ARCore SDK Preview 2

2018-01-30 EDIT: Added link to ARCore supported devices.

This is a developer preview SDK that enables prototyping AR experiences using ARCore on supported Android devices.

ARCore enables AR applications to track a phone’s motion in the real world, detect planes in the environment, and understand lighting in the camera scene.

This developer preview release is not intended for use in production apps that are shipped to customers. Future versions of ARCore might introduce breaking changes to the API.

Supported Devices

This release is only supported on, and should only be installed on, qualified devices running Android N and later. See the list of ARCore supported devices for specific device models.

Installing this release on phones other than the supported devices might break other software on the device.

Notes

Multiple Versions of ARCore

• Apps built with the original ARCore SDK Developer Preview are only compatible with the original arcore-preview.apk. Apps built with ARCore SDK Developer Preview 2 are only compatible with arcore-preview2.apk.
• A developer can install versions of ARCore provided in both the original ARCore SDK Developer Preview and ARCore SDK Developer Preview 2 at the same time to facilitate using apps built with either SDK. The original arcore-preview.apk will be visible as “Tango Core” in the Android Apps settings whereas arcore-preview2.apk will be visible as “ARCore” in the Android Apps settings.

Known Issues

• ArSession_configure() does not apply settings changes if called while in a resumed state.

What’s New In Developer Preview 2

New SDK

• Developer Preview 2 features an Android NDK (C API)

API Changes

• The API interfaces have undergone significant revision and changes. All key functionality from the original Developer Preview remains, but the methods and function calls have changed.
• The API now natively supports attaching Anchors to Planes.
• ARCore Developer Preview 2 includes Unity and Java samples for camera image data access on the CPU. The provided Computer Vision sample is useful for running custom computer vision algorithms or image analysis on the camera image data.

Lifecycle Improvements

• ARCore apps may recover planes and anchors if the session is paused and then resumed, provided the user is in roughly the same location and the environment or lighting has not changed significantly. This enables scenarios where AR content is not lost when the user briefly switches apps.

Fixes

• Many stability and performance improvements.