2.1 SLAM vs. Static Scanning

How Traditional Static Scanning Works

Traditional terrestrial laser scanning (TLS) uses a scanner mounted on a tripod in a fixed position. The scanner rotates and fires laser pulses in all directions, capturing a dense sphere of point data around that single point. When coverage of a large space is needed, the scanner is moved to a new position and the process repeats. Every position becomes a separate scan.

The scans are then registered together in post-processing by identifying common reference targets or overlapping geometry between adjacent setups. The quality of the final dataset depends largely on the density of scan positions and the accuracy of the registration between them.

For a typical commercial building floor, a TLS workflow might involve 20 to 100 or more scan positions, each requiring the operator to stop, level the instrument, wait for the scan to complete, and move to the next position. The work is methodical and predictable. Errors tend to be localized to specific setups or specific registrations.

How XGRIDS Mobile SLAM Works

XGRIDS devices use Simultaneous Localization and Mapping, or SLAM. Instead of capturing data from fixed positions, the scanner moves continuously through the space while capturing. The system uses three data sources working together in real time: LiDAR (laser ranging), visual cameras, and an IMU (inertial measurement unit). These three sources are fused continuously to track the scanner's position as it moves and build the point cloud simultaneously.

A single operator walking a planned route through a 50,000 sq ft building can capture the entire floor in one continuous session. There are no tripod setups, no scan positions to plan, and no registration targets to place. The output of a single session is a single continuous dataset.

The tradeoff is that the system's ability to stay accurately localized depends entirely on having enough environmental features to track against. When the scanner moves through a feature-poor area, a long blank corridor, a large open room with no furniture, tracking quality degrades. This is fundamentally different from TLS, where feature-poor environments slow you down but do not affect data quality.

The other key mechanism is loop closure. As drift accumulates over distance, the system detects when it has returned to a previously scanned area and uses that recognition to measure and correct the accumulated error. Without loop closure opportunities built into the scan route, drift compounds without correction. This is why route planning is the single most important field skill in mobile SLAM.

What XGRIDS Devices Capture

All four current XGRIDS devices (L3, L2 Pro, K2, and PortalCam) use the Multi-SLAM system combining LiDAR, visual cameras, and IMU. The LiDAR provides geometric measurement. The cameras provide visual tracking features and, in the case of the L3, L2 Pro, and K2, color data for the point cloud. The IMU provides continuous motion sensing between LiDAR pulses. All three operate simultaneously from the moment scanning begins.

A real-time preview of the developing point cloud is visible in LixelGO (L3, L2 Pro, K2) or LCC Scan (PortalCam) during scanning, allowing the operator to monitor coverage and identify gaps before leaving the site. This is one of the most important practical advantages of mobile SLAM: you know what you captured before you walk out the door. The L3 shows capture status, a model preview and a real-time quality check on its built-in 4.3-inch display.

How Planning Changes

The most consequential planning decision is the route. A well-planned SLAM route creates multiple loop closure opportunities throughout the scan. A poorly planned route, typically a linear path from one end of a building to the other, accumulates drift continuously with no correction.

The second major planning difference is the coordinate strategy. In TLS, georeferencing is often applied after scanning by transforming the registered point cloud to a coordinate system using control points. In XGRIDS mobile SLAM, the georeferencing strategy must be decided before scanning begins. RTK or PPK must be configured and connected before the scan starts. GCPs must be physically placed and marked during the scan. A scan captured without georeferencing cannot be retrospectively tied to a real-world coordinate system.

How Technique Changes

In TLS, the operator's physical technique during scanning is largely irrelevant. The scanner is on a tripod. The operator waits.

In mobile SLAM, the operator is the instrument platform. How you move, how fast, how you hold the device, and where you direct it all affect the quality of the data. This is the biggest adjustment for anyone coming from a TLS background.

The Key Technique Requirements

1. Initialization Before Moving

   Every session begins with a stationary initialization. The scanner must remain completely still during the 15-second countdown. Any movement during this period corrupts the starting reference frame and the error persists for the entire session. TLS has no equivalent to this step.

2. Controlled Walking Speed

   Walking too fast reduces point density and degrades visual tracking. The system needs adequate dwell time to capture sufficient geometry at each location. The general maximum is 3.3 ft/s. In indoor environments, dim lighting, doorways, and corners, reduce to 1.6 ft/s.

3. Device Posture

   The L2 Pro should be held within 20 degrees of vertical along the walking direction, and the K2 within 15 degrees. Either may briefly reach 30 degrees in small spaces or for ground targets. LixelGO displays a real-time posture indicator. Excessive tilt affects IMU accuracy and point cloud geometry.

4. Full-Body Turns

   When changing direction, rotate your entire body rather than twisting just your arms. Arm-only rotation introduces angular inconsistency in the trajectory that appears as distortion in the point cloud at corners and doorways.

5. Safe Shutdown

   End the scan using the app's End Recording button or the device's stop function (double-press the power/device button). Then wait for the LED to reach solid green, confirming the save is complete. Powering off while data is writing corrupts the session and the data is not recoverable.

None of these requirements exist in TLS.

A Note for Experienced TLS Users

If you are coming from a TLS background, you already understand coordinate systems, registration, accuracy specifications, and delivery workflows. What requires deliberate adjustment is the error model and the field execution approach.

In TLS, errors are local. A bad setup affects that position's data. A failed registration affects the connection between two positions. Problems are isolated and often correctable in post-processing.

In mobile SLAM, errors compound. A bad initialization affects the entire session. A route without loop closure allows drift to accumulate across the entire dataset. These errors are difficult or impossible to correct after the fact. Prevention in the field is the only reliable strategy.

The practical implication: experienced TLS operators sometimes scan too quickly on their first XGRIDS project because they are accustomed to a workflow where speed in the field has no quality cost. With mobile SLAM, moving through a building faster than the system can reliably track produces drift and gaps.

For the complete field technique reference, including device-specific posture, route planning patterns, and speed guidelines, see Module 3: Field Technique.

Next: the two output types XGRIDS produces and why the choice between them must be made before you scan.

The Two Workflows →