held in 2017, that aims at letting citizens and tourists discover “hidden” parts of the city. Those areas are often not visible from the main streets and shopping districts. To be able to discover these areas, a “3D map” is necessary. Unfortunately, at this moment, such 3D maps are built manually. The aim of the project is therefore to investigate methods to automatically construct such maps, and to investigate how socio-economical data based on mobile technologies can be used to improve the map.

The reconstruction of a 3D part of a city can be performed automatically by combining different datasets, for instance 2D geographical datasets and using the elevation points obtained by laser-scanning. However, there are several unknown:
– How to deal with conflicting 2D datasets?
– How to deal with conflicts between 2D datasets and the elevation dataset?
– The quality of the models obtained depends greatly on the quality of the elevation dataset, especially where the helicopter/airplane that flew over Amsterdam to collect the data was. Several gaps and missing information will thus be created.

We plan to use a mobile surveying device (the Google Tango) to quickly scan small occluded areas and objects to obtain more information. We also plan to investigate methods to automatically infer the interior of buildings (apartments within buildings) by combining different datasets that are publicly available.

Another objective of the project is to develop a method for the targeted collection and analysis of social media streams. Spatial and temporal patterns in these streams, combined with geographical and statistical data will be exploited to provide the socio-economic context (eg accessibility, types of facilities, amount and types of people around during specific periods of the day). Patterns of movement, combined with socio-economic characteristics of specific locations can provide more detailed insights into the real potential and the various viable functions that could be given to specific. Moreover, it can be used as a monitoring and evaluation tool over time and indicate if new functions of hidden spaces cause different (desirable or undesirable) patterns of movement and behaviour.

Scientifically speaking, the two main results will be
1. a methodology to reconstruct the 3D map of Amsterdam.
2. a template to “measure” the state of (hidden) areas by analyzing movement patterns and location specific characteristics through the use of social media streams and socio-economic statistics and maps.

For the festival, the 3D model of one part of the city centre will be delivered. This model will allow us to gain insight into the hidden places, and it will be used in the smartphone app that will be developed.

Project duration: 1 February 2016 – 31 December 2016

Partners: TU Delft, Wageningen UR, Stichting Hidden Amsterdam

Project leader / contact:
Dr Hugo Ledoux, assistant-professor in 3D geoinformation
TU Delft
Phone: +31 (0)15 2786 114