How to Extract Raw Data and What Is Included

Export Raw Data from LiDAR Captures

Extract comprehensive raw data from your LiDAR captures to enable advanced 3D reconstruction, precise measurements, and custom processing workflows. Access depth maps, camera parameters, mesh information, and more for professional analysis and development.

Important: You must use the same device that originally captured the LiDAR data to export raw data. Raw data export is only available for captures made after enabling Developer Mode.

Prerequisites

Before you can extract raw data from your LiDAR captures, you'll need to enable Developer Mode on your device. For detailed instructions, see How to Access Developer Mode.

Developer Mode Required: Once enabled, all new LiDAR captures will include a Raw Export option. Captures made before enabling Developer Mode cannot be exported as raw data.

How to Export Raw Data

Follow these steps to export raw data from your LiDAR captures:

1Enable Developer Mode

Ensure that Developer Mode is enabled on your device. This must be done before capturing to enable raw data export functionality.

2Open Your Capture

Navigate to your capture library and select the LiDAR capture you want to export.

3Access Export Options

Tap the export option (arrow icon) located on the right-hand side of the capture editor to open the export window.

4Select Raw Data Export

Scroll to the bottom of the export options and select Raw Data export.

5Download the File

Click Export to download the file. The export will download as a zip file. Once downloaded, you'll need to unzip the file to access the raw data contents.

What's Included in the Raw Data Export?

The raw data export includes several folders and files containing detailed information about your LiDAR capture. Below is a comprehensive guide to understanding each component.

Cameras (JSON File)

This JSON file provides a compact way to store both camera information and environmental data for each frame or scan. It enables 3D reconstruction, accurate measurements, and proper alignment across multiple captures.

Image Dimensions

width and height - The resolution of the captured image in pixels. 

{
  "width": 1024,
  "height": 768
}

Timestamp

timestamp - A numerical identifier marking when the frame or scan was captured. Used to order frames or sync data, but may not represent real-world time unless tied to a specific internal clock. 

{
  "timestamp": 638443502
}

Image Sharpness

blur_score - Reflects how blurry or sharp the image is. Higher values indicate more blur, while lower values suggest a sharper image. 

{
  "blur_score": 115.28
}

Camera Intrinsic Parameters

These values describe the internal characteristics of the camera lens:

  • fx and fy - The focal length of the camera in pixels along the x and y axes
  • cx and cy - The principal point (optical center) of the camera image in pixel coordinates, typically near the center of the image

Together, these values form the camera intrinsic matrix: 

[fx  0  cx]
[0  fy  cy]
[0   0   1]

Example: 

{
  "fx": 712.3738,
  "fy": 712.3738,
  "cx": 516.26404,
  "cy": 384.30533
}

Camera Pose (Extrinsics)

t_00 to t_23 - These values represent transformation matrices used to describe the camera's position and orientation in 3D space. Most commonly, this is a 4×4 matrix that defines how to convert points between world space and camera space.

The numbers are labeled row-wise:

[t_00  t_01  t_02  t_03]
[t_10  t_11  t_12  t_13]
[t_20  t_21  t_22  t_23]

This structure is for understanding how 3D data is aligned and projected.

Additional Fields

manual_keyframe - A boolean indicating whether this frame was manually selected as a keyframe (used in keyframe-based scanning or tracking). 

{
  "manual_keyframe": false
}

angular_velocity - A measurement of how fast the camera was rotating when the frame was captured. Often used for motion tracking. May show a placeholder value (3.4028235e+38) if unused or not captured.

center_depth - The depth (distance) from the camera to the center of the image, typically measured in meters. 

{
  "center_depth": 2.93
}

This means the object in the center of the image is 2.93 meters away from the camera.

Confidence Images

A confidence map or confidence image is commonly used in LiDAR scanning and depth sensing to represent how reliable or accurate each pixel's depth value is.

Understanding Confidence Maps: The image is in grayscale (or pseudocolored), and each pixel's brightness indicates the confidence of the depth measurement at that location.

Interpreting Confidence Values

  • White = High confidence
    These areas have depth values that the LiDAR sensor is very sure about.
  • Black = Low confidence
    These pixels have depth values that are likely unreliable or noisy (e.g., due to distance, reflectivity, or occlusion).
  • Gray values = Medium confidence
    These might still be usable but should be treated with caution for critical measurements.
Best Practice: Use confidence images to filter or weight your depth data. Areas with low confidence should be excluded from precision measurements or flagged for review.

Depth Images

A depth image encodes how far each pixel in the image is from the camera or LiDAR sensor. Each pixel's brightness (grayscale value) corresponds to a depth value.

Understanding Depth Encoding

  • White pixels = far away
  • Black pixels = very close
  • Grayscale = in between
Technical Note: Depth images provide pixel-by-pixel distance measurements that can be used for 3D reconstruction, object detection, measurement calculations, and spatial analysis.

Mesh Info (JSON File)

This file contains metadata about a 3D mesh generated from your LiDAR capture, including its geometry, location, and alignment.

Mesh Information Fields

Key Meaning
vertexCount Number of vertices (points) in the mesh.
faceCount Number of triangle faces that make up the mesh surface.
totalArea Total surface area of the mesh in square meters.
bboxSize The width, height, and depth of the bounding box containing the mesh. Format: [x, y, z] in meters.
bboxCenter Center point of the bounding box. Often [0, 0, 0] when unshifted.
alignmentTransform A 4×4 matrix (flattened) describing how the mesh is aligned in space. This matrix includes rotation, translation, and possibly scaling. Useful for reconstructing global orientation.
xPlusArea, xMinusArea, yMinusArea, zPlusArea, zMinusArea, horizontalUpArea Surface area (in m²) of the mesh portions facing a certain direction. Helps analyze wall, floor, ceiling, or terrain coverage.
georeferenceLatitude / Longitude / Altitude The real-world location (in degrees and meters) where the mesh was captured or aligned.
georeferenceRotation The rotation (in radians) applied to align the model with the real-world compass.
yAlignmentRotation Indicates how the Y-axis of the mesh was rotated during alignment (likely relative to gravity).
Georeference Example: The mesh information includes precise location data:
  • Latitude: 41.721196
  • Longitude: -81.353154
  • Altitude: 183.56 m
This enables accurate placement of your 3D model in real-world coordinate systems.