Skip to content

AR Cloud

rokru edited this page May 20, 2019 · 1 revision

AR Cloud

A 3D digital copy (representation) of the physical world that enables persistent overlays of augmented reality experiences that can be shared with many users and devices.

A one or multi-tiered layer of digital constructs that can exist as many composite bits and bytes of information architecture. By linking this data to the real world spaces, and places we can accurately represent the world and objects around us in digital formats desired by the system designer.

From VRARA whitepaper :

The AR Cloud is a continuously updated collection of machine-readable datasets, primarily sparse or dense point clouds plus a feature descriptor and other meta-data for each point or groups of points. AR Cloud data is any dataset which can aid in accurately determining the position and pose of AR-enabled devices, semantically understanding the scene and positioning digital content in physical spaces (this can include polygonal geometry, visual descriptors, images, video, ‘simple geographic features’, etc). Because of the ever-changing nature of real-world spaces, the AR Cloud must constantly be updated with new data to enable AR devices to accurately localize within these dynamically changing environments. AR Cloud data is crowdsourced from an array of technologies (sparse slam maps, point clouds, occlusion meshes, photorealistic textures, BIM data) and versioned to reflect multiple states including time of day, environment changes, weather or lighting changes.

Clone this wiki locally