Mobile mapping generic

Deliver 360° panorama images, point clouds, or both. Panoramas need JPEG images and one CSV listing each photo's filename, position, orientation and capture time. Point clouds are delivered as LAS or LAZ files.

Prepare your delivery#

• Panoramas: JPEG images and one pose CSV, using the layout below.
• Point clouds: LAS or LAZ files in a point cloud folder.
• Both: include both folders in the same dataset.

Deliver at least one of these. Before submitting for validation, check the dataset's coordinate system and required settings.

Folder layout#

Use this layout for a new delivery. Omit the folder for any data you are not delivering.

<dataset bucket>/
├── panorama/
│   ├── poses.csv              one CSV; the filename is your choice
│   └── jpg/
│       ├── PANO_0001.jpg
│       └── PANO_0002.jpg
└── pointcloud/
    └── scan_001.laz

Put the CSV directly inside panorama/ and the images directly inside panorama/jpg/. Do not put them in further subfolders. Deliver complete, valid JPEG files; an incomplete upload or a different image format renamed to .jpg fails the delivery.

Point clouds may be in subfolders of pointcloud/. See Point cloud generic for file requirements, including checks for incomplete files and LAS Extra Bytes metadata. Here, point clouds are optional.

Alternative folder names#

Existing deliveries may use these names instead, in the order shown:

Use only one folder for each kind of content. For panoramas, the first folder containing any file is used. For point clouds, the first folder containing LAS or LAZ files is used. Files in later folders are not included.

Each panorama layout needs a CSV directly in the chosen folder and a jpg/ subfolder containing the images. A missing CSV fails the delivery. A missing or empty jpg/ subfolder leaves the panoramas out; another panorama folder is not used as a fallback.

Coordinate reference system#

Supply data in the coordinate system configured for the dataset. Coordinates are used as delivered; they are not reprojected automatically.

• Photo positions (x, y, z) must use the dataset's 3D CRS.
• Point clouds must use the same horizontal system and vertical datum as the photo positions. Their boundaries use the dataset's 2D CRS.
• z is the absolute height in that CRS, not the camera's height above the road or ground.

For example, with EPSG:7415, x and y are RD coordinates in metres and z is height in metres relative to NAP. Camera height above ground is a separate dataset setting.

See Coordinates for more background.

The pose CSV#

Create one UTF-8, comma-separated CSV with a header row, followed by one row per photo. For example:

filename,x,y,z,heading,pitch,roll,timestamp
PANO_0001.jpg,182518.396,477934.743,37.326,0.0,0.0,0.0,2025-12-13T15:09:06.704Z
PANO_0002.jpg,182519.204,477936.881,37.341,1.4,0.1,-0.2,2025-12-13T15:09:08.113Z

Keep the columns in this order. Header names may differ (for example, heading is also called pan), but changing a header does not change the meaning of its column. Always include the header, or the first photo will be left out.

Use a decimal point as in the example. Decimal commas are also accepted, but a value containing a comma must be enclosed in double quotes so it stays in one CSV column.

Deliver only these columns. Positions after the optional floor column are reserved for future fields. Include exactly one CSV in the panorama folder; CSVs in subfolders are not used.

Filename matching#

Use the photo's filename, including .jpg or .jpeg, in the first column. Names are case-sensitive: PANO_0001.jpg and pano_0001.jpg do not match. Only images directly inside jpg/ are included.

Filenames without an extension are also accepted: .jpg is tried first, then .jpeg. If an extension is supplied, it must match the delivered file.

Timestamps#

Use a capture time with an explicit timezone where possible, such as 2025-12-13T15:09:06.704Z. The Z means UTC. An explicit offset such as +01:00 is also accepted.

Accepted timestamp formats are listed below. In the date formats, [.fff] means optional fractional seconds.

• epoch seconds (whole or fractional)
• RFC 3339 (2026-05-07T09:18:00Z)
• YYYY-MM-DD HH:MM:SS[.fff] (read as UTC)
• YYYY-MM-DDTHH:MM:SS[.fff] (read as UTC)
• YYYY/MM/DD HH:MM:SS[.fff] (read as UTC)
• DD-MM-YYYY HH:MM:SS[.fff] (read as Dutch local time, Europe/Amsterdam)
• YYYY-MM-DD (UTC midnight)

Epoch seconds means Unix time, counted from 1970-01-01 UTC, not GPS Time or milliseconds. Date-time formats without a timezone use UTC, except DD-MM-YYYY HH:MM:SS, which uses Dutch local time (Europe/Amsterdam, CET or CEST depending on the date).

If you also deliver point clouds, check that photo and scanner capture times agree. This matters for measurements in panoramas.

Heading, pitch and roll#

Deliver all three angles in degrees. They describe the centre of the panorama: the middle column of the equirectangular image.

Check the roll sign before submitting for validation. A wrong sign can pass the delivery checks but leave the panoramas tilted in the viewer.

Data from Orbit and CoPre: use Orbit mobile mapping for an Orbit generic traject file, which uses the opposite roll sign. For a CHCNAV CoPre *_OrbitPos.txt file, arrange the values in the CSV layout above, copy Pan and Tilt unchanged, and negate Roll.

Floors (optional)#

For indoor imagery, add a floor label as the ninth column:

filename,x,y,z,heading,pitch,roll,timestamp,floor
PANO_0001.jpg,182518.396,477934.743,37.326,0.0,0.0,0.0,2025-12-13T15:09:06.704Z,Ground
PANO_0002.jpg,182519.204,477936.881,41.012,1.4,0.1,-0.2,2025-12-13T15:09:08.113Z,1

Use consistent labels: B1 and b1 appear as different floors in the viewer's floor menu. Labels may contain up to 64 characters. Leave the cell blank, or omit the ninth column, for a photo without a floor.

Dataset settings#

Set camera height above ground and anonymisation on the dataset before submitting the delivery. Both settings are required for this dataset type, including when you deliver point clouds only.

Camera height above ground is in metres; for example, enter 2.3 for a camera mounted 2.3 m above the road. It does not replace the absolute z coordinate in the CSV.

Missing and unlisted photos#

Aim to deliver every photo listed in the CSV. The following allowance is for missing files; a corrupt JPEG that is present still fails the delivery.

• Up to 10 % of a recording's CSV rows may reference missing photos. Beyond that the run fails, naming the recording.
• Photos in the delivery that the CSV does not reference are not indexed (noted, not an error).

Measurements in panoramas#

When you deliver panoramas and point clouds together, the point cloud can provide depth for measurements in the panoramas. There is no extra depth file to upload. Coverage, coordinates and capture times must agree; see Depth for the matching requirements.

Before submitting#

• Confirm the dataset's CRS matches the coordinates and heights you supply.
• Check filenames, including capitalisation, against the CSV.
• Include the CSV header and keep the documented column order.
• Check angle conventions and timestamp timezones.
• Finish all uploads and set the two required dataset settings.

Troubleshooting#