A Web-based platform, Google's Project Sunroof is rolling out in three U.S. cities and regions. Below, well explain how we used deep neural networks to obtain accurate three-dimensional models of residential roofs. ConclusionSunPower empowers homeowners to understand the amount of energy they can generate with solar, now with just a few clicks. The data: guiding the design with both color and depth imageryIt is probably possible to design a three-dimensional model of a roof with satellite imagery alone, but design accuracy improves greatly with the use of a height map. Our team was able to start work on this exciting project due to advances in aerial imagery and machine learning. First, we model the roof in three dimensions to account for obstructions such as chimneys and vents. Specifically, we use deep learning and high-resolution imagery as inputs to models that design and visualize solar power systems on residential roofs. Florida Polytechnic University, Designed Is Architecture Fueling this Pandemic, and IKEA Launches Augmented Reality Application, Next Progressives: IUA Ignacio Urquiza Arquitectos, Meet the Winners of the 2022 AIA Election, Harvard GSD Introduces Its 2023 Loeb Fellowship Class, Harvard GSD Announces 2022 Wheelwright Prize Shortlist, AIA Publishes Justice Supplement to Guides for Equitable Practice, IIDA Expands 'Design Your World' Education Program to Miami, WoodWorks Names 2022 U.S. Wood Design Award Winners, Wan Bridge Secures Site for Newest Build-to-Rent Community in Greater Houston, NKBAs Residential Kitchen and Bath Market Outlook Update Forecasts Double-Digit Increase Over 2021, Existing-Home Sales Slide for Fifth Consecutive Month in June, Our Competitor Is the Status Quo: How Agorus Plans to Transform the Construction Industry, Residential Construction Starts Decline Another 2% in June, Inventory in Overvalued Housing Markets Is Rising, Decoding the Framing Square for Rafter Layout, New Tool: Metabo HPT MultiVolt Cordless Compressor, 390,000 Nonfarm Payroll Jobs Added in June. Announced yesterday, the company'sProject Sunroof Speed and scale via Cloud AI PlatformOnce we had built a satisfactory proof of concept, we quickly realized that we would need to iterate on our model in order to deliver an experience that was ready for homeowners. For Instant Design, we partnered with Google Project Sunroof for access to both satellite and digital surface model (DSM) data. It is currently available in the greater Boston area, Washington, D.C., Washington County, Ore., and Lo Barnachea and Vitacura, both in Chile. encounter who think that 'my roof isnt sunny enough for solar,' or 'solar is It also offers information on purchasing or leasing the panels, as well as taking out a loan to cover the installation expenses, and the projected payback period in energy-cost savings. For our prediction, we used NVIDIA V100 GPU-enabled virtual machines on GCP with nvidia-docker, which helped us achieve prediction times of around one second. Architect Magazine: Architectural Design |ArchitectureOnline: The premier site for architecture industry news & building resources for architects and architecture industry professionals. Roof segmentationTo reconstruct a roof, we model each roof segment with its corresponding pitch and azimuth in three dimensions. combines aerial 3D models from Google Maps, historical weather data, the cost Our main challenge here was that we had to handle both the quantity and size of the obstructions, and address any imbalance in class representation. Have you ever wondered what solar panels would look like on your roof? Copyright 2022 Zonda Media, a Delaware corporation. She describes how SunPower uses AI Platform to provide users with useful models and proposals of solar panel layouts for their home, with only a street address for user input. While we have more work to do, we are optimistic that SunPower Instant Design will transform the solar industry when our first product featuring this technology launches this summer. We used our database of manually generated designs as a base for our labeled data, and projected those onto the RGB and depth channels for the training, validation, and test sets. Indeed, there are more roof pixels than obstruction pixels in our images. The Coming Renewable Energy Revolution in the Middle East, USGBCLA Reveals Lineup for 21st Annual My Green Building Conference, Why Standby Power is Imperative for Residential Healthcare, This Week in Tech: A Floating Park Made Entirely of Recycled Plastic Waste, The Acceleration of Building Sensors Networks. Google knows about usiswhether our homes and businesses could benefit from the addition of solar panels. With SunPower Instant Design, homeowners can create their own designs in seconds, which improves their buying experience, reduces barriers to going solar, and ultimately increases solar adoption. To learn more about how SunPower is using the cloud, read this blog post from Google Cloud CEO Thomas Kurian. For instance, we added batch normalization to each convolutional layer for regularization and selected the Wide Residual Network as our encoder for improved accuracy. We also generated augmentationsincluding rotation and translationto reduce overfitting. What gets in the way: chimneys, vents, pipes, and skylightsIn an effort to avoid mistakenly placing panels on obstructions such as chimneys, vents, pipes, skylights, and previously-installed panels, our next step is to detect those obstructions as separate items on the roof. There are no quick answers because every roof is different and every house requires a customized design. Follow her on Twitter at@HallieBusta. After some experimentation, we chose to perform semantic segmentation, and then selected a version of a U-net that works well with our type of imagery at high speeds. Mithun, Waggonner & Ball Architecture/Environment. The journal of the American Institute of Architects. The app, in turn, shares a recommended solar installation size in square feet and kilowatts. For now, however, we'll have to wait and see where the app goes next. Moreover, we took full advantage of the ability to train multiple models simultaneously to conduct a vast hyperparameter search. Adding to the growing list of things She holds a bachelor's degree in journalism from Northwestern University's Medill school and a LEED Green Associate credential. We needed to build a development pipeline that could quickly bring modeling ideas from conception to deployment, so we chose AI Platform to help us scale. In contrast, on AI Platform, we were able to train and test a new model in a single day. just too expensive. Certainly The tech giant isn't the first to put this type of information into consumers' hands. Finally, the app connects homeowners with local installers to do the workalong with, we hope, a tried-and-true site evaluationof getting the PVs on the roof. of utilities, and the value of local incentives to gauge whether covering a rooftop with photovoltaics (PVs) would result in energy-cost savings. Hallie Busta is an associate editor of products and technology atARCHITECT. The online app is the project of Google engineer Carl Elkin, who previously volunteered with Solarize Massachusetts, a Boston-based solar-adoption program targeting residences and businesses. Yetthe scale of Google's resources gives Project Sunroof a leg up in reaching the masses. What Happens After the Work is Completed? We began to identify roof segments by applying image processing and edge detection on both the satellite and depth data, but we quickly realized that semantic segmentation would yield much better results, as similar edges were detected successfully with that method in research literature. Due to the difference in shape and scale of chosen classes we decided to use a separate model from the segmentation model to detect obstructions, although both models are similar in structure. The U-net architecture was a solid starting point, with a few tweaks for better results. In ablog postannouncing the project, Elkin explained why he developed the tool:Ive always been surprised at how many people I Read on to learn how and why we built this technology for our customers, called SunPower Instant Design. Editors Note: Todays post comes from Nour Daouk, Product Manager at SunPower. And AI Platform helped us focus on the core design problem, achieve our goals faster, and create designs quickly. We are changing how we offer solar power to homeowners by giving them immediate answers to their questions. A screenshot of Project Sunroof's analysis of Google's San Francisco offices. Second, we lay legally-mandated access walkways and place solar panels on the roof segments. Homeowners typically spend a significant amount of time online researching solar panels and running calculations to understand their potential savings and the number of panels they need for their home. Mapdwell, a project from MIT's Sustainable Design Lab, for example, launched in 2013 to provide homeowners and businesses with the costs and benefits of adding PVs to a building. many of them are missing out on a chance to save money and be green.. Google is rolling out the platform as a consumer tool in Boston, the San Francisco Bay area, and Fresno, Calif. Users plug in their address and how much they typically spend on electricity. At SunPower, were helping homeowners create solar designs from the comfort of their home. Our initial training setup was on our own servers, and the training process was slow: training a new model took a week. With Instant Design, we replicate this same process by leveraging tools including machine learning and optimization. Hallie Busta is a former associate editor of products and technology at ARCHITECT, Architectural Lighting, and Residential Architect. How we helpDesigning a solar power system for a home is a process that relies on factors unique to each home. Start building on Google Cloud with $300 in free credits and 20+ always free products. We also created a domain-specific loss function to get the model to converge to meaningful outcomes. Finally, we model the angle and exposure of sunlight hitting the roof to calculate the systems potential energy production. Previously, she wrote about building-material sales and distribution at Hanley Wood.