Artificial intelligence shows where the sun is shining

In a new innovation project, Aither and NORCE will develop an interactive sun diagram from an image to illustrate how the sun casts light and shadows at different times of the day and year. The purpose is to provide a visual understanding of the course of the sun and the potential for the utilization of solar energy.

The technology is useful for private individuals, but also for builders and architects when planning residential areas or planning where to place solar panels. A larger number of private individuals and organizations can then be encouraged to invest in solar energy, which helps to reduce greenhouse gas emissions. Renewable energy companies can use the solution to identify potential locations for solar power plants.

Innovation in image processing

– I will develop new algorithms for image processing. The purpose is to use artificial intelligence to automatically detect and remove shadows from a 2D image and create a 3D image. The shades can then be applied depending on where the home or building is placed and you can see what will be the optimal location in terms of solar conditions for the home or for solar panels on the roof, says Xuan Zhang, Senior Researcher at NORCE.

Imagine, for example, that you have bought a plot of land and are going to build a house on the plot, but you are unsure exactly where on the plot you should build. With the new technology, you will be able to see on the image of the plot how the sun conditions will be in different places so that you can more easily choose the best location for your house.

There are already algorithms for this, but they are very simple. The new solution has an advanced shadow removal algorithm that can interpret images with several objects, for example, images that contain houses and buildings within natural surroundings.

NORCE has specialist expertise in these areas and will work closely with Aither to design the right solutions.

– We will create a tool that is easy to use and accessible to everyone. Many existing solar courses and solar energy calculation tools require technical or scientific knowledge. An automated process is crucial for the solution to be so fast and affordable that everyone can use it, says Karsten Ersland, in Aither.

Advanced techniques

First, the researcher must remove shadows from the original image. This involves developing artificial intelligence that can identify and remove shadows from an image in an accurate and efficient manner, which places great demands on advanced techniques in image processing and machine learning.

She will then convert the image from 2D to 3D, which provides the basis for the interactive solar diagram. This involves developing techniques to gain depth in the 2D image, a task that in turn requires advanced techniques in artificial intelligence, especially in machine learning and data processing.

– The project can lead to increased use of solar energy, reduced energy consumption, and reduction of CO2 emissions, in line with the UN's sustainability goals, says Ersland.

Today, there are no good tools that show where the panels should be placed. Incorrect placement of the solar panel can lead to little energy gain. The EU also comes with requirements for solar panels on both large and small buildings.

Research and policy development

By offering a tool that can visualize the course of the sun and calculate solar energy potential based on images, the project can support researchers in climate science, renewable energy, urban planning, and architecture in their work. The tool can help collect and analyze large amounts of data on solar processes and solar energy potential, which can provide new insights and inform future research and policy development.

This innovation project has been developed in the Agder region, but the technologies and methods that are developed can be used both nationally and internationally. The project is based on results from a preliminary project financed by Agder County Council through Research Mobilization Agder.

– This project is exciting and fully in line with our intention, says Bjarne Sverkeli, project manager for Research Mobilization Agder.

– In Agder, we have a very strong research environment, and especially in the area of ​​AI. When, through our policy apparatus, we can connect an innovative startup like Aither with such a strong research environment as NORCE, we have created a connection with a lot of power in it. There is great power in the link between innovator and researcher, and it is this power that we work with to trigger, says Sverkeli.

In the video below you can see how the solar diagram can be used.