How does the AI work?
- Our AI is “scans” laser scans of buildings and finds objects like walls, windows or pipes (premium customers only). This way, e.g., single scan points of windows are grouped to individual objects of type window.
- The user can now clean-up the point cloud automatically, e.g., by only viewing wall, windows, doors or pipes. Other stuff like furniture is removed automatically.
- In the AI output, also walls are detected which the user can check off wall-by-wall. That way, modeler and AI work together in a continuouse quality checking workflow.
Does it already exist?
Yes. Please register under https://app.aurivus.com/ to download an example project or upload your own scan to the AI. Some features are only available for premium customers.
What is the pricing model?
- We offer Pay-Per-Use. When you upload the point cloud, the price will be shown which you can pay via credit card, SEPA and many other mayment methods. The price is 0,25 EUR for roughly 1 m² and you get a discount if you buy credits for large or multiple projects. See the “Credits” section in the online portal for more.
- You find our free plug-in for Revit 2020, 2021, and 2022 in the “Downloads” section in the online portal. For other CAD programs, third-party plug-ins can be used.
- If you are interested in a annual license, please contact sales.
Is there a free trial?
The Revit plug-in is free and with pay-per-use for the AI you can test aurivus with a small example for a low price: A 100 m² scan results in roughly 25 EUR. Feel free to contact us if you are a student.
What file format is supported?
Please save all scans into a single E57 file. The format (unified, structured, unstructured, …) doesn’t matter. The maximum file size is 20 GB.
We also support RCP. Please create a .zip archive which includes the RCP file and the “support”-folder where the rcs files of that rcp project are stored.
I get the error “No data was found”.
This can happen when the uploaded point cloud is no supported file format. E.g., an .xyz file in a .zip archive.
Please only upload supported file formats. E.g., E57 or RCP.
What kind of objects can your AI understand?
- Our AI is trained for INDOOR scenes.
- Best supported classes are walls, roofs, roof trusses, doors, windows, columns. More classes, like furniture, are processed. Other classes, like pipes, are only available for premium customers.
Can I use your AI for piping?
We have a separate neural network for piping. It is currently in the beta phase and available for premium customers only.
What buildings can I use with aurivus AI?
The aurivus AI is trained to understand office spaces, apartments, and similar context. The AI is best familiar with
- Walls
- Floors and Ceiling
- Roofs and Roof Trusses (ger. Dachstühle)
- Windows and Doors
- Columns and Beams
Our AI is continuously trained to know new objects. Classes which are not yet possible are, e.g., pipes, ducts, or trees. These are coming soon and will be announced on our social media channels.
What scanners do you recommend?
Our AI can read points from a variety of different scanner types, e.g., terrestrial, handheld or mobile devices. It doesn’t matter if the point cloud contains multiple scans, a single scan, unified or structured data. We tested our AI on data from (alphabetic order) Faro, GeoSLAM, Hexagon/Leica, NavVis, and many others including the iPad Pro 12.
How much costs one credit?
Credits represent the work of our AI. It correlates with the building area and complexity. A house with 400 m² floor area results roughly in 40 credits. Scenes which are more complex for the AI cause more credits. Outdoor area, e.g., isn’t trained at the moment and usually no credits or only a few are charged for outdoor AI effort.
As soon as the exact amount of credits is shown in the project preview, you can buy the exact amount of credits per project (€2,50 per credit) or you can buy a credit block in advanced to receive a discount.
How can I load the AI file into Revit?
You find our free revit plugin in the download area at https://app.aurivus.com/ and a quick start guide here.
I’m mostly interested in other classes, e.g., pipes.
The neural network can be trained to detect other objects and object categories. It is currently trained on pipes and pipe-related parts. If you are interested in pipes or other classes, we are happy to receive your mail.
I can’t see my point cloud in Revit
If you loaded an aurivus point cloud with the aurivus plug-in for Revit, the coordinates of the scan might be too far away from the Revit document origin. In this case you can follow the instruction in this tutorial video: