3.2 Route Planning and Coverage Strategy
Plan scan routes that use backbone-first structure, effective loop closures, and serpentine coverage to produce complete, drift-corrected datasets in any space.
The Backbone-First Approach
For any building with corridors connecting rooms, scan the main corridors first before branching into individual spaces. The corridor network is the backbone of the scan. Establishing it early means every room you branch into has a registered starting point already built into the dataset. When you exit a room and return to the corridor, you create a loop closure automatically.
Walk the Main Corridors First
Cover the full length of every primary corridor on the floor before entering any rooms. Walk at a moderate pace and focus on establishing the spine of the scan with clean, uninterrupted coverage.
Branch Into Rooms From the Corridor
Enter each room from the corridor, scan its interior completely, and return to the corridor through the same doorway you entered. Each entry and exit creates a loop closure between the room and the corridor backbone.
Return to the Initialization Area at the End
After all branches are complete, walk back to within 15 to 30 ft (5 to 10 m) of the initialization point from a similar angle to your starting position. This final loop closure corrects any drift that accumulated across the full session.
Do not stop the scan or create separate sessions for each room. SLAM performance depends on continuous trajectory data. Scan all rooms in a single continuous session where possible, and plan the route so you end near the starting point. Segmenting by room eliminates the loop closures that correct drift between spaces.
Loop-Based Routes vs. Linear Paths
The single most impactful route planning decision is whether your route creates genuine loops. SLAM drift accumulates with every meter traveled. Loop closure is the correction mechanism. A route with no loops provides no mid-session corrections, all drift accumulates until the end, when it is too large to fully correct.
Loop-Based Route
Multiple loops throughout the session. Each crossing of a previously-scanned area triggers a correction. Drift is corrected continuously rather than only at the end. Large spaces and long sessions remain accurate because errors are caught early and distributed.
Linear Route
Travels from point A to point B in one direction with no crossings. Drift accumulates the entire way. The final result may appear to close when processed, but the intermediate geometry carries uncorrected error. Dimensions measured between points far apart will be least accurate.
Ending at the starting point is not the same as creating a loop. A route that walks the full perimeter of a floor plan and returns to the start from the opposite direction does not reliably trigger loop closure. The viewing angle on return must be within 40 degrees of the original viewing angle at that location. Returning from the opposite direction fails this condition. Plan routes with genuine interior crossings, not perimeter circuits.
Creating Effective Loops
- For any scan lasting more than 10 minutes, plan at least one midway loop by returning through an already-scanned area before continuing to new areas. For large-scale sites, plan loops every 5 to 10 minutes of continuous walking.
- Use doorways as natural loop closure points. Enter and exit through the same door after covering the room, then cross the corridor to continue.
- When returning to a previously-scanned area for loop closure, approach from the same general direction you originally scanned it, not the reverse. For maximum loop closure confidence, return to the same position and face the same direction as your original pass through that area.
Midway loops are especially important in three scenarios: when the 40-degree return angle condition is difficult to meet, when scanning large-scale scenes, and when scanning spaces with significant obstructions or narrow structures.
Serpentine and Saturated Scanning
For any open area that cannot be covered by simple loops, including large rooms, parking floors, and exhibition halls, use serpentine (S-shaped) routes rather than straight parallel passes. Serpentine routes naturally generate loop closures at each turn while providing overlapping coverage from adjacent paths.
Within-Room Scan Pattern
Within rooms and open spaces, use a serpentine (S-shaped) or lawn-mowing pattern rather than random paths. Keep a consistent pace and maintain a slight scan tilt of 10 to 15 degrees up or down throughout each pass. Avoid random zigzags, abrupt vertical tilts, or height changes during a single pass. In extremely narrow corridors where an S-pattern is not physically possible, a single linear pass is acceptable.
Saturated Scanning for Complex Geometry
For areas with significant occlusion (complex machinery, dense shelving, heavy equipment) or where scanning quality is critical, use saturated scanning: cover the same area from multiple different routes and directions rather than a single pass. This is more time-consuming but produces substantially better coverage of occluded surfaces.
Unlike terrestrial laser scanning, where additional time at a station improves results, standing completely still with a mobile SLAM device adds no value after the first few seconds. If an area needs better coverage, move through it again from a different route and angle. For 3DGS (LCC Studio) workflows, controlled figure-8 motions from a near-stationary position can improve visual coverage and splat density even though the LiDAR geometry does not benefit. See 3.1 Movement and Posture for figure-8 technique details.
PortalCam bidirectional coverage: Because the PortalCam front cameras have a fixed forward orientation (unlike the 360-degree L2 Pro and K1), a single pass captures only the forward-facing perspective of each surface. Walk the same route forward and then in reverse to capture both directions. Also scan at multiple heights (low, chest-level, and overhead) for complete coverage.
Equipment and Object Coverage
Mechanical equipment, HVAC units, tanks, columns, and any object with surfaces facing multiple directions requires deliberate multi-angle coverage. A single pass captures the surfaces facing your direction of travel. Surfaces parallel to your path and surfaces facing away from you are missed entirely.
For standalone equipment and objects, circle the target at a distance of 3 to 10 ft (1 to 3 m), moving at 1.5 ft/s (0.5 m/s). Complete at least one full revolution. For complex equipment with many internal surfaces, walk figure-eights around it to capture geometry from crossing angles. For tall objects, vary your height across passes where practical and transition between heights gradually over several feet of travel rather than in a single step.
- Circle standalone objects rather than passing them once. Complete at least one full revolution at consistent distance.
- Use figure-eight walking patterns for equipment with internal complexity. The crossing paths capture geometry from intersecting angles that a simple circle misses.
- For ceiling-mounted equipment, approach from directly below and scan with the device tilted upward. Keep the tilt under 30 degrees for L2 Pro and K1.
- Do not attempt to capture a surface by tilting the device to an extreme angle. Reposition your body to face the surface instead.
Height Variation
Varying the height at which you hold the scanner during a pass adds geometry from angles that a single-height route cannot capture: tops of low partitions, undersides of mezzanines, equipment at floor level. This is particularly useful in complex industrial environments.
Height changes must be gradual. Raising or lowering the device sharply over a short distance creates an abrupt geometry and lighting change that can disrupt SLAM tracking. Make height transitions slowly and over several meters of travel, not in a single step. Extension pole accessories designed for XGRIDS devices handle height variation without requiring any special technique adjustment beyond keeping movement smooth and continuous.
Multi-height strategy for complete coverage: For thorough interior documentation, scan the same space at multiple heights in separate passes. Pass 1 at approximately 5 ft captures core wall geometry and the primary SLAM loop. Pass 2 at approximately 7 to 8 ft captures overhead views, ceiling lines, and light fixtures. Pass 3 at approximately 2 to 3 ft captures baseboards, floor textures, and under-furniture geometry. Maintain consistent height within each pass for even quality.
Site Reconnaissance Before Scanning
Walk the site without the scanner before beginning. 5 minutes of reconnaissance prevents route planning decisions made under pressure once the scan is running and the clock is ticking. On a reconnaissance walk, identify the following.
- Feature-poor zones: long corridors, open areas, and low-texture surfaces. Plan to slow down and use serpentine routes in those areas.
- Obstructions and occlusion: dense equipment, shelving, or structures that will require additional passes from multiple angles.
- Lighting conditions: very dark areas that need supplemental lighting, and bright transitional zones (indoor-outdoor thresholds) that require reduced speed.
- Reflective or transparent surfaces: glass walls, mirrors, and polished floors. Plan how to approach them (see 3.4 Challenging Environments).
- Access constraints: doors that must remain closed, areas with moving machinery, and personnel who may be present during the scan.
- Loop closure opportunities: identify where your route can naturally cross itself to create corrections.
©2026 Alpine Reality Capture LLC • XGRIDS Pro Guide™ • Site Disclaimer