This thesis focused on the development of a mobile application that implements augmented reality (AR) technologies to visualize solar panel installations on real-world buildings.
Mobile, Unity, Vuforia, Image Detection, C#
The primary objective of this thesis was to create a proof-of-concept application that integrates AR frameworks with image recognition to accurately place virtual solar panels on building roofs and facades. The application aims to enhance the customer experience by allowing solar installers to present a realistic AR visualization of solar installations on-site. Key challenges addressed include extracting relevant information from camera feeds, accurately positioning 3D models, and creating an intuitive user flow and experience that does not interfere with the immersion of the scene.
Solextron AG offers online tools with which solar installers can create scenes of houses with optimized solar installations. In collaboration with the FHNW, Solextron wishes to examine the viability of displaying solar panels to customers using an augmented reality experience. Displaying solar installations in AR helps customers gain insight into how the installation will look in real life.

The proof-of-concept developed uses image-based object recognition using
Vuforia, along with manual controls, to place virtual solar panels on the
roofs and facades of real buildings in augmented reality. It also
has a logically intuitive user flow which does not interfere with
the displayed scene and user experience.
The application highlights limitations in using Vuforia image targets to
automatically place virtual content. The accuracy of the image target's
placement is limited by the need for precise measurements of the building's
dimensions. This is needed to ensure correct model placement and must
currently be manually adjusted by the user. Furthermore, the obligation to
turn off Vuforia's image target behavior when trying to view the solar panels
from different angles, forces the use of AR Anchors. These AR Anchors are
resource intensive to use, and it is probable that the tested hardware in
this study could not facilitate the use of AR Anchors, causing the displayed
objects to drift. This drift compromises the immersion and use of the
application. Finally, an unintended behavior created by using Vuforia image
targets is causing tracked objects to be momentarily displaced and possibly
compromising the immersion of the scene which can negatively impact the user
experience.
Video Demo
Project duration: 19.02.2024 - 16.08.2024
Effort per person: 360 hours
Team size: 2
Solextron AG
Alte Spinnerei 9,
5210 Windisch, Switzerland
design.solextron.com
Jarno Degano jarno.degano@students.fhnw.ch
Kyle Egli kyle.egli@students.fhnw.ch
Advisors
Prof. Hilko Cords hilko.cords@fhnw.ch
Cloe Hüsser cloe.huesser@fhnw.ch