Released in closed beta at the end of last year, ARCore Depth is now available in ARCore 1.18. Since the initial announcement, Google has been working with selected partners to create compelling use cases of this technology.
ARCore Depth uses a depth-from motion algorithm to creates a depth map of your environment and provide support for occlusion detection. A depth map enables determining which objects are closer to the viewer and which objects should appear as overlapped by other real-world objects. This feature had been previously available, but only on devices equipped with a depth sensor, such as Apple's TrueDepth camera.
The official release of ARCore Depth API includes minimal changes from last year's beta. Instead of adding new features, Google focused on partnering with a number of collaborators to create samples and use cases showing how depth mapping can improve AR experiences.
"We’re beginning to understand what kinds of depth capabilities are exciting for developers to build with", remarked Sam Hare, research engineering manager at Snap Inc.
Google lists a number of applications integrating ARCore Depth that have been published on Google Play. Those include Play Five Nights at Freddy’s AR: Special Delivery, which uses occlusion to allow game characters to hide behind objects, and TeamViewer Pilot, a remote assistance solution using AR annotations to let experts better understand the environment they are virtually interacting with.
Google also worked closely with Snap, Inc. to create Snapchat lenses. This led to new Snapchat Lens templates which can be used by lens developers. In particular, the Depth Toybox template uses depth data to estimate surface orientation and occluded objects; the Depth material template provides examples of using depth in conjunction with materials; and the Underwater depth template uses depth to create an underwater scene.
In addition to this, Google open sourced two GitHub repositories: Lines of Play and ARCore Depth Lab. Lines of Play uses ARCore Depth to integrate virtual dominoes with the real world. Specifically, besides using occlusion, this sample app shows how to use a depth map to provide more realistic collision detection.
ARCore Depth Lab focuses isntead on how depth maps can be used to provide advanced realistic effects besides occlusion, such as character locomotion on uneven terrain, laser beam reflections, rain and snow particle collision, pain splat, AR fog, surface retexturing, and many others.