Information Technology Reference
In-Depth Information
Fig. 4.6
Detecting and
cleaning environmental
threats:
1
detection,
2
cleaning,
3
overlaying of the
environmental threat
The eVision virtual character, Snowkin, invites users to participate in a game
activity that consists in using their mobile device to detect and clean environmental
threats in their vicinity, such as cars, airplanes and factories. Detection is automat-
ically achieved by eVision as the user moves his mobile phone and points around.
Once an environmental threat is detected, the player should clean it by using his
finger to rub the mobile phone display over the detected threat (Fig.
4.6
), which will
then be automatically overlaid with pro-environmental objects (e.g., overlaying a car
with a bicycle).
When completing each activity, users are awarded with points and green leaves
(eVision's virtual currency), as well as with environmental information regarding
the corresponding threat. eVision's green leaves can then be spent to buy items, in
the in-game store, to customize the Snowkin. Positive reinforcement techniques are
also used to keep the user engaged. This aspect is very important and it is achieved
with the assistance of Snowkin by establishing motivating and pro-environmental
dialogues every time the users clean a threat.
eVision was implemented using Apple's Xcode IDE (Integrated Development
Environment) in Objective-C language. The image processing module was developed
using the OpenCV library compiled for iOS 5.1. For the Facebook integration it was
used the proper Facebook mobile API designed for iOS with minor refinements.
eVision current implementation allows users to detect different pollution sources:
cars, airplanes and factories. Knowing the static location of factories, GPS was
used to identify factories in the users surroundings. Image detection methods were
investigated in order to detect the dynamic position of cars and airplanes during
eVision's game activity. Using the OpenCV library (
2013
), the airplane detection
method was designed based on contour detection and analysis. It worked quite well
on a clean sky. For vehicle detection, recognition of cars' registration plates was used.
These methods for recognition of dynamic threats still need optimization, however
given the eVision's modular software architecture, it is easy to replace the current
image recognition algorithms by more precise ones.
Figure
4.7
a illustrates the eVision's main menu interface, where the users can
navigate through all the features of the application. The Scanner mode is the eVision's
main feature, leading the users to the core game activity, where they can inspect their
surroundings and gather information regarding environmental threats. Snowkin is
always present, guiding the users through the game, encouraging them to play and
to help the environment, as well as congratulating the users for their achievements.
