PoseTracker Architecture
What is PoseTracker?
PoseTracker configures Frame data such as the device's camera image and pose information, and then passes it to VLSDK's internal logic modules.
Frame Data
Frame data refers to the device information retrieved for each frame. The information that PoseTracker must pass to VLSDK's internal modules is as follows:
| Data | Type | Description | Note |
|---|---|---|---|
| texture | Texture | Camera preview of the current frame | |
| localPosition | Vector3 | Camera position of the current frame calculated by the device | |
| localRotation | Quaternion | Camera rotation of the current frame calculated by the device | |
| intrinsic | ARIntrinsic | Camera intrinsic parameters provided by the device | |
| projection matrix | Matrix4x4 | Projection matrix for content rendering | |
| display matrix | Matrix4x4 | Rotation matrix that rotates the original camera preview to the correct orientation | |
| yuvBuffer | UnityYuvCpuImage | YUV image buffer data of the camera preview. This field is used when useFaceBlurring is enabled on iOS and Android devices. | Optional |
The format of frame data varies depending on the framework being used. When using ARFoundation, it is provided in the form of ARCameraFrameEventArgs, and when using ARDataset, it is provided in the form of FrameData. These different formats must be converted into a unified ARFrame format so that PoseTracker can utilize them.
