At its I/O 2022 development conference, Google is presenting two Google Maps innovations, the ubiquitous display and the geospatial API, which could prove to be important components of the future of augmented reality.
Google Global View is an approximation of a visual digital twin of the world, or a further development of the 3D view of major cities or well-known landmarks from Google Earth VR.
Neural Rendering transforms 2D images into a 3D perspective
The 3D perspective of the panoramic view is generated by the computer. Google uses neural rendering to combine 2D satellite imagery and Street View images into 3D scenes. In the already impressive 3D view, Google integrates additional information that is visible in real time, such as traffic, visitor numbers or the weather.
Thanks to AI-powered rendering, Google can also create 3D rooms from 2D images of restaurants, for example. Dive into individual streets with the panoramic view and keep navigating inside – all in 3D.
According to Google, the new offering runs on every smartphone via an “immersive stream” from Google Cloud. There’s one small advantage with this amazing technology: It’s available now, but initially only for select areas of San Francisco, New York, Los Angeles, London and Tokyo. Other cities follow.
Will Google Maps become the next big digital infrastructure?
Immersive Stream shows that the mapping platform has long since surpassed its status as a navigation app. Google is now mining the real world as intensely as the Internet
As a result, maps are increasingly becoming a digital picture of the real world, also known as “cloud AR” in the jargon of the augmented reality industry. This is amazing 3D coordinate system in the real world It is a building block for a common AR future.
Anyone who has mastered cloud AR, for example, can determine – and pay for it – when and where digital audiovisual information appears in real life. Meta is also looking for such a digital twin of the world, but thanks to Maps it is likely that Google is ahead.
With the ARCore Geospatial API, Google brings another step in map development for cloud AR at I/O: developers can use the programming interface to place digital content at real locations in 87 countries — without having to visit or scan that real location. Digital games or exhibits that appear in fixed locations are in fact possible.
According to Google, the experience gained from developing Google Maps for about 15 years is in the new interface. The AI-powered VPS visual positioning method, which was previously used in the Google Maps app for augmented reality navigation stocks, among other things, aims to help locate digital content in real life as accurately as possible. VPS in turn relies on images obtained from Street View along with artificial intelligence image analysis.
Geospatial API is a further development of Cloud Anchors, which was first introduced in 2019. Examples of applications and code information for the new interface are found in Google’s dev blog.
At I/O, Google also introduced new technology glasses for real-time visual translations and made a clear commitment to AR glasses as a new computing platform.