Rooftop photovoltaic panel detection

How to identify solar panels on rooftops from satellite imagery?

In this post, we discuss how you can identify solar panels on rooftops from satellite imagery using Amazon Rekognition Custom Labels. High-resolution satellite imagery of urban areas provides an aerial view of rooftops. You can use these images to identify solar panel installations.

Which Visualization Library is used for rooftop photovoltaics?

The library for visualization is matplotlib. The project target is to segment in aerial images of Switzerland (Geneva) the area available for the installation of rooftop photovoltaics (PV) panels, namely the area we have on roofs after excluding chimneys, windows, existing PV installations and other so-called ‘superstructures’.

Can U-nets be used to segment roof-top PV panels in satellite images?

In 2020, Zhuang et al. proposed a cross-learning driven U-Net (CrossNets) method to segment roof-top PV panels in satellite images. However, the above studies focused on using the universal machine learning frameworks such as CNN, U-Net, DeepLabv3 and etc., lacking analyzing the characteristics of PV image data and improving the models.

How accurate is rooftop PV estimation based on Geographic Information?

Table 1. Previous studies for rooftop PV estimation based on geographic information data. Area Estimation: Precision: 93.1% for residential areas and 90.5% in non-residential areasRecall: 88.5% in residential areas and 90.5% in non-residential areas Count Estimation: MeanRelativeError: 3.0% for residential areas and 2.1% for non-residential areas.

How can machine learning improve the knowledge of rooftop PV installations?

Small-scale, rooftop PV installations are deployed at an unprecedented pace, and their safe integration into the grid requires up-to-date, high-quality information. Overhead imagery is increasingly being used to improve the knowledge of rooftop PV installations with machine learning models capable of automatically mapping these installations.

How can we detect rooftop area at pixel level?

We used a Convolutional Neural Network (CNN) model based on U-net and an adaptive learning algorithm to train it. Iou and Acurrancy are computed to evaluate the performances. We are able to automatically detect in test images the available rooftop area at pixel level with performances comparable the state-of-the-art.