Page 156 - Proceedings of the 2018 ITU Kaleidoscope

UNDECLARED CONSTRUCTIONS: A GOVERNMENT’S SUPPORT DEEP LEARNING
SOLUTION FOR AUTOMATIC CHANGE DETECTION

Pamela Ferrari Lezaun; Gustavo Olivieri

Department of Systems Information Engineering, Universidad Tecnológica Nacional FRSF, Santa Fe, Argentina

ABSTRACT Bearing in mind that with all the aforementioned, much less
tax revenue will be obtained, thus, representing great
In our cities, in particular those with high demographic economic losses for the governments of the cities. Due to
density, the proliferation of buildings goes so fast that it is the nature of the taxes, in one way or another will end up
not possible or -in the best scenarios- very difficult to be affecting the citizens.
handled by the government departments regulating the
legitimacy -in term of safety and taxes- of those
constructions.
In this paper, we propose a deep learning tool for computer
vision trained with a corpus of satellite images provided by
SpaceNet to detect changes in cities, showing the most
recent constructions automatically, which allows different
municipal officers to check if they have been -or haven’t
been- declared. To achieve this, we implemented the
layered architecture of the SpaceNet Challenge Round 2
winning solution, and decided to improve it with an output
comparison which gives us a high value final result for the
end-user in the detection of changes, giving him the
possibility to appreciate in the graphic user interface how
many new buildings and square meters were detected.

Keywords – building detection, undeclared constructions,
illegal construction, Machine learning, computer vision
Figure 1 - Statistics of illegal constructions in the province
1. INTRODUCTION of Buenos Aires.

In Argentina, millions of undeclared square meters are built In the same way, the lack of regulation in building
every year. Just in the province of Buenos Aires, 14 million constructions would lead to the non-detection of structural
square meters were detected in the last few years without problems that may end in great tragedies such as landslides.
registration [1]. As can be seen in Figure 1, just in 2018
half million square meters with irregularities were detected The number of provinces in Argentina -and many regions
in that province [2]. Many of these infractions involve around the world- facing the same problem are rising
parcels where industries and active businesses operate but constantly. The problem has been mentioned in different
are still registered as vacant parcels.
academic articles [6, 7, 8], allowing to determine that it is
present in many places around the world, all of them with a
In 2017, The Direction of land registry office of Plottier in point in common: the complexity to detect illegal square
Neuquén Province indicated that estimated clandestine meters is given by the lack of a tool that can assist different
constructions hovered 250.000 square meters and in the municipal or provincial officers in the detection of
capital of that province about half on that number [3]. This undeclared constructions, in process or already finished.
situation could also be reflected, to a different extent, in
many countries around the world. In Hong Kong, for It should be noted that in some countries, the government is
example, 25% constructions have illegal structures [4]. investing and implementing projects of this magnitude, as is
Another case is in Bulgaria, where the illegal buildings in the case of Spain through the public company of
preserved areas are a threat to the biological diversity [5].
cartography of Canarias (GRAFCAN), who on June 1 of
2017 released a statement in its official page that says ―The

978-92-61-26921-0/CFP1868P-ART @ 2018 ITU – 140 – Kaleidoscope