Introduction
Use this guide to achieve a high level of accuracy when processing datasets with the CDA feature. Prepare and execute your scans so that CDA can align the datasets correctly. If you want more information about the CDA feature, see How Cross-Dataset Alignment (CDA) works.
CDA works with surveyed control points and alignment markers, or a combination of both. If you are not familiar with control point types, see Control Point Types: Surveyed and Alignment Markers before continuing.
Advantages of Using CDA in Your Workflow
All NavVis devices can capture datasets that can then be processed using the CDA feature.
The feature does not require extra processing units beyond standard dataset processing.
When processing multiple datasets with only alignment markers, no survey equipment is required.
Note: CDA reduces SLAM drift but cannot fix SLAM breaks. If a SLAM break occurs, perform a rescan of the area. See SLAM breaks.
Creating a Dataset Group Using CDA
Three scenarios apply when capturing a dataset with control points: scanning and capturing with alignment markers only, with surveyed control points only, or a combination of the two.
1. Scanning and capturing with alignment markers only
Recommended layouts for placing alignment markers:
Hub layout: One master dataset in the center with adjacent datasets around it. Best for lobbies, atriums, and central halls.
Multiple floors: Place alignment markers on vertical overlapping areas (for example, stairs) to link the floors.
Note: Place at least two alignment markers in the overlapping areas; more markers improve results.
Note: Each adjacent dataset must share overlapping control points. When using alignment markers, larger overlapping areas improve results.
2. Scanning and capturing with surveyed control points only
If you are using all surveyed control points, CDA can detect overlaps without common alignment markers between datasets.
All datasets must have at least three surveyed control points, measured in the same reference system (local or georeferenced).
All adjacent datasets must have full coverage of the overlapping scanned areas.
3. Scanning and capturing with alignment markers and surveyed control points
If a dataset has three or more surveyed control points, you can extend the group with alignment-marker-only datasets; those datasets will be georeferenced. Recommended layouts for placing surveyed control points and alignment markers.
Linear chain layout: Datasets arranged sequentially. For long chains, add surveyed control points at both ends to prevent drift.
Loop / ring layout: An exterior loop with adjacent datasets on the inside. Closing the loop constrains every dataset from both sides.
General scanning information
Regardless of which control point type or combination you use, the following guidelines are important when capturing datasets for CDA.
Spread control points out. Avoid placing surveyed control points along a single axis. Place surveyed control points or alignment markers on permanent, visible features such as wall corners, columns, fixed signage, and door frames. For more information, see Placing control points.
Overlap with areas, not edges. When scanning overlapping areas, ensure they cover full sections, not just corners. Greater overlap improves accuracy.
Perform a 360° turn in the overlapping area. This ensures thorough coverage and helps CDA achieve stronger alignment.
Check the quality of the scan before leaving the site. Confirm full scanning coverage.
Processing
All datasets in one processing task share the same point cloud, panorama, and control point settings, though some settings can be grouped by device type (for example, NavVis VLX and NavVis MLX datasets in the same task). If you need different settings, split datasets into separate tasks. CDA does not align datasets across separate tasks.
You cannot add a dataset to a group later without reprocessing it. Double-check that the datasets you want to group together are correct.
Process related datasets in the same task. CDA aligns only within a task.
Provide the surveyed control point file in metric units, in a Cartesian, right-handed coordinate system (x, y, z in meters).
After Processing
Visually check the linked datasets in Dataset Alignment. Walls and features must line up.
Check the dataset for accuracy. Datasets with drifts indicate a problem that automatic alignment cannot fix. You may need to recapture the scanning session.
Note: Unlinking a dataset from its CDA group cannot be undone. To restore the original state, unload the datasets and reload them.