The technology behind FRAMENCE

The groundbreaking FRAMENCE technology turns ordinary photos into 3D-models that act as three dimensional photorealistic digital twins across the assets’ entire life cycle.

Technologie

For elevating the 2D-photos onto the third dimension, an AI-powered mathematic approach is necessary. For each and every photo, these algorithms determine the precise location, direction, and distance from the camera when the photo was taken. All these calculations are of course completely automated. All the user has to do is upload the photos onto the server where the model is located, and the software does the rest.

Any type of graphical material can be part of these 3D-calculations, a process we call adjustment. Typical sources are pictures taken with mobile phones, pictures or videos captured by drones and of course high-resolution panoramic images. The quality and precision of the corresponding 3D-model is determined solely by the resolution of the photo. A decent image resolution, that any off-the-shelf digital camera will yield, leads to an accuracy that rivals laser scanning. By adding additional photos containing more details, the accuracy can be increased further.

The 3D-photorealistic digital twin contains a hybrid data structure. Therefore it can be combined with any existing 2D-plans like floorplans as well as 3D-surface and volumetric models, such as BIM. Further, floorplans, 3D-surface and volumetric models can be derived directly from the photos themselves and are exported as IFC. We have nicknamed this function “Photo-to-BIM”.

An integral part of this hybrid data structure are the Points-of-Information (PoI). A PoI is a real 3D-coordinate that has any type of information attached to it. Using these PoI, any dataset from external databases, cloud-systems or 3rd party software may be accesses and displayed. Also, the PoI can represent dynamic information and thereby react to changes or updates. The PoI are located in the same 3D-navigation system as the images but are independent from them. They only happen to be in the same geometric location. This leads to the fact that the PoI can be displayed in all photos geometrically located nearby. Further, the PoI will persist even if the photos are replaced due to changes in the environment. As mentioned, the PoI are generated from a photographic representation of the reality. This allows for the PoI to also be superimposed onto the reality – thereby they can form the input for augmented reality use cases. Nowadays it is common practice to document changes or alteration made to buildings or assets by simply taking a photo using your mobile phone. FRAMENCE is capable of including those photos as well. They are localized automatically within the images in the right location at the right point in time. Every type of graphical information is always located on a timeline. If a photo needs to be replaced because of changes in the realty, it is terminated, not deleted. Therefore, one can simply go back in time using the timeline or a list of images and restore any previous situation.

The model may be accessed using any standard web browser. All information and data are now available from anywhere in the world. To ensure usability and to cater for any bandwidth, we have developed methods for effectively loading and displaying even large amounts of images and 3D-data without any loss in quality.