Module 3 Quick Field Guide

Speed and Posture Reference

Speed and posture are device-dependent. The L2 Pro and K2 share most rules. The PortalCam has different tolerances due to its multi-camera array and visually-dependent SLAM.

Walking Speed (All Devices)

• Outdoor maximum: 3.3 ft/s (1 m/s) in open feature-rich areas
• Indoor target: 1.6 ft/s (0.5 m/s) for all indoor environments, dim lighting, doorways, corners, and feature-poor areas
• Tunnels and long corridors: 1.6 ft/s (0.5 m/s) or below throughout
• Doorway approach: begin slowing 6 to 10 ft (2 to 3 m) before the threshold
• Reflective surfaces: 1.6 ft/s (0.5 m/s) past any glass or mirror

Walking too fast reduces point density and degrades visual tracking. The system needs adequate dwell time at each location to accumulate sufficient geometry. There is no in-app warning for moderate excess speed: degraded accuracy is discovered during processing or after delivery.

Figure-8 Scanning Technique

Use controlled figure-8 motions with the scanner to improve SLAM tracking and Gaussian Splat density. Keep the movement fluid and deliberate. Avoid erratic hand motions.

Route Structure Checklist

• Walk all main corridors first before entering any rooms. Establish the spine of the floor plan before branching
• Branch into each room from the corridor. Scan the room interior completely, then return through the same doorway you entered. Each room entry and exit is a loop closure
• Within rooms and open spaces, use a serpentine (S-shaped) or lawn-mowing pattern rather than random paths. This maximizes multi-angle coverage. Exception: extremely narrow corridors where an S-pattern is not physically possible
• Close loops continuously. Return to previously scanned areas at the end of every major space. For maximum loop closure quality, re-enter the previously scanned area from the same direction and position you originally used. Retracing from the same spot and facing the same direction gives SLAM the highest-confidence match
• For corridors over 160 ft (50 m) (SLAM-only): branch into at least one room and return before reaching the far end. A straight out-and-back walk without branching accumulates uncorrected drift. With RTK Fixed status, corridors up to 500 ft (150 m) are viable because drift is bounded by the coordinate stream
• Return to within 15 to 30 ft (5 to 10 m) of the initialization point at the end of the session from a similar angle to your starting position. This is the final loop closure
• 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. Start at the main entry, move through all rooms in one continuous loop, and finish where you started

Transition Procedures

Doorways, stairwells, and confined-space exits are the highest-failure transitions in indoor scanning. The procedures below apply to all three devices unless explicitly noted otherwise.

Interior Door Transitions (Door Pre-Opened)

• Begin slowing 6 to 10 ft (2 to 3 m) before the doorway. Target speed: 1.6 ft/s (0.5 m/s)
• Leave the door half open. A fully open door leaves one side of the frame unscanned. A half-open door captures both sides
• Turn 90 degrees sideways before stepping through the threshold. One side of the scanner faces back into the room you are leaving, one side faces into the new room
• Pause on each side of the doorway threshold until the LixelGO or LCC Scan preview shows point cloud data from both rooms. Typically a few seconds per side. Do not use a fixed timer; the required duration depends on the geometric complexity of each space
• Resume normal speed 6 to 10 ft (2 to 3 m) past the doorway

Closed Door Transitions

• If the door cannot be pre-opened, slowly turn so the scanner faces away from the door. Your body must block the LiDAR's line of sight to the closed door surface
• Open the door with your back to it. The scanner must not see the door opening, which would introduce a rapidly changing geometry that degrades SLAM tracking
• Stand sideways under the doorframe and pause until point cloud data from both rooms appears in the preview
• Enter the new room slowly and resume normal scanning

Stairwells

• Scan the landing at the base fully before beginning the ascent
• Walk up one side of the staircase, then return down the other side for bidirectional coverage
• Angle the scanner slightly to capture the underside of the flight above you as you ascend. Stay within 30 degrees of vertical
• Scan each landing fully on both the ascent and descent
• For maximum coverage, include one pass walking sideways (crab-walking) up or down the stairwell. This provides the LiDAR with a different geometric perspective on tread and riser geometry
• Budget 5 to 10 minutes for a standard three-flight stairwell with proper coverage. Complex stairwells with wide landings or multiple branches may take longer. Do not rush stairwells; they are a high-failure zone and time-consuming to rescan

Exiting Confined Spaces

When exiting a confined space, align the scanner's field of view with feature-rich areas (walls with geometry, equipment, doorframes). Do not simply turn around and exit. Abrupt reversal without maintaining LiDAR contact with structural features causes reference loss and SLAM misalignment.

Environment-Specific Rules

Long Corridors and Tunnels

• Maximum 1,600 ft (500 m) per scan segment. Divide longer runs before starting
• Enable Narrow Scene mode in LixelStudio before processing any tunnel or long corridor scan
• For multi-segment tunnels, plan a shared overlap point between sessions. When marking the shared point in the second scan, device position must be within 4 in (10 cm) and 10 degrees of its position in the first scan
• Minimum spacing: 15 ft (5 m) between any two shared points
• Multi-segment control point placement: the first segment gets two control points at the end only. The last segment gets two at the start only. All middle segments need two at the start and two at the end (four total). Give the same physical location the same point ID in every segment
• Slow to 1.6 ft/s (0.5 m/s) or below throughout. In low-light tunnels, slower is better
• Place high-contrast SLAM anchor targets every 100 to 150 ft (30 to 45 m) in long featureless runs. The L2 Pro ships with magnetic steel targets that attach to ferromagnetic surfaces. The K2 ships with 30 reflective sticker targets that adhere to any flat surface. Both serve the same SLAM-anchor function; neither is a GCP and neither requires surveyed coordinates. They give the system reliable high-contrast features to match between passes. If you also need absolute coordinates, supply those targets with surveyed coordinates in LixelStudio, at which point they become GCPs

Reflective Surfaces (Glass and Mirrors)

• Never stop in front of glass or mirrors. Keep moving at 1.6 ft/s (0.5 m/s) past any reflective surface. Maintain at least 3.3 ft (1 m) distance from highly reflective objects
• Approach glass walls at an angle rather than head-on. An angled approach reduces direct reflection exposure to the LiDAR
• Keep the scanner out of mirror reflections by positioning your body between the scanner and the mirror. Stand with your back to the mirror when possible
• In bathrooms and vanities, scan low over counters or approach from side angles to avoid capturing the scanner's own reflection
• Capture mirror and glass surfaces in a single sweep. Multiple passes in front of the same reflective surface introduce artifacts
• For fully-glazed office facades, plan your route to cross glass zones rather than walk parallel to them at close range
• PortalCam is more severely affected by reflective surfaces than the L2 Pro or K2 because the visual component of its SLAM system can track reflections as false geometry

Dark Environments

• Slow to 1.6 ft/s (0.5 m/s) or below in any dark area
• Bring a portable light panel and angle it forward into the area you are walking toward, not downward at the floor at your feet. SLAM is continuously matching new geometry against what it has already seen. The upcoming area needs to be illuminated before you reach it, not after
• Camera coloring in dark environments produces poor results. Consider disabling coloring for dark spaces and re-enabling it for lit areas, or plan dark zones as separate no-color sessions
• Place high-contrast SLAM anchor targets in dark featureless areas before scanning. L2 Pro magnetic steel targets or K2 reflective sticker targets both serve this role. Even without surveyed coordinates, these targets give SLAM high-contrast features to match against when visual texture is minimal

Outdoor Scanning

• Outdoor areas require consistent feature density. Do not scan open plazas or parking lots without including surrounding buildings, walls, or fixed structures in the scan path
• Do not point cameras at the sun or scan into direct sunlight. Overexposure corrupts camera coloring data
• Wind affects PortalCam stability and visual tracking. In windy conditions, walk with the wind rather than against it where possible
• L2 Pro 32-300 is required for tall structures and large outdoor sites. The 16-120 and 32-120 L2 Pro models are limited to 400 ft (120 m) range. The K2 has a 100-meter LiDAR maximum range, making it appropriate for indoor and close-quarters outdoor work but not for tall facades or large open exteriors

Featureless Corridors

• In narrow, uniform hallways, walk sideways (side shuffle). This keeps both walls in the LiDAR field of view simultaneously, which is especially important in featureless corridors where SLAM has fewer anchors
• Avoid scanning long blank walls with no features. Use diagonal paths to increase perspective variation when a side shuffle is not practical

Data Centers

Data centers are scanned repeatedly over time for change tracking, equipment audits, and capacity planning. The challenge is repetitive geometry: row after row of identical racks with no distinguishing features. Scan quality and target placement here directly affect how usable rescan comparison is.

• Speed: 1.6 ft/s (0.5 m/s) maximum in all aisles. The proximity of rack faces on both sides creates a low-feature environment throughout
• Walk every aisle. Do not skip cold aisles if hot aisles were already covered. Each aisle provides different rack face geometry
• Vary scanner height between aisles where possible. Walk one aisle at chest height and the adjacent aisle at overhead height. Change heights gradually over several feet of travel
• Complete loop closures at the end of every row by looping back through the cross-aisle at the far end before starting the next aisle
• Space targets every 100 to 165 ft (30 to 50 m) along the scan path, or one per 3 to 5 aisles. L2 Pro magnetic steel targets attach directly to rack panels and are the preferred target type in data center applications. K2 reflective sticker targets are functional in this role but require adhesive placement and may leave residue on rack surfaces. Neither type is a survey GCP. Both serve as high-contrast SLAM anchors in this repetitive-geometry environment
• Place each target where the scanner can approach within 3 to 6 ft (1 to 2 m) and circle it during the scan
• Keep targets away from hot-aisle vents and active airflow paths. Turbulent air near cooling exhaust can cause vibration. For L2 Pro, magnetic attachment directly to a steel rack panel is the most stable mount. For K2, choose sticker placement on flat panel faces away from airflow paths; an adhesive sticker on a vibrating surface can shift mid-scan
• Mark every target in LixelGO during the scan. Unmarked targets are visible in the point cloud but are not used for SLAM correction or georeferencing in processing
• Place targets in the same positions as the original scan. For recurring rescans, record target locations permanently with floor coordinates or photos
• Plan rescan session boundaries so the new session covers at least 50 ft (15 m) of geometry that is also present in the session it must merge with
• Check the LixelGO preview before leaving site. Visible kinks in aisle lines or discontinuities at pod boundaries indicate a tracking failure that will require recapture

Critical Warnings

• Bad data Moderate excess speed produces no immediate warning in the app. LixelGO does display a "Record Warning Only" alert when SLAM confidence drops severely, but normal excess speed degrades data silently. Degraded accuracy is discovered during processing or after delivery. Slow down before the environment forces you to
• Bad data Loop closure requires a return viewing angle within 40 degrees of the original. Walking back through a room facing the opposite wall is not a loop closure. SLAM will not detect the overlap and drift continues accumulating
• Bad data Do not allow objects to block more than 50% of the LiDAR field of view within 3.3 ft (1 m). Sustained obstruction from your body, carried equipment, or close obstacles degrades point cloud accuracy. Reposition your carry posture or route to maintain a clear field of view
• Warning Do not stop the scan or create separate sessions for each room. SLAM performance depends on continuous trajectory data. Segmenting by room breaks continuity and degrades alignment between spaces
• Warning Tunnels and long corridors over 1,600 ft (500 m) must be segmented before scanning. A continuous scan beyond this length is not recoverable in post-processing. Plan segments at 1,600 ft (500 m) or less
• Warning Enable Narrow Scene mode before processing any tunnel or long corridor scan. Processing tunnel data with the default SLAM mode produces significantly degraded results. This setting is in LixelStudio, not in LixelGO
• Warning Stairwells take longer than their footprint suggests. A standard three-flight stairwell with correct technique requires 5 to 10 minutes. Complex stairwells with wide landings take longer. Allocate time before arriving on site
• Warning Standing completely still adds no data after the first few seconds. For LiDAR processing (LixelStudio), only physical movement through space generates new geometry. Circle objects at 1.6 ft/s (0.5 m/s) instead. For 3DGS processing (LCC Studio), controlled figure-8 motions from a near-stationary position can improve visual coverage, but standing motionless does not improve reconstruction
• Warning Do not point cameras at the sun or direct light sources. Overexposure corrupts camera coloring data for the entire scan, not just the overexposed frames. Avoid scanning toward windows in bright sunlight
• Warning When exiting confined spaces, do not simply turn around and walk out. Maintain LiDAR line of sight to feature-rich geometry (walls, doorframes, equipment) as you exit. Abrupt reversal without structural features in view causes SLAM reference loss