Reference

This page provides a guide to the BigQuery tables used for training models, including their schemas and field descriptions. You'll also find a breakdown of the core concepts—snapshots, assets, and observations—that structure the data, along with details on how to use the tables where relevant.

Table names are prefixed by product type (cropped_, full_frame_, or pano_), and only the tables for the products included in your subscriptions are present in your dataset. Tables ending in _all include data from all snapshots, and tables ending in _latest include data only from the most recent snapshot. LiDAR metadata tables (*_lidar_metadata_all and *_lidar_metadata_latest) are included in your dataset only when LiDAR is enabled for your subscription.

Core concepts

  • Snapshot: A fixed, unchangeable copy of a dataset at a specific moment in time.
  • Asset: An object in the real world, such as a road sign or a utility pole. Assets are available for the cropped and full-frame products.
  • Observation: An image record from a capture event. For the cropped and full-frame products, an observation is a detection of an asset, and contains metadata for that detection, such as the time the image was captured and the position of the camera. For the panoramic product, an observation is an individual image frame of a 360-degree capture, and isn't associated with an asset.
  • LiDAR metadata: Metadata and Cloud Storage file locations for 3D laser point clouds (.laz files) associated with asset observations. Available for the cropped and full-frame products when LiDAR is enabled, each point cloud is spatially cropped to the immediate 3D region surrounding a detected asset.

snapshots

A snapshot is a fixed, unchangeable copy of a dataset at a specific moment in time. The snapshot table provides metadata associated with snapshots, to allow you to understand the temporal state of the data at a high level.

Column name Type Description
snapshot_id STRING Unique identifier for the snapshot. Used as a key to join tables.
subscription_id STRING Unique identifier for the subscription.
creation_time TIMESTAMP Timestamp of when the snapshot was created. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.
product_types ARRAY<STRING> The product types associated with the snapshot.
Valid values are SV_CROPPED_ASSETS, SV_FULL_FRAME_ASSETS, and SV_PANO.

cropped_observations_all

The cropped_observations_all table provides cropped asset observations from all snapshots. You can use this table to detect differences in observations between snapshots.

Column name Type Description
snapshot_id STRING Unique identifier for the snapshot which collected this observation.
asset_id STRING Unique identifier for the asset associated with this observation.
asset_type STRING Major classification of the asset.
Valid values are ASSET_CLASS_ROAD_SIGN, ASSET_CLASS_UTILITY_POLE, and ASSET_CLASS_MAINTENANCE_COVER.
location STRUCT Struct containing latitude/longitude coordinates as floats. Location where the observation was captured.
detection_time TIMESTAMP The most recent observation capture_time for the given asset. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.
observation_id STRING String that uniquely identifies the observation.
bbox STRUCT Struct of structs containing the x/y coordinates aligning the asset's bounding box.
camera_pose STRUCT The position of the camera LLA and Euler angles. Struct containing floats for latitude/longitude, altitude (in meters), pitch, heading, and roll (in degrees).
capture_time TIMESTAMP Timestamp of when the observation was captured. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.
pano_id STRING Unique identifier for a 360-degree stitched panoramic image. This value is nullable; it may not be present. If present, it can be used as an accessor key with the Google Maps Street View Static API.
gcs_uri STRING Google Cloud Storage URI where the image is hosted.
map_url STRING Google Maps URL that shows the location of the asset.

cropped_assets_all

The cropped_assets_all table provides cropped assets from all snapshots. You can use this table to detect differences in assets between snapshots.

Column name Type Description
asset_id STRING Unique identifier for the asset.
snapshot_id STRING Unique identifier for the snapshot which collected this asset.
asset_type STRING Major classification of the asset.
Valid values are ASSET_CLASS_ROAD_SIGN, ASSET_CLASS_UTILITY_POLE, and ASSET_CLASS_MAINTENANCE_COVER.
location STRUCT Struct containing latitude/longitude coordinates as floats that represents the asset's approximate location.
detection_time TIMESTAMP The most recent observation capture_time for the given asset. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.

cropped_lidar_metadata_all

The cropped_lidar_metadata_all table provides LiDAR point cloud metadata for cropped asset observations from all snapshots. You can use this table to detect differences in LiDAR point clouds between snapshots.

Column name Type Description
uri STRING Google Cloud Storage URI (gs://...) where the compressed LiDAR point cloud (.laz) file is hosted.
observation_id STRING String that uniquely identifies the observation.
asset_id STRING Unique identifier for the asset associated with this observation.
snapshot_id STRING Unique identifier for the snapshot which collected this observation.
device_number INT64 Index of the LiDAR sensor on the capture vehicle that produced this point cloud. Multiple LiDAR sensors may record observations for the same capture event.
device_specs STRUCT Hardware specifications of the spinning LiDAR sensor. Struct containing shots_per_beam_rotation (INT64) and rotation_rate_hz (INT64).
count_points INT64 Total number of 3D laser return points contained in the point cloud file.
centroid_lat FLOAT64 WGS84 latitude (in degrees) of the spatial centroid of all laser returns in the point cloud.
centroid_lon FLOAT64 WGS84 longitude (in degrees) of the spatial centroid of all laser returns in the point cloud.
start_time TIMESTAMP Timestamp of when the first point in the LiDAR sweep was recorded. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.
end_time TIMESTAMP Timestamp of when the final point in the LiDAR sweep was recorded. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.

cropped_observations_latest

The cropped_observations_latest table provides cropped asset observations only from the most recent snapshot.

Column name Type Description
snapshot_id STRING Unique identifier for the snapshot which collected this observation.
asset_id STRING Unique identifier for the asset associated with this observation.
asset_type STRING Major classification of the asset.
Valid values are ASSET_CLASS_ROAD_SIGN, ASSET_CLASS_UTILITY_POLE, and ASSET_CLASS_MAINTENANCE_COVER.
location STRUCT Struct containing latitude/longitude coordinates as floats. Location where the observation was captured.
detection_time TIMESTAMP The most recent observation capture_time for the given asset. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.
observation_id STRING String that uniquely identifies the observation.
bbox STRUCT Struct of structs containing the x/y coordinates aligning the asset's bounding box.
camera_pose STRUCT The position of the camera LLA and Euler angles. Struct containing floats for latitude/longitude, altitude (in meters), pitch, heading, and roll (in degrees).
capture_time TIMESTAMP Timestamp of when the observation was captured. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.
pano_id STRING Unique identifier for a 360-degree stitched panoramic image. This value is nullable; it may not be present. If present, it can be used as an accessor key with the Google Maps Street View Static API.
gcs_uri STRING Google Cloud Storage URI where the image is hosted.
map_url STRING Google Maps URL that shows the location of the asset.

cropped_assets_latest

The cropped_assets_latest table provides cropped assets only from the most recent snapshot.

Column name Type Description
asset_id STRING Unique identifier for the asset.
snapshot_id STRING Unique identifier for the snapshot which collected this asset.
asset_type STRING Major classification of the asset.
Valid values are ASSET_CLASS_ROAD_SIGN, ASSET_CLASS_UTILITY_POLE, and ASSET_CLASS_MAINTENANCE_COVER.
location STRUCT Struct containing latitude/longitude coordinates as floats that represents the asset's approximate location.
detection_time TIMESTAMP The most recent observation capture_time for the given asset. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.

cropped_lidar_metadata_latest

The cropped_lidar_metadata_latest table provides LiDAR point cloud metadata for cropped asset observations only from the most recent snapshot.

Column name Type Description
uri STRING Google Cloud Storage URI (gs://...) where the compressed LiDAR point cloud (.laz) file is hosted.
observation_id STRING String that uniquely identifies the observation.
asset_id STRING Unique identifier for the asset associated with this observation.
snapshot_id STRING Unique identifier for the snapshot which collected this observation.
device_number INT64 Index of the LiDAR sensor on the capture vehicle that produced this point cloud. Multiple LiDAR sensors may record observations for the same capture event.
device_specs STRUCT Hardware specifications of the spinning LiDAR sensor. Struct containing shots_per_beam_rotation (INT64) and rotation_rate_hz (INT64).
count_points INT64 Total number of 3D laser return points contained in the point cloud file.
centroid_lat FLOAT64 WGS84 latitude (in degrees) of the spatial centroid of all laser returns in the point cloud.
centroid_lon FLOAT64 WGS84 longitude (in degrees) of the spatial centroid of all laser returns in the point cloud.
start_time TIMESTAMP Timestamp of when the first point in the LiDAR sweep was recorded. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.
end_time TIMESTAMP Timestamp of when the final point in the LiDAR sweep was recorded. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.

full_frame_observations_all

The full_frame_observations_all table provides full-frame (uncropped) asset observations from all snapshots. You can use this table to detect differences in observations between snapshots.

Column name Type Description
snapshot_id STRING Unique identifier for the snapshot which collected this observation.
asset_id STRING Unique identifier for the asset associated with this observation.
asset_type STRING Major classification of the asset.
Valid values are ASSET_CLASS_ROAD_SIGN, ASSET_CLASS_UTILITY_POLE, and ASSET_CLASS_MAINTENANCE_COVER.
location STRUCT Struct containing latitude/longitude coordinates as floats. Location where the observation was captured.
detection_time TIMESTAMP The most recent observation capture_time for the given asset. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.
observation_id STRING String that uniquely identifies the observation.
bbox STRUCT Struct of structs containing the x/y coordinates aligning the asset's bounding box.
camera_pose STRUCT The position of the camera LLA and Euler angles. Struct containing floats for latitude/longitude, altitude (in meters), pitch, heading, and roll (in degrees).
capture_time TIMESTAMP Timestamp of when the observation was captured. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.
pano_id STRING Unique identifier for a 360-degree stitched panoramic image. This value is nullable; it may not be present. If present, it can be used as an accessor key with the Google Maps Street View Static API.
gcs_uri STRING Google Cloud Storage URI where the image is hosted.
map_url STRING Google Maps URL that shows the location of the asset.

full_frame_assets_all

The full_frame_assets_all table provides full-frame (uncropped) assets from all snapshots. You can use this table to detect differences in assets between snapshots.

Column name Type Description
asset_id STRING Unique identifier for the asset.
snapshot_id STRING Unique identifier for the snapshot which collected this asset.
asset_type STRING Major classification of the asset.
Valid values are ASSET_CLASS_ROAD_SIGN, ASSET_CLASS_UTILITY_POLE, and ASSET_CLASS_MAINTENANCE_COVER.
location STRUCT Struct containing latitude/longitude coordinates as floats that represents the asset's approximate location.
detection_time TIMESTAMP The most recent observation capture_time for the given asset. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.

full_frame_lidar_metadata_all

The full_frame_lidar_metadata_all table provides LiDAR point cloud metadata for full-frame (uncropped) asset observations from all snapshots. You can use this table to detect differences in LiDAR point clouds between snapshots.

Column name Type Description
uri STRING Google Cloud Storage URI (gs://...) where the compressed LiDAR point cloud (.laz) file is hosted.
observation_id STRING String that uniquely identifies the observation.
asset_id STRING Unique identifier for the asset associated with this observation.
snapshot_id STRING Unique identifier for the snapshot which collected this observation.
device_number INT64 Index of the LiDAR sensor on the capture vehicle that produced this point cloud. Multiple LiDAR sensors may record observations for the same capture event.
device_specs STRUCT Hardware specifications of the spinning LiDAR sensor. Struct containing shots_per_beam_rotation (INT64) and rotation_rate_hz (INT64).
count_points INT64 Total number of 3D laser return points contained in the point cloud file.
centroid_lat FLOAT64 WGS84 latitude (in degrees) of the spatial centroid of all laser returns in the point cloud.
centroid_lon FLOAT64 WGS84 longitude (in degrees) of the spatial centroid of all laser returns in the point cloud.
start_time TIMESTAMP Timestamp of when the first point in the LiDAR sweep was recorded. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.
end_time TIMESTAMP Timestamp of when the final point in the LiDAR sweep was recorded. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.

full_frame_observations_latest

The full_frame_observations_latest table provides full-frame (uncropped) asset observations only from the most recent snapshot.

Column name Type Description
snapshot_id STRING Unique identifier for the snapshot which collected this observation.
asset_id STRING Unique identifier for the asset associated with this observation.
asset_type STRING Major classification of the asset.
Valid values are ASSET_CLASS_ROAD_SIGN, ASSET_CLASS_UTILITY_POLE, and ASSET_CLASS_MAINTENANCE_COVER.
location STRUCT Struct containing latitude/longitude coordinates as floats. Location where the observation was captured.
detection_time TIMESTAMP The most recent observation capture_time for the given asset. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.
observation_id STRING String that uniquely identifies the observation.
bbox STRUCT Struct of structs containing the x/y coordinates aligning the asset's bounding box.
camera_pose STRUCT The position of the camera LLA and Euler angles. Struct containing floats for latitude/longitude, altitude (in meters), pitch, heading, and roll (in degrees).
capture_time TIMESTAMP Timestamp of when the observation was captured. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.
pano_id STRING Unique identifier for a 360-degree stitched panoramic image. This value is nullable; it may not be present. If present, it can be used as an accessor key with the Google Maps Street View Static API.
gcs_uri STRING Google Cloud Storage URI where the image is hosted.
map_url STRING Google Maps URL that shows the location of the asset.

full_frame_assets_latest

The full_frame_assets_latest table provides full-frame (uncropped) assets only from the most recent snapshot.

Column name Type Description
asset_id STRING Unique identifier for the asset.
snapshot_id STRING Unique identifier for the snapshot which collected this asset.
asset_type STRING Major classification of the asset.
Valid values are ASSET_CLASS_ROAD_SIGN, ASSET_CLASS_UTILITY_POLE, and ASSET_CLASS_MAINTENANCE_COVER.
location STRUCT Struct containing latitude/longitude coordinates as floats that represents the asset's approximate location.
detection_time TIMESTAMP The most recent observation capture_time for the given asset. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.

full_frame_lidar_metadata_latest

The full_frame_lidar_metadata_latest table provides LiDAR point cloud metadata for full-frame (uncropped) asset observations only from the most recent snapshot.

Column name Type Description
uri STRING Google Cloud Storage URI (gs://...) where the compressed LiDAR point cloud (.laz) file is hosted.
observation_id STRING String that uniquely identifies the observation.
asset_id STRING Unique identifier for the asset associated with this observation.
snapshot_id STRING Unique identifier for the snapshot which collected this observation.
device_number INT64 Index of the LiDAR sensor on the capture vehicle that produced this point cloud. Multiple LiDAR sensors may record observations for the same capture event.
device_specs STRUCT Hardware specifications of the spinning LiDAR sensor. Struct containing shots_per_beam_rotation (INT64) and rotation_rate_hz (INT64).
count_points INT64 Total number of 3D laser return points contained in the point cloud file.
centroid_lat FLOAT64 WGS84 latitude (in degrees) of the spatial centroid of all laser returns in the point cloud.
centroid_lon FLOAT64 WGS84 longitude (in degrees) of the spatial centroid of all laser returns in the point cloud.
start_time TIMESTAMP Timestamp of when the first point in the LiDAR sweep was recorded. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.
end_time TIMESTAMP Timestamp of when the final point in the LiDAR sweep was recorded. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.

pano_observations_all

The pano_observations_all table provides panoramic (full scene) observations from all snapshots. Unlike asset observations, panoramic observations contain frame-level camera metadata and spatial orientation for 360-degree capture events rather than localized asset detections. You can use this table to detect differences in panoramic observations between snapshots.

Column name Type Description
capture_id STRING Unique identifier representing a single panoramic capture event.
observation_id STRING Unique identifier for an individual image frame that forms a single segment of a 360-degree panoramic capture.
snapshot_id STRING Unique identifier for the snapshot which collected this observation.
capture_location STRUCT Struct containing latitude/longitude coordinates as floats representing the centroid of the vehicle when the observation was captured.
capture_time TIMESTAMP Timestamp of when the observation was captured. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.
camera_pose STRUCT The position of the camera LLA and Euler angles. Struct containing floats for latitude/longitude, altitude (in meters), pitch, heading, and roll (in degrees).
gcs_uri STRING Google Cloud Storage URI where the image is hosted.
map_url STRING Google Maps URL that shows the capture location of the observation.
pano_id STRING Unique identifier for a 360-degree stitched panoramic image. This value is nullable; it may not be present. If present, it can be used as an accessor key with the Google Maps Street View Static API.

pano_observations_latest

The pano_observations_latest table provides panoramic (full scene) observations only from the most recent snapshot.

Column name Type Description
capture_id STRING Unique identifier representing a single panoramic capture event.
observation_id STRING Unique identifier for an individual image frame that forms a single segment of a 360-degree panoramic capture.
snapshot_id STRING Unique identifier for the snapshot which collected this observation.
capture_location STRUCT Struct containing latitude/longitude coordinates as floats representing the centroid of the vehicle when the observation was captured.
capture_time TIMESTAMP Timestamp of when the observation was captured. ISO 8601 formatted timestamp e.g., 2019-09-25 17:26:27.757171 UTC.
camera_pose STRUCT The position of the camera LLA and Euler angles. Struct containing floats for latitude/longitude, altitude (in meters), pitch, heading, and roll (in degrees).
gcs_uri STRING Google Cloud Storage URI where the image is hosted.
map_url STRING Google Maps URL that shows the capture location of the observation.
pano_id STRING Unique identifier for a 360-degree stitched panoramic image. This value is nullable; it may not be present. If present, it can be used as an accessor key with the Google Maps Street View Static API.

Deprecated tables