Executive Summary
Key Takeaways
- Aurora Solar is a solar software company aiming to transform the process of designing, selling, and delivering solar projects.
- One of Aurora’s products, Aurora AI, leverages computer vision and machine learning algorithms to generate 3D roof models, automating the initial site modeling work.
- By comparing the roof models generated by Aurora AI’s algorithm with a dataset of human-produced models of the same roof, Aurora conducted a study aimed to test how accurate Aurora AI was.
- Aurora found that Aurora AI could produce models of equivalent or higher accuracy than human-produced models.
About Aurora
Aurora Solar is a solar software company working to make solar energy accessible for all by streamlining the process of designing, selling, and delivering solar projects. Using our years of experience and real-world data, Aurora has developed Aurora AI, a proprietary machine learning engine leveraging computer vision algorithms. Aurora AI enables solar businesses to move faster by automating the initial site modeling work, creating AI-generated, reliable 3D roof models, including obstructions, in about 30 seconds using satellite imagery, LIDAR data, and insights from over half a million site models built by Aurora’s Expert Design Service.
Study Overview
In a study conducted by Aurora to assess the accuracy of Aurora AI, Aurora ran a dataset of 4,000 residential homes across the United States through Aurora AI and compared them to human-produced models of the same homes in Aurora. Overall, Aurora found that for all three metrics, Aurora AI’s performance was essentially tied, if not slightly better, than human performance. Overall, Aurora’s AI accuracy is not only effectively serving its purposes right now, but will only improve from here.