DepthCapture Lab

DepthCapture Lab is an iPhone / iPad app for recording RGB images and depth data by selecting from available depth sources such as LiDAR, TrueDepth, stereo vision, and AI depth estimation.

It lets you capture RGB-D data in formats suitable for research and development, 3D measurement, robotics, computer vision, AR validation, and related workflows, while previewing the results directly on the device.

Main Features

- RGB + Depth capture using LiDAR, TrueDepth, stereo vision, or AI depth estimation
- Selection of available depth sources
- Still image capture and continuous recording
- Live depth map preview
- Live point cloud preview
- Browsing, previewing, and playback of saved sessions
- RGB image export as PNG / JPEG
- Depth data export as BIN / 16-bit PNG / TIFF / EXR
- Recording as image sequences or MCAP files
- Frame-level and session-level metadata storage
- Settings for FPS, resolution, depth range, synchronization method, and depth filtering
- Data storage including camera intrinsic parameters
- Storage of acquisition settings, conditions, and metadata for each depth source

Depth data is saved in formats designed for post-processing. Each session includes a manifest.json file and frame-level metadata, allowing you to inspect saved images, depth maps, camera intrinsic parameters, depth sources, and other information required for reconstruction.

Recorded data can be saved either as image and depth sequences or as ROS 2-compatible MCAP files.
The data structure is designed to be easy to integrate into computer vision, point cloud processing, machine learning, and robotics workflows.

Supported Devices
- Rear depth capture using LiDAR: iPhone / iPad models with LiDAR
- Front depth capture using TrueDepth: iPhone / iPad models with TrueDepth
- Depth capture using stereo vision: iPhone / iPad models with supported camera configurations
- AI depth estimation: Metal-compatible devices capable of running supported models
- Metal-compatible devices

Available depth sources, depth formats, resolutions, and frame rates vary depending on the device, camera configuration, OS, and model settings.
Even on devices without LiDAR or TrueDepth, some depth capture features may be available if stereo vision or AI depth estimation is supported.

DepthCapture Lab is a practical RGB-D capture tool for recording depth data from multiple depth sources, visualizing it, and passing it on to post-processing workflows.