Part 8: Specialized Applications

---

Quick Field Guide

Multi-Floor Buildings

Pre-Scan Preparation:

• Open all doors on all floors
• Identify and plan stairwell/elevator routes
• Decide: single scan (2-3 floors) or map fusion (3+ floors)

Single-Scan Approach:

• Initialize ground floor, central location
• Complete ground floor loop
• Slow to 0.3-0.5 m/s approaching stairs
• Scan stairwell from multiple angles before ascending
• Pause at each landing, scan from landing position
• Ascend steadily, no rushing
• Complete each floor as loop
• Descend through stairwell to close vertical loop
• End near initialization point

Map Fusion Approach:

• One segment per floor (5-10 scans per batch max)
• Include stairwell in adjacent floor segments
• 15-30m overlap in stairwell zones
• Mark control points on landings OR maintain RTK coverage
• Name segments clearly: Building_A_Floor1, Building_A_Floor2

Critical Points:

• Stairwells stress SLAM - extra coverage essential
• Do NOT scan in moving elevators
• Scan elevators as static environments only

Learn more about multi-floor buildings →

Large Outdoor Sites

Primary Strategy:

• Use RTK for continuous absolute positioning
• Verify with GCP at site corners/key features

Route Planning:

• Perimeter first (establish boundaries)
• Grid of main pathways
• Fill details with small loops returning to main paths
• Each segment: 10-20 minutes, 15-30m overlap

Segment Division:

• Use map fusion for sites exceeding 30 minutes
• 5-10 scans per batch
• Under 200 minutes total per batch
• Logical divisions: buildings, blocks, zones

Dynamic Object Management:

• Scan during low-activity periods (early morning, weekends)
• Face scanner away from active areas
• Accept some dynamic objects (filter in post-processing)

Battery Management:

• L2 Pro: ~90 minutes runtime
• Plan swaps at logical segment breaks
• Bring multiple batteries

Control Point Strategy:

• RTK primary, GCP backup at corners/features
• Verify RTK accuracy with independent GCP
• GCP spacing if RTK fails: under 100m (L2 Pro), under 50m (K1)

Learn more about large outdoor sites →

Narrow Corridors and Tight Spaces

Processing Settings:

• Enable Narrow Scene mode in LixelStudio (Special SLAM Mode)
• For corridors over 500m: break into segments

Control Point Requirements:

• Under 100m corridor (50m for K1): optional
• Over 100m corridor: control points every 50-100m

Scanning Technique:

• Hold scanner to SIDE (not front) - maintains both wall visibility
• Walk slowly: 0.3-0.5 m/s
• Vary position: close to one wall, return close to other wall
• Creates different viewpoints aiding SLAM

Best Practices:

• Accept that long linear paths accumulate drift
• Control points are critical for accuracy
• Multiple passes at different positions help

Learn more about narrow corridors →

As-Built Documentation

Accuracy Tiers:

• Rough (10cm): Basic SLAM, no RTK/GCP
• Standard (3-5cm): RTK or moderate GCP spacing (50-100m)
• High-precision (1cm): Dense GCP (30-50m spacing), total station survey

Coverage Planning:

• Confirm requirements: full building vs. specific areas
• Confirm features: all MEP, structural only, exterior only
• Don't overscan, but capture everything required

Standard Deliverables:

• Point cloud (LAS or E57 format)
• 2D floor plans extracted from cloud
• Elevation views/cross-sections
• Critical dimension measurements
• Accuracy report (if RTK/GCP used)
• Check point verification (independent measurements)

Metadata Requirements:

• Project date
• Scanner model used
• Coordinate system (WGS84 UTM zone, State Plane, etc.)
• Accuracy assessment (control point residuals, check points)
• Processing settings used

Best Practices:

• Document existing conditions, not ideal conditions
• If installation errors discovered, scan as-is, note in report
• Never alter scan data to "fix" problems
• Coordinate early with stakeholders on requirements, access, timing
• Include independent check points for accuracy validation

Learn more about as-built documentation →

Quick Decision Matrix

Application	Duration	RTK/GCP	Segments	Key Challenge
Small building (2-3 floors)	15-25 min	Optional	Single scan	Stairwell loops
Large building (4+ floors)	10-20 min/floor	RTK if exterior walls, else GCP	Map fusion	Floor connections
Outdoor site <100m data-preserve-html-node="true"	15-25 min	RTK preferred	Single or 2-3 segments	Sparse features
Outdoor site >100m	10-20 min/segment	RTK + GCP verify	Map fusion	SLAM drift
Corridor <100m data-preserve-html-node="true"	10-15 min	Optional	Single	Limited features
Corridor >100m	10-15 min/segment	GCP every 50-100m	Multiple segments	Linear drift
As-built (rough)	Per building size	None	Per building	Coverage completeness
As-built (precise)	Per building size	Dense GCP	Per building	Accuracy verification

Common Mistakes to Avoid

Multi-Floor Buildings:

• ❌ Rushing through stairwells
• ❌ Scanning in moving elevators
• ❌ Insufficient stairwell coverage from different angles
• ❌ Forgetting to close vertical loop

Large Outdoor Sites:

• ❌ Exceeding battery life without planning swap points
• ❌ No RTK backup plan for GPS-denied areas
• ❌ Segments exceeding 200 minutes combined duration
• ❌ Inadequate overlap between segments

Narrow Corridors:

• ❌ Not enabling Narrow Scene mode in processing
• ❌ Walking too fast (over 0.5 m/s)
• ❌ Holding scanner in front instead of to side
• ❌ No control points in corridors over 100m

As-Built Documentation:

• ❌ Unclear requirements before scanning
• ❌ No independent check points for verification
• ❌ Insufficient metadata documentation
• ❌ Altering scan data to match drawings

---

Full Explanatory Guide

Multi-Floor Buildings

Multi-floor building documentation presents three fundamental challenges: vertical transitions through stairwells stress the SLAM algorithm's trajectory estimation, floor-to-floor connections must be established to create a unified coordinate system, and cumulative scan duration often exceeds practical single-scan limits.

Understanding Vertical Transition Challenges

The SLAM algorithm relies on continuous feature tracking and IMU integration to estimate scanner trajectory. Stairwells create conditions that challenge both mechanisms simultaneously.

As the scanner moves through a stairwell, rapid elevation change means the LiDAR continuously observes new features while losing sight of features that were visible moments before. Unlike horizontal movement through a corridor where wall features remain visible throughout, vertical movement through stairs means first-floor features disappear as the scanner ascends, replaced entirely by second-floor features at the top.

This continuous feature turnover reduces the SLAM algorithm's ability to correct accumulated drift through feature matching. Simultaneously, rapid elevation change creates high angular rates in the IMU as the scanner tilts and rotates while climbing stairs. These high angular rates increase IMU drift accumulation during the stairwell transition.

Loop closure through stairwells provides the mechanism to detect and correct accumulated drift. If the scanner returns down through the same stairwell after completing upper floors, the SLAM algorithm can recognize features observed during initial ascent. The difference between where those features were mapped during ascent and where they appear during descent reveals accumulated drift. The algorithm distributes this drift correction across the entire trajectory, improving accuracy throughout all floors.

Single-Scan Approach for Smaller Buildings

Buildings with 2-3 floors and straightforward layouts can often be captured in a single continuous scan lasting 15-30 minutes. This approach simplifies workflow by avoiding fusion processing while maintaining adequate accuracy through proper loop closure technique.

Pre-scan preparation:

• Ensure all doors on all floors are open before initialization
• Closed doors create impenetrable barriers preventing scanner from observing spaces beyond
• Verify door status on all floors during reconnaissance

Route planning for single-scan multi-floor work follows a vertical loop pattern:

1. Initialize on ground floor in central location with good feature visibility
2. Complete full ground floor loop returning to initialization point
3. Approach stairwell, slow to 0.3-0.5 m/s
4. Scan stairwell from multiple angles before ascending (walk around stairwell landing capturing it from 3-4 different positions)
5. Begin ascent, maintaining vertical scanner orientation
6. Pause at each landing (3-5 seconds) to allow SLAM to stabilize
7. Complete each upper floor as loop returning to stairwell
8. Descend through same stairwell to close vertical loop
9. End scan near ground floor initialization point

Stairwell scanning technique:

• Before ascending: Walk around stairwell landing capturing from multiple positions
• During ascent: Maintain 0.3-0.5 m/s speed, pause at landings
• Scanner orientation: Keep scanner vertical (RTK antenna up, LiDAR facing forward)
• Body position: Hold scanner to side or front, not blocking LiDAR view of stairs
• Multiple angles: Capture each flight from at least 2-3 different positions

Floor completion loops: On each upper floor, walk a complete loop returning to stairwell before ascending to next floor. This creates loop closures at each level that help correct drift accumulated while exploring that floor.

Descent for loop closure: After completing highest floor, descend through the same stairwell used for ascent. This creates the critical vertical loop closure. Walk slowly (0.3-0.5 m/s) during descent to ensure adequate feature matching with ascent data.

Map Fusion Approach for Larger Buildings

Buildings with 4+ floors or complex layouts should use map fusion with one segment per floor.

Segmentation strategy:

• One segment per floor (typically 5-10 scans covering that floor)
• Include stairwell in adjacent segments: Floor 1 segment scans ground floor plus stairwell to Floor 2. Floor 2 segment scans Floor 2 plus stairwell (creating 15-30m overlap in stairwell zone)
• Name clearly: Building_A_Floor1, Building_A_Floor2, etc.

Control points for floor connections:

• Option A: Mark control points on stairwell landings (mark in both adjacent floor segments)
• Option B: Maintain RTK coverage through stairwell if exterior walls have windows providing satellite visibility

Processing workflow:

1. Process each floor segment independently first
2. Verify individual floor quality
3. Create map fusion batch combining all floors
4. Enable appropriate georeferencing (RTK, GCP, or both)
5. Verify floor-to-floor alignment in overlap zones

Elevator considerations:

• Do NOT scan in moving elevators (creates tracking failures)
• Scan elevators as static environments: Enter elevator, stop scan, close doors, open doors on target floor, resume scan
• Document elevator geometry with separate static scans if needed

Large Outdoor Sites

Large outdoor sites including construction sites, infrastructure projects, industrial facilities, and campus environments require different workflow strategies due to sparse natural features, large distances, variable terrain, and weather exposure.

Understanding Outdoor Scanning Challenges

Feature density differences: Interior spaces provide abundant geometric features (walls, corners, doorways, equipment) creating rich point cloud geometry for feature tracking. Outdoor spaces often present large expanses of relatively uniform surfaces (pavement, gravel, grass, dirt) providing minimal geometric variation.

Distance and drift: A 10,000 sq ft building interior might be captured in 15 minutes walking at 1 m/s. A comparable outdoor area might require 30-40 minutes because the perimeter is larger and route must cover more distance. The combination of sparse features and long distances creates conditions where accumulated drift can reach unacceptable levels.

RTK solution: RTK positioning addresses outdoor drift by providing continuous absolute position measurements throughout the scan. Instead of relying solely on relative position estimates from SLAM, the scanner measures absolute coordinates from satellite signals every second. These measurements constrain trajectory, preventing drift accumulation beyond RTK accuracy specification (~3-5cm horizontal).

RTK as Primary Strategy

When to use RTK: Consider RTK the default approach for outdoor sites exceeding 100 meters in any dimension. Smaller sites might achieve acceptable accuracy with pure SLAM if feature density is adequate, but larger sites benefit significantly from RTK's drift prevention.

RTK network selection: Choose correction service provider covering project geographic area with adequate base station density (national networks, regional networks, or commercial providers). Account setup requires subscribing and obtaining login credentials.

RTK coverage planning: Examine site for areas where satellite signals may be blocked (buildings, dense tree canopy, overhead structures). RTK requires clear sky view from ~10° elevation upward. Areas showing blue or red RTK light indicate GPS-denied zones requiring alternative strategies.

Hybrid approach: For sites with partial GPS denial, use RTK in open areas and supplement with ground control points in GPS-denied zones. This provides absolute positioning accuracy across entire site despite satellite visibility variations.

Hierarchical Route Planning

Perimeter-first strategy: Scan site boundary before addressing interior areas. The perimeter route creates:

• Outer boundaries preventing disorientation during interior scanning
• Longest-distance loop closure in dataset
• Context for interpreting interior features

Walk outermost accessible boundary creating closed loop. Maintain RTK fixed solution throughout (stay several meters from buildings/trees to maintain sky view).

Grid approach: After perimeter, create network of main pathways through site. Identify primary vehicle routes, pedestrian paths, or logical grid lines dividing site into sections. Scan these pathways systematically, creating framework of scanned corridors segmenting site into zones.

Detail loops: Fill interior zones between grid lines with detail loops. Start from grid pathway, explore interior zone, return to same grid point before beginning next loop. These frequent returns provide regular loop closure opportunities.

Segmentation for Very Large Sites

When to segment: Sites requiring >30 minutes continuous scanning should divide into segments for map fusion processing.

Segment boundaries: Follow logical divisions (buildings, functional areas, ~500m sections for linear infrastructure). Each segment should target 10-20 minutes scanning time.

Overlap requirements: Must include 15-30m overlap with adjacent segments. Overlap zones should occur in areas with:

• Good RTK coverage
• Rich geometric features (buildings, equipment, vegetation, terrain features)
• NOT featureless parking lots or open grass fields

Control points in overlaps: Use either RTK positioning as primary with GCP providing verification, or GCP as primary if RTK unreliable. Place 2-4 ground control points at each overlap zone, marked during scanning of both adjacent segments.

Naming conventions: Site_A_Segment_01, Site_A_Segment_02, etc. Including area identifiers prevents confusion when managing multiple site areas.

Managing Dynamic Objects

Timing: Scan during low-activity periods (early morning before work crews arrive, weekend scanning at industrial facilities).

Scanner orientation: Face LiDAR away from active areas when possible (doesn't eliminate dynamic objects but reduces duration LiDAR dwells on them).

Walking speed: Can increase slightly to 1.2 m/s in zones with heavy dynamic activity (reduces time spent near moving objects). Don't exceed 1.5 m/s as this degrades SLAM performance.

Accepting presence: Active sites will inevitably capture some vehicles, people, and equipment in motion. Document in project notes that scan was conducted during active operations. Post-processing can filter some objects but complete removal is impossible.

Battery Management

Capacity monitoring: L2 Pro provides ~90 minutes continuous operation. Plan battery change before capacity drops below 20% (change at ~70-75 minutes into scanning).

Change locations: Coincide with logical segment breaks, not mid-segment. End one segment, change battery, start next segment.

Multiple battery strategy: For full-day outdoor work, carry 2-4 spare batteries. Each provides ~60-75 minutes usable scanning time. Four hours scanning needs at least 4 batteries total.

Cold weather: Battery capacity may decrease 30-50% in freezing conditions. Plan shorter segments and more frequent changes. Keep spares insulated until needed.

Narrow Corridors and Tight Spaces

Tunnels, underground utilities, long hallways, mine drifts, and similar confined linear spaces present unique SLAM challenges requiring specialized processing modes and field techniques.

The Narrow Scene SLAM Challenge

Normal SLAM algorithms expect features distributed in all directions around the scanner. In wide-open rooms or outdoor areas, the LiDAR observes walls, objects, and terrain in 360 degrees horizontally and from floor to ceiling vertically.

Narrow corridors violate this assumption. In a 2-meter-wide tunnel, the LiDAR only observes features to the left and right (the tunnel walls). Forward and backward directions show empty tunnel extending to limits of scanner range. Vertical directions show floor and ceiling but these are often featureless smooth surfaces.

This limited lateral feature distribution reduces the number of geometric constraints available for SLAM trajectory estimation. Mathematically, the algorithm has strong constraints in the lateral direction (perpendicular to the tunnel axis) from the side walls, but weak constraints along the tunnel axis. This asymmetry in constraint strength makes the trajectory estimation poorly conditioned, leading to drift accumulation primarily along the tunnel direction.

Narrow Scene Processing Mode

LixelStudio includes a specialized "Narrow Scene" SLAM mode designed for corridors, tunnels, and mines. This mode adjusts the SLAM algorithm to:

• Be more conservative about position estimates along the weak axis
• Increase reliance on IMU data for forward motion estimation
• Adjust feature matching parameters for limited lateral geometry
• Modify loop closure detection thresholds

When to use: Enable Narrow Scene mode for:

• Tunnels and mines
• Long corridors with plain walls (<2m data-preserve-html-node="true" wide)
• Underground utilities
• Linear spaces under 500 meters length

How to enable: LixelStudio → Project Processing → Advanced Settings → Special SLAM Mode → Select "Narrow Scene"

Limitations: For corridors/tunnels longer than 500 meters, break into segments with control points every 500m or less. Even Narrow Scene mode cannot prevent drift accumulation over extremely long linear paths.

Field Scanning Technique for Narrow Spaces

Scanner position: Hold scanner to SIDE (not in front of body). This maintains visibility of both walls simultaneously. If scanner is held in front, your body blocks view of one wall, reducing feature availability.

Walking speed: Reduce to 0.3-0.5 m/s (slower than normal 1.0 m/s). Slower movement:

• Increases point density on walls
• Provides more time for SLAM to identify subtle features
• Reduces IMU drift contribution

Vary walking position:

• Walk close to left wall for first pass
• Return close to right wall for second pass
• Creates different viewpoints and observation angles
• Provides SLAM with more geometric diversity
• Improves feature triangulation

Multiple passes: For critical corridors, scan multiple times:

• First pass: One direction along left side
• Second pass: Return along right side
• Third pass (optional): Center position for additional coverage

Control Point Requirements

Under 100m corridor (50m for K1): Control points optional if Narrow Scene mode used and multiple passes captured.

Over 100m corridor: Control points critical. Place every 50-100m to constrain drift accumulation. Linear spaces naturally accumulate drift along the axis; control points provide external position references that correct this drift.

Control point placement: Place on corridor floor or walls at regular intervals. Mark during scanning using standard control point procedure. Ensure adequate point cloud coverage around each control point (dwell 5 seconds, circle if space permits).

As-Built Documentation

As-built documentation captures existing physical conditions of buildings, infrastructure, or facilities for design, renovation, regulatory compliance, or facilities management purposes.

Establishing Accuracy Requirements

As-built projects span wide accuracy spectrum. Confirming accuracy requirements with client before beginning work determines appropriate georeferencing method and workflow approach.

Rough documentation (10cm accuracy):

• Suitable for: Space planning, furniture placement, general renovation budgeting, preliminary design
• Method: Basic SLAM without RTK or GCP
• Achievement: Single well-executed scan with good loop closures
• Represents: Minimum acceptable tier for professional work

Standard as-built (3-5cm accuracy):

• Suitable for: Detailed renovation design, MEP coordination, structural assessment, construction verification
• Method: RTK positioning OR ground control points at 50-100m spacing
• Achievement: RTK provides 3-5cm with minimal field effort; GCP provides similar accuracy with more surveying effort but works in GPS-denied environments

High-precision (1cm accuracy):

• Suitable for: Industrial equipment installation, manufacturing facility documentation, historic preservation, structural monitoring baseline
• Method: Dense ground control point networks at 30-50m spacing surveyed using total station or static GPS
• Achievement: Dense control constrains SLAM drift aggressively

Critical dimensions: Some projects specify looser overall accuracy but require tighter tolerances on specific features (structural column locations, equipment anchor points). Verify these using independent measurement methods (total station, tape measurements) rather than relying solely on point cloud extraction.

Defining Coverage Requirements

Full building coverage: Captures every accessible room, corridor, stairwell, and exterior surface. Requires systematic route planning ensuring no areas missed while maintaining efficient scan paths.

Document which areas were scanned versus which were inaccessible. Buildings almost always contain some locked areas, occupied private spaces, or hazardous areas that cannot be accessed. Final deliverable should include coverage map showing captured areas and exclusions.

Specific area coverage: Limits scanning to designated portions (one floor, one wing, mechanical spaces only, exterior facades only). Coverage boundary definition must be very clear before scanning begins. Obtain marked floor plans showing exactly which rooms/areas should be scanned.

Edge cases: Clarify boundary details during scoping. If scanning first floor of three-floor building, how much of stairwell to second floor should be captured? If scanning exterior facades, should window reveals and overhangs be captured or only outer facade plane?

Feature-specific coverage: Focuses on particular building systems rather than complete spatial documentation. MEP documentation might capture only systems suspended from ceilings and mounted on walls, ignoring room geometry. Structural documentation might capture columns, beams, load-bearing walls while ignoring partitions and finishes.

Standard As-Built Deliverables

Point cloud data (foundation deliverable):

• Formats: LAS (uncompressed), LAZ (compressed, smaller file), E57 (vendor-neutral supporting metadata), RCP (Autodesk ReCap)
• Content: RGB color if camera used, intensity values (always present)
• Coordinate system documentation: Local arbitrary origin, WGS84 UTM zone XX, State Plane Zone XXXX, or other system

2D floor plans: Extracted using LixelStudio Horizontal Slice tool. Create cross-sections at each floor level (typically 1.2m above floor). Export slices, trace walls using AI-Extract and manual drawing tools, export as DXF for CAD software.

Elevation views and sections: Created using Vertical Slice or Profile Analysis tools. Show exterior facades, interior wall elevations, or building cross-sections. Export as DXF for integration with architectural drawings.

Critical dimension measurements: Specific measurements extracted from point cloud for verification purposes. Typical: floor-to-floor heights, structural column spacing, equipment clearances, door/window dimensions, ceiling heights. Compiled into tables in documentation package.

Independent verification: Select subset of critical dimensions (~20%) and measure independently using tape measures or laser distance meters. Compare field measurements to point cloud measurements. Agreement within project accuracy tolerance validates data quality.

Accuracy reports (when RTK or GCP used):

• Coordinate system definition
• Georeferencing method (RTK, PPK, GCP)
• Number and distribution of control points
• Control point residual statistics
• Check point residuals if established
• For RTK: percentage of scan with fixed solution, maximum RTK gaps and locations, overall achieved accuracy estimate

Metadata documentation:

• Project name and address
• Scan date or date range
• Scanner model used
• Weather conditions (outdoor scanning)
• Processing software and version
• Who performed scanning and processing
• Special conditions or limitations

Best Practices for Professional As-Built Work

Early stakeholder coordination: Meet with client, architect, engineer, facility manager, or other stakeholders before mobilizing. Review project scope, accuracy requirements, coverage areas, deliverable formats, schedule constraints, site access procedures, and safety/security requirements.

Document as-is conditions: Scanner operator's role is accurate documentation of reality, not interpretation or correction. If wall is 10cm out of plumb, scan should capture it 10cm out of plumb. If equipment installed in different location than design drawings, scan should show actual installation location.

Resist pressure to "fix" scan data: If stakeholders discover discrepancies between scan data and design documents, document those discrepancies in project notes and reports, but do NOT alter scan coordinates or geometry to force agreement with drawings.

Problem discovery documentation: When scanning reveals issues (structural damage, equipment misalignment, code violations, missing components), note and report but do not affect scan data itself. Create field notes documenting discovered issues. Take photographs if appropriate. Inform client about significant discoveries. But do not alter point cloud to compensate.

Quality assurance through independent verification: After processing, select 5-10 easily accessible locations throughout scanned area. Return to site with tape measure or laser distance meter and measure distance between clearly defined features. Compare field measurements to corresponding measurements extracted from point cloud. Agreement within accuracy specification validates scan quality.

Check point distribution: Identify 10-20 specific locations that will be surveyed independently using total station or high-quality RTK GPS. Mark during scanning like control points but do not use in georeferencing transformation. After processing, measure coordinates of check points in point cloud and compare to surveyed values. Residuals provide objective accuracy assessment.

Distribute check points throughout project area rather than clustering. For multi-floor buildings, include check points on each floor. For large sites, include check points near perimeter and in interior.

Professional reporting: Package deliverables with appropriate documentation enabling client to use data effectively. Complete package includes:

• Point cloud file in specified format
• 2D plan views as DXF if required
• Elevation and section views as DXF if required
• Critical dimension table as spreadsheet or PDF
• Accuracy report documenting georeferencing quality
• Metadata document describing project details
• Coverage map showing captured areas
• Field notes documenting observations or limitations

Organize in clear folder structure. Use descriptive file naming: ProjectName_Floor1_Plan.dxf rather than plan1.dxf. Include date stamps for multi-date projects.

Archival copies: Maintain all project data separately from client deliverables. Retain raw scanner data files, processing project files, control point survey data, field notes, and photographs for at least 2 years after project completion.