Automated Planning of Rooftop PV Systems with Aerial Image Processing

• Published in: 2018 IEEE PES Asia-Pacific Power and Energy Engineering Conference (APPEEC) pp. 736 - 740
• Main Authors: dos Santos, Francis Carlos, Thornburg, Jesse, Ustun, Taha Selim
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.10.2018
• Subjects: Buildings; demand forecasting; Graphical user interfaces; Microgrids; object detection; photovoltaic systems; Planning; Renewable energy sources; Satellites; Software; solar panels
• ISBN: 9781538656853; 153865685X
• DOI: 10.1109/APPEEC.2018.8566666

The increasing prevalence of photovoltaic (PV) panels for microgrid and off-grid energy applications makes affordable PV planning an important issue. Since recording rooftop area and dimensions traditionally required on-site measurements, the process was expensive, slow, and hard to scale. This research develops software that uses image processing for roof detection. Satellite images feed into the software, which estimates the rooftop area receiving solar exposure in that area as well as the number of individual buildings receiving solar exposure. In this way, entire villages can be analyzed automatically, and PV installations planned from afar, rather than requiring a human taking measurements of each building from the ground. This research further develops a GUI to accomplish this rooftop classification for users around the globe, making this capability available even to parties with low resources who would benefit from access to electricity. In this way, the study makes planning PV systems feasible and affordable for many scales of installation, from a single home to a city of numerous assorted buildings.