Occlusion

ARKit provides support for occlusion based on depth images it generates every frame.

There are three types of depth images that ARKit exposes through the provider's implementation of the XROcclusionSubsystem implementation:

  • Environment depth: distance from the device to any part of the environment in the camera field of view.
  • Human depth: distance from the device to any part of a human recognized within the camera field of view.
  • Human stencil: value that designates, for each pixel, whether that pixel is part of a recognized human.

Optional feature support

ARKit implements the following optional features of AR Foundation's XROcclusionSubsystem. The availability of features depends on device hardware and software. Refer to Requirements for more information.

Feature Descriptor Property Supported
Environment Depth Image environmentDepthImageSupported Yes
Environment Depth Confidence Image environmentDepthConfidenceImageSupported Yes
Environment Depth Temporal Smoothing environmentDepthImageSupported Yes
Human Segmentation Stencil Image humanSegmentationStencilImageSupported Yes
Human Segmentation Depth Image humanSegmentationDepthImageSupported Yes
Note

Refer to AR Foundation Occlusion platform support for more information on the optional features of the occlusion subsystem.

Environment Depth

The occlusion subsystem provides access to two types of environment depth: raw and smoothed. These correspond to the following ARKit APIs:

Note

You must enable smoothed depth by setting environmentDepthTemporalSmoothingRequested to true. Otherwise, TryAcquireSmoothedEnvironmentDepthCpuImage will return false.

Requirements

Environment depth requires Xcode 12 or later, and it only works on iOS 14 devices with the LiDAR scanner, such as the new iPad Pro.

Human depth and human scencils requires Xcode 11 or later, and it only works on iOS 13+ devices with the A12 Bionic or higher.

Apple and ARKit are trademarks of Apple Inc., registered in the U.S. and other countries and regions.