Data Owner: Jia Ying Lin, Costin Ograda-Bratu and S. Keshav
Data Description: There are many anomalies that may affect a solar panel’s power production, including cloudiness, snow, dust, and shadows. Images of the solar panels are taken as input to detect snow, since it can be easily identified through images. This dataset contains a set of measurements from a solar installation on top of the ERC building at the University of Waterloo, and some photos of the system which could allow image recognition to identify snow on the panels.
Data Size: The dataset is 172.2 MB in total.
Data Field Description: In the Images folder, there are three folders: no_snow, partial, and all_snow containing images of the solar panels with no snow on it, partially covered by snow, and completely covered in snow respectively. Every jpg file is named with the timestamp that the image is taken. Images are taken at January 26th, 27th, 28th, February 2nd, 3rd, 28th, and March 1st, 2nd, 21st, 22nd, 23rd, 24th, 25th of 2019.
In the Power_Measurements folder, each file is named after the date when the measurements are taken. Each entry consists of the timestamp, battery voltage and charging current. Power measurements are collected from February 27th 2019 to July 22nd 2019.
Funding: Cisco Systems and the Natural Science and Engineering Research Council of Canada (NSERC).
