Skip to Content
DocumentationReferenceDatabase Schema

Database Schema

The tables are organized in modules.

Acquisition System Module

Platform

  • Table: platform
  • Definition: A sensor platform represents the physical vehicle or mounting system that carries one or more 3D sensors during data collection (e.g., survey vehicle, UAV, tripod).
  • Purpose: Define the spatial relationships between sensors (lever arms, boresight angles) and platform dynamics.
  • Attributes: Name, platform type (vehicle, uav, tripod, static), description.
  • Examples:
    • “Survey Vehicle – VW Transporter #MUC-042”
    • “DJI Matrice 300 RTK – UAV Unit 3”
    • “Static Tripod Mount – Marienplatz Station A”

Sensor

  • Table: sensor
  • Definition: A sensor represents a 3D sensor mounted on a platform that captures geometric data of the environment. Supported sensor types are LiDAR scanners, cameras, and radar sensors.
  • Purpose: Track sensor characteristics, calibration data, and metadata required for post-processing and data fusion.
  • Attributes: Name, type (lidar, camera, radar), manufacturer, model number, specification (JSONB).
  • Examples:
    • “Riegl VUX-1LR LiDAR Scanner (Serial #12345)”
    • “Velodyne HDL-64E LiDAR Scanner”
    • “Basler ace2 Camera (Body #B007)“

Operations Module

LevelScopeTypical DurationExample EntityFocus
CampaignProject-wideWeeks–months”Munich Street Network Survey 2023”Overall data collection goal
MissionOne sortie or scan sessionHours”Mission 03 – Maxvorstadt Drive”Operational run
RecordingSingle sensor streamMinutes–hours”LiDAR_Maxvorstadt_2024-07-15_0830”Raw sensor data

Campaign

  • Table: campaign
  • Definition: A campaign represents an entire surveying project or data collection effort for a specific purpose, region, or time period. It contains many missions, possibly recorded by different platforms.
  • Attributes: Name.
  • Examples:
    • “Munich Street Network Survey 2023”
    • “Ingolstadt Road Space Mapping Campaign, Q2 2024”

Mission

  • Table: mission
  • Definition: A mission represents a single data collection run — typically corresponding to a drive or scan session — conducted under specific conditions. It belongs to a campaign and contains one or several recordings.
  • Attributes: Name.
  • Examples:
    • “Mission 03 – Maxvorstadt Drive (Morning Scan)”
    • “Mission 07 – Schwabing North Traverse”

Recording

  • Table: recording
  • Definition: A recording represents a continuous data acquisition session from one or more 3D sensors (LiDAR, camera, or radar) within a mission.
  • Purpose: Capture raw sensor data for post-processing (e.g., trajectory correction, point cloud generation).
  • Attributes: Name, start and end timestamps, references to mission, platform, and sensor. May reference a parent recording for segmented datasets.
  • Examples:
    • “LiDAR_Maxvorstadt_2024-07-15_0830”
    • “LiDAR_Recording_Mission03_SegmentA”

Trajectory Module

Trajectory

  • Table: trajectory
  • Definition: A trajectory represents the sequence of poses of a platform or sensor over time, derived from a recording.
  • Attributes: Domain (timed or sequence), interpolation type (step or linear), extrapolation type.

Trajectory Pose

  • Table: trajectory_pose
  • Definition: A trajectory pose represents a single position and orientation sample within a trajectory.
  • Attributes: Position (3D point), orientation (quaternion), timestamp (seconds + nanoseconds) or sequence index.

Point Cloud Module

Point Cloud

  • Table: point_cloud
  • Definition: A point cloud represents a spatial dataset of discrete 3D points generated from one recording or a defined time slice of it, typically sized to allow efficient processing on a single computer.
  • Attributes: Name, start and end timestamps.

Point Cloud Cell

  • Table: point_cloud_cell
  • Definition: A point cloud cell represents a single octant within the point cloud’s octree hierarchy, providing the spatial subdivision used for efficient storage, indexing, and querying.
  • Attributes: Octree level, grid coordinates (x, y, z), cell envelope, point envelope, point count, start and end timestamps.

Point Cloud Cell Data

  • Table: point_cloud_cell_data
  • Definition: A point cloud cell data record stores a named attribute for a point cloud cell.
  • Purpose: Enable flexible, schema-less storage of per-cell attributes such as point coordinates, intensity values, return numbers, or sensor beam geometries.
  • Attributes: Namespace, name, datatype, and a value column matching the data type (3D geometry, scalar, array, or quaternion).

The client is designed to support the maximum field size for each column, meaning that complex data structures — such as multi-point geometries or intensity value arrays — are transmitted as single, atomic requests without additional fragmentation or metadata.

Point Cloud Cell Metadata

  • Table: point_cloud_metadata
  • Definition: A point cloud metadata record stores a named scalar attribute at the point cloud level (not per cell), such as acquisition parameters or processing provenance.
  • Attributes: Namespace, name, datatype, scalar value (integer, float, or string).
Last updated on
ExportQueries