Agisoft sample datasets¶

Agisoft hosts ten free sample datasets at https://www.agisoft.com/downloads/sample-data/. Each ships as a .zip on download.agisoft.com. They cover representative capture scenarios and are the canonical material to validate procedures against and to illustrate articles in this manual.

Use them when you can. Hand-rolled or contributor-private datasets are fine if a sample doesn't fit the article, but a sample-data validation is reproducible by any reader.

Quick reference¶

Dataset	Volume	Best illustrates
Witcher	239 images	Close-range / turntable, background-mask workflows
Building	50 images	Terrestrial facade modelling, planar orthomosaic
Coded targets	6 images, 3 targets	Marker-based scaling, Detect Markers
Aerial images (with GCPs)	444 images, 18 GCPs	UAV workflow, georeferencing, GNSS offset, GCP optimisation
Aerial LiDAR	148M points + trajectory	Aerial LiDAR + photogrammetry fusion, classification
Terrestrial LiDAR	9 E57 scans + panoramas	Terrestrial laser-scan workflows, scan registration
Spherical panorama (single)	73 images	Single-station panorama generation
Spherical panoramas (interior)	9 panoramas	Interior modelling from panoramas
Depth images	29 images	iPad-Pro depth-sensor capture, RGB+depth fusion
Interior	391 images	Indoor capture, low-feature surfaces

Naming this manual uses¶

"the Witcher dataset" → the close-range turntable set
"the Building dataset" → the terrestrial facade set
"the Coded targets dataset" → the markers/scaling set
"the Aerial-with-GCPs dataset" → the 444-image UAV set
"the Aerial LiDAR dataset" → the airborne lidar set
"the Terrestrial LiDAR dataset" → the ground-based lidar set
"the Spherical-panorama dataset" → the single-station 73-image set
"the Spherical-panoramas dataset" → the 9-panorama interior set
"the Depth-images dataset" → the iPad Pro set
"the Interior dataset" → the 391-image indoor set

When citing a dataset for the first time in an article, link to the section anchor below; subsequent references can use the short name.

Witcher¶

Close-range images taken with a turntable with a background image for masks generation. 239 images. Download: https://download.agisoft.com/datasets/the_witcher.zip

Best illustrates

Turntable / close-range object capture
Masks generated from a background image (Tools → Import Masks → From Background)
Aligning images that share a moving foreground over a static background

Used by

Exclude stationary tie points: when and why
Scale-bar-free workflows where measurement matters less than topology

Articles in this manual

(none yet — link from the next article that uses it.)

Building¶

Terrestrial images for facade 3D model and planar orthomosaic generation. 50 images. Download: https://download.agisoft.com/datasets/building.zip

Best illustrates

Terrestrial / non-aerial photogrammetry
Planar (vertical-projection) orthomosaic generation
Mesh quality on man-made flat surfaces
A small, fast-converging alignment for tooling smoke tests

Articles in this manual

(none yet)

Coded targets¶

Images with coded targets to scale a model. 6 images, 3 targets. Download: https://download.agisoft.com/datasets/coded.zip

Best illustrates

Detect Markers with coded targets
Scale-bar definition and measurement
Marker-based georeferencing without GCPs
The minimum useful project size (6 images)

Articles in this manual

Removing "blue flag" marker projections cleanly — uses the Coded targets dataset to demonstrate the bulk-removal pattern after chunk.detectMarkers().
Programmatic marker placement and pinning — uses the Coded targets dataset to demonstrate that all three Marker.Projection constructor input forms (Vector, list, tuple) produce equivalent projections.

Aerial images (with GCPs)¶

Aerial images with GCPs and GNSS offset, including separate CSV file with camera coordinates. 444 images, 18 GCPs. Download: https://download.agisoft.com/datasets/aerial_images_with_gcps.zip

Best illustrates

UAV photogrammetry pipeline end-to-end
Reference Preselection with geotagged photos
GCP import, marker placement, and Optimize Cameras after GCPs
GNSS antenna offset handling
Reference-pane CRS configuration

Articles in this manual

The Clean Tie Points → Optimize Cameras loop — uses Aerial-with-GCPs (or Building as a faster alternative) to demonstrate the alignment + cleanup-loop scripted end-to-end.
Reproducing chunk-info statistics in Python — uses Aerial-with-GCPs (georeferenced, has 18 GCPs) to demonstrate all three chunk-info formulae and verify them against the GUI display.
Setting chunk.region to bound the tie-point cloud — uses Aerial-with-GCPs to demonstrate the simple internal- coordinates region resize after alignment.
Exporting depth maps from Python (32-bit, scaled) — uses Aerial-with-GCPs to demonstrate the float32 export with chunk.transform.scale applied.

Aerial LiDAR¶

Aerial LiDAR point cloud with trajectory data. 148 million points, 1 trajectory file. Download: https://download.agisoft.com/datasets/aerial_lidar.zip

Best illustrates

LiDAR import and trajectory alignment
Point cloud classification (Tools → Classify Points…)
Filter by Class / Filter by Return / Filter by Confidence
Memory-bound workflows (148 M points is non-trivial on consumer hardware)

Articles in this manual

(none yet)

Terrestrial LiDAR¶

Terrestrial laser scans with built-in panoramas. 9 scans in E57 format. Download: https://download.agisoft.com/datasets/terrestrial_lidar.zip

Best illustrates

E57 import workflows
Laser-scan registration and Align Laser Scans
Panorama-from-scan colourisation
Scan-vs-photo hybrid alignment

Articles in this manual

(none yet)

Spherical panorama¶

Images taken from single point for panorama generation. 73 images. Download: https://download.agisoft.com/datasets/pano.zip

Best illustrates

Single-station panorama generation
Stitching from a non-spherical capture rig
Workflow → Build Panorama

Articles in this manual

(none yet)

Spherical panoramas (interior)¶

Spherical panoramas for interior modeling. 9 spherical panoramas. Download: https://download.agisoft.com/datasets/spherical_panoramas.zip

Best illustrates

Interior reconstruction from spherical panoramas
Spherical camera type configuration
Multi-station alignment in a confined space

Articles in this manual

(none yet)

Depth images¶

Images from front camera of iPad pro with a depth sensor. 29 images. Download: https://download.agisoft.com/datasets/depth_images.zip

Best illustrates

Depth-sensor + RGB fusion
iPad / mobile capture pipelines
Pre-existing depth maps fed into the photogrammetry pipeline

Articles in this manual

(none yet)

Interior¶

Terrestrial images for interior modeling. 391 images. Download: https://download.agisoft.com/datasets/interior.zip

Best illustrates

Indoor capture with low-feature surfaces (walls, floors)
Multi-room alignment without GPS
Lighting variation across rooms

Articles in this manual

(none yet)

Maintenance¶

This page is hand-maintained. When you write an article that uses one of these datasets:

Add a one-line entry under that dataset's Articles in this manual section linking to the article.
In the article, link the dataset by its anchor here on first mention.
Captions on dataset-derived screenshots should name the dataset and the Metashape version (per STYLE.md).

Adding a new dataset (if Agisoft publishes more): copy the section template above, add a row to the Quick reference table, and add the short name to the Naming this manual uses list.