Aerial image dataset automatically maps rooftop solar arrays

Scientists at Mines Paris-PSL University in France have created a dataset of aerial images, segmentation masks, and installation metadata for rooftop PV systems. They conceived the dataset to set up installation registries by extracting small-scale PV metadata from overhead imagery.

“Our dataset provides ground truth installation masks for 13303 images from Google Earth and 7686 images from the French national institute of geographical and forestry information (IGN),” the researchers said, noting that the metadata includes installed power, surface, tilt, and azimuth angles.  “To address architectural differences, researchers can either use the coarse-grained location included in our dataset or use our dataset in conjunction with other training datasets that mapped different areas.”

The dataset provides thumbnails with a resolution of 400 × 400 pixels centered around the locations of PV systems. The thumbnails are based on the geolocation of a PV database for deep learning known as Base de données apprentissage profond PV (BDAPPV).

The user initially performs image classification by clicking on an image if it depicts a PV system and then annotators delineate the PV panels on the images.

“Once we generate our PV panels polygons, we match them with the installations’ metadata reported in the BDPV dataset.” the French researchers said, adding that this step involves internal consistency, unique matching, and external consistency.

The file containing PV installations’ metadata and the notebooks used to generate the masks are available in two separate datasets. The dataset provides installation metadata for more than 28,000 installations and segmentation masks for 13,000 installations.

“Dataset applications include end-to-end PV registry construction, robust PV installations mapping, and analysis of crowdsourced datasets,” the scientists said.

They described the dataset in “A crowdsourced dataset of aerial images with annotated solar photovoltaic arrays and installation metadata,” which was recently published in Scientific Data.

“To the best of our knowledge, it is the first time a training dataset contains PV panel images, ground truth labels, and installation metadata,” they concluded.