Start Here

Two Core Workflows

XGRIDS produces two fundamentally different output types. The workflow you choose determines which software you use, how you process your data, and what your deliverable can and cannot do. These outputs cannot be converted between each other after the fact, and confusing the two pipelines is one of the most common early errors.

On many projects, both workflows run in parallel. The same raw capture data from the L2 Pro or K1 feeds both pipelines. Plan for both before you scan if both are needed. The decision cannot be made after the fact. See 2.2 The Two Workflows for a full comparison.

Software

Four applications, two pipelines. Each has a specific and non-interchangeable role. Sending data to the wrong application will either produce an error or a result that does not match the deliverable requirement. See 2.4 Software Map for the full ecosystem overview.

Which App Goes With Which Device

Processing Time and RAM Requirements

Undersized hardware causes mid-run failures, not errors you can recover from easily. Set these expectations before every project. See 2.5 Critical Basics for the full requirements list.

LixelStudio holds the entire scan trajectory in memory during SLAM optimization. If your project requires scans longer than your RAM supports, break them into shorter sessions and use map fusion to stitch the segments in post-processing.

Before Your First Real Scan

Most first-job failures are preparation failures. The scanner worked fine. The operator left site with unusable data. Run through this before every project. The complete procedure is in Module 1: Getting Started.

Pre-Field Checklist

• LixelStudio installed, activated, and opened at least once on the processing computer
• LCC Studio installed if Gaussian Splat output is required. NVIDIA GPU driver version 520.0 or later confirmed
• LixelGO or LCC Scan installed on your mobile device and signed in with your XGRIDS credentials
• Scanner batteries fully charged (3 to 4 hours from empty)
• Storage confirmed on scanner. K1: TF card. L2 Pro: 1TB SSD. Budget approximately 60 to 80 GB per hour of scanning
• Processing computer storage confirmed. Budget approximately 50 GB per project including workspace
• Coordinate system and georeferencing method decided before leaving for site
• RTK module installed and NTRIP credentials entered and tested if using RTK
• Ground control points physically placed, surveyed, and coordinate CSV prepared if using GCPs
• GCP names in the CSV confirmed to match exactly what you will type in LixelGO. Case-sensitive
• Site walkthrough completed: access points, hazards, scanning route, known GNSS obstructions identified
• Mobile device charged. Portable power bank packed for sessions over 60 minutes

Decide Your Coordinate Strategy Before You Scan

This decision cannot be made after the fact. RTK must be configured before scanning begins. GCPs must be placed and marked during the scan. A scan captured without georeferencing cannot be retrospectively tied to a real-world coordinate system. See 4.1 When Georeferencing Matters before committing to an approach.

Ground Control Point Rules That Matter in the Field

• L2 Pro: spacing at most 100 m, evenly distributed. K1: at most 50 m, evenly distributed
• Points cannot all be collinear. Three-dimensional distribution is required
• GCP names in the coordinate CSV must exactly match what you type in LixelGO, character for character, including capitalization
• Set the scanner down gently when marking. Dropping or jarring the device at the marker reduces accuracy at that point
• After marking, circle the point 1 to 2 times or remain nearby for a few seconds to capture additional geometry around the marker
• If ending the session immediately after marking a point, wait at least 15 seconds before stopping the recording

The complete GCP procedure is in 4.5 Control Point Marking During Scanning.

Field Technique

SLAM accuracy is earned in the field. Post-processing cannot compensate for poor technique. These principles directly determine the quality of the data you bring back. The full technique reference is Module 3: Field Technique.

The Concept That Changes Everything

SLAM drift accumulates over distance. Every meter you travel, a small amount of positional error compounds. The correction mechanism is loop closure: returning to an area you have already scanned so the system can detect the match, measure the accumulated error, and distribute the correction back across the trajectory.

Initialization

The initialization sets the origin and calibrates the SLAM starting reference. Errors introduced here persist for the entire session. Do it correctly every time. See 3.1 Movement and Posture Fundamentals for the complete procedure.

Walking Speed

Standing still does not improve scan quality. The system needs motion to generate coverage. Move continuously, just slowly where it matters. See 3.3 Transitions: Doors, Stairs, and Between Spaces for specialized guidance.

Device Posture

• Hold vertically at chest to shoulder height. Tilt must stay under 20 degrees from vertical. LixelGO shows a real-time posture indicator
• In corridors narrower than 2 meters, turn the device sideways so the LiDAR captures both walls simultaneously
• Maintain at least 0.5 meters clearance from walls and obstacles during normal movement
• Rotate your whole body when turning. Twisting only your arms creates angular inconsistency in the trajectory

Route Planning

• Scan backbone corridors first, then branch into individual rooms. Return to the corridor between each room, not directly to the next room
• For large projects, build midpoint loops into the route, not just a final return to start
• For complex or narrow environments, plan at least 4 loop intersections throughout the session
• Return to your initialization point at the end of every session when possible
• For corridors or tunnels exceeding 300 meters, break into shorter segments and use map fusion in post-processing
• Never plan a pure out-and-back linear path. Drift accumulates with no opportunity for correction

Stopping the Scan

To stop: double-tap the power button, both presses within 500 milliseconds. Wait for the indicator to transition. Do not touch the device again until solid green confirms the save is complete.

Before Leaving Site

• Solid green confirmed. Scan saved successfully
• LixelGO preview reviewed for coverage gaps across all critical areas
• All control point names confirmed as entered correctly during scanning
• Coordinate CSV file accessible and ready for processing if using GCPs
• Raw data transfer plan confirmed. Know exactly how the data is getting to your processing machine

The complete post-scan procedure is in 1.5 Post-Scan Review and Data Transfer.

Common Early Mistakes

These are the errors that appear consistently with first-time users. Each one creates a downstream problem that is harder to fix than it was to prevent. The full troubleshooting reference is 10.3 Troubleshooting Index.

• Moving during initializationAny movement during the 15-second countdown corrupts the reference frame. The entire session carries that error. The only fix is to restart.
• Lifting the scanner immediately after the countdownThe countdown ending is not the signal to move. Hold for another 15 seconds. SLAM has not established motion-stable tracking yet. Moving too soon introduces early drift that compounds throughout the session.
• Walking too fastMotion blur degrades visual SLAM. Reduced dwell time means lower point density. In tight or feature-poor areas, excessive speed is the fastest way to produce unusable data.
• Linear routes without loopsOut-and-back paths accumulate drift with no correction mechanism. The further from the start, the worse the error. No post-processing step fixes a scan that never had loop opportunities built into the route.
• Trusting RTK without verifying Fixed statusRTK requires Fixed status, not Float, not Single Point. It also requires more than 10 valid satellites, more than 10 meters of movement while Fixed, and more than 100 valid points logged. Indoors, RTK rarely achieves Fixed. Verify in LixelGO before relying on it for absolute coordinates. See 4.2 RTK Positioning.
• Mismatched GCP namesThe point names in your coordinate CSV must match what you typed in LixelGO exactly, every character, every capital letter. A single mismatch causes that control point to fail at the processing stage. See 4.5 Control Point Marking.
• Processing with insufficient RAMLixelStudio holds the entire scan trajectory in memory. A 30-minute scan on a 32 GB machine will fail mid-processing. Plan hardware requirements before the project. See 2.5 Critical Basics.
• Powering off during saveFast-flashing green means data is still writing. Pressing the power button in this state corrupts the scan. Wait for solid green. Every time, without exception.
• No preview before leaving siteThe LixelGO preview shows coverage gaps. Use it before you leave. Discovering a gap at the office means a rescheduled project.
• Deciding on coordinate strategy after the scanRTK must be configured before scanning begins. GCPs must be placed and marked during the scan. There is no way to georeference data after the session without control that was captured during it. See 4.1 When Georeferencing Matters.

Understanding Accuracy

Accuracy in mobile SLAM is not a single number. The published specifications reflect results achieved with proper technique and, for absolute accuracy, proper georeferencing. Understanding what these numbers mean prevents misrepresentation to clients and unrealistic expectations in the field.

Relative accuracy is internal consistency: whether a wall is planar, dimensions are repeatable, geometry holds across the scan. Achieved through good technique and loop closure. No georeferencing required.

Absolute accuracy is how closely the point cloud registers to a real-world coordinate system. Requires RTK, PPK, or surveyed GCPs. Technique alone cannot produce it.

SLAM drift is the primary threat to both. It compounds over distance and is corrected only through loop closure and control points. The specifications above are achievable results when proper technique and control were applied. They are not guaranteed outcomes from any scan regardless of which device was used.