Part 1: Understanding Your XGRIDS System

---

Quick Field Guide

Pre-Flight Checklist

• Lixel charged and powered on
• iPad fully charged (iOS 17.5+)
• XGRIDS app installed and updated
• WiFi/Bluetooth enabled on iPad
• Location Services enabled for XGRIDS app
• Camera permissions granted to XGRIDS app

Connection & Setup

1. Power on Lixel (press power button 3 seconds)
2. Wait for WiFi icon on Lixel screen
3. iPad Settings → WiFi → Connect to "Lixel-XXXX"
4. Open XGRIDS app → Allow permissions
5. App auto-detects and connects

Starting a New Scan

1. Tap "+" in bottom center of app
2. Name your project (include location/date)
3. Select scan mode (Default: Standard)
4. Tap "Start Scanning" (green button)

Scanning Technique

• Speed: Slow walking pace (~1-2 ft/sec)
• Motion: Smooth, no jerky movements or rapid turns
• Camera: Hold steady, avoid shaking
• Coverage: 30% overlap between passes
• Distance: 3-10 feet from surfaces

Real-Time Monitoring

• Green dots = good point capture
• Check point density for adequate coverage
• Watch for low battery/connection warnings
• Review preview mesh for completeness

Device Initialization

1. Place on flat, stable surface (use steel control point base)
2. Ensure 360° clear view
3. Aim toward feature-rich areas (not blank walls)
4. Keep completely stable during 15-second countdown
5. Wait additional 15 seconds before moving

Learn more about initialization →

Active Scanning Best Practices

• Hold device upright (max 20° tilt)
• Maintain smooth, consistent walking speed
• Keep device >1 meter from objects
• Create loops returning to previous areas
• Slow down at doorways and stairs

Learn more about scanning workflow →

Control Point Marking (GCP Method)

1. Approach point slowly
2. Align control point base corner with physical marker
3. Lower device smoothly to ground
4. Trigger marking function in app
5. Wait 5 seconds (keep stable)
6. Slowly pick up and circle point 1-2 times
7. Continue scanning

Note: Control point names in app must exactly match coordinate file names.

Learn more about georeferencing →

RTK Monitoring (RTK Method)

• Green light = fixed solution (good)
• Blue light = float solution (pause if possible)
• Red light = no RTK (pause or note section)
• Keep sections without RTK <100m data-preserve-html-node="true" (L2 Pro) or <50m data-preserve-html-node="true" (K1)
• Hold antenna vertical

Learn more about RTK →

Session Completion

1. Wait for indicator light sequence:
   • Slow-flashing green = scanning active
   • Fast-flashing green = saving data (DO NOT power off)
   • Solid green = save complete
2. Power off safely

Processing Quick Steps

LixelStudio:

1. Import project folder from scanner
2. Configure coordinate transformation (RTK/PPK/GCP if needed)
3. Select camera coloring settings
4. Enable loop closure if scan returned to start
5. Process (2-24 hours depending on settings)

LCC CyberColor Studio (for photorealistic models):

1. Import same project folder
2. Select quality level (fast/standard/slow)
3. Configure portability settings
4. Process

Learn more about processing →

Common Output Formats

• LAS/LAZ - General point cloud use
• E57 - Structured archival
• RCP - Autodesk Recap/AutoCAD
• LCC - XGRIDS ecosystem (smallest file)
• PLY - Open-source viewers, game engines
• DXF - 2D drawings for CAD

Learn more about deliverables →

---

Full Explanatory Guide

The Lixel Ecosystem

The XGRIDS Lixel system is a comprehensive mobile SLAM (Simultaneous Localization and Mapping) platform for high-precision 3D data capture and photorealistic reconstruction. The ecosystem consists of four integrated components:

Hardware Foundation
Handheld LiDAR scanners with integrated IMU (Inertial Measurement Unit), high-resolution cameras, and optional RTK/PPK modules for georeferencing. Unlike traditional survey methods requiring stationary setups and line-of-sight measurements, the Lixel system enables continuous mobile scanning where the operator walks through spaces while the device automatically captures geometry and visual information.

Mobile Application Layer
LixelGO provides real-time control and monitoring on Android and iOS devices, displaying live point cloud previews, managing scan sessions, configuring RTK connections, and marking control points. Real-time feedback allows operators to verify coverage and quality during collection rather than discovering issues later.

Processing Software
Two specialized applications handle different outputs:

• LixelStudio - Point cloud generation from raw scanner data, performing SLAM optimization, coordinate transformations, and traditional point cloud workflows for engineering and design applications
• LCC CyberColor Studio - Uses 3D Gaussian Splatting technology to create photorealistic models from scan data, generating immersive visualizations for client presentations and web sharing

Visualization and Sharing
Processed data delivers as conventional point clouds for CAD/BIM integration, photorealistic LCC models for web browsers, or hybrid deliverables combining both formats. This flexibility allows the same field data to serve multiple project needs without requiring separate collection passes.

Key Differentiators from Competing Systems

Real-time RTK Integration
Provides direct absolute coordinate output during scanning rather than requiring post-processing transformation. This eliminates point cloud layering issues that occur when RTK data is applied after SLAM processing. The L2 Pro can output point clouds directly in projected coordinate systems, ready for immediate GIS use.

SLAM-Based Efficiency
Enables large area coverage through continuous mobile scanning. A single operator can capture 10,000+ square meters per day, compared to traditional static laser scanning which typically covers 1,000-2,000 square meters daily due to setup time at each station. The mobile workflow naturally captures transitions between spaces that static scanning often misses.

High-Density Point Cloud Generation
Achieves up to 1,000,000 points per square meter when using point cloud enhancement. This density exceeds most mobile mapping systems and approaches static terrestrial scanner quality. The combination of LiDAR measurement with visual SLAM produces consistent density across varying distances, unlike photogrammetry which degrades with distance.

Integrated Coloring
Merges geometric accuracy with photorealistic appearance. The dual 48-megapixel panoramic cameras on the L2 Pro capture texture information simultaneously with LiDAR geometry, ensuring perfect alignment between color and structure. This eliminates registration challenges inherent in systems that capture geometry and photos separately.

Survey-Grade Accuracy
3cm RMSE (Root Mean Square Error) for both horizontal and vertical dimensions matches survey-grade requirements when proper workflow is followed. This accuracy applies with control point spacing up to 100 meters or RTK disconnection up to 100 meters, making the system viable for engineering documentation and as-built verification.

Hardware Lineup

Lixel L2 Pro Specifications

The L2 Pro is available in three configurations based on LiDAR capabilities:

16-Channel 120-Meter Configuration
Uses a Velodyne-style rotating LiDAR with 16 laser channels and 120-meter maximum range. Scans at 320,000 points per second. Suits general documentation projects where maximum range is not critical. The lower channel count produces adequate density for most architectural and construction applications at walking speed.

32-Channel 120-Meter Configuration
Doubles the laser channels to 32 while maintaining 120-meter range. Scans at 640,000 points per second, producing denser point clouds particularly beneficial for outdoor environments with vegetation or distant objects. The increased channel count improves feature detection for SLAM algorithms, resulting in more stable tracking.

32-Channel 250-Meter Configuration
Premium option with 32 laser channels and extended 250-meter range. Scans at 640,000 points per second with enhanced distance capability. Required for corridor mapping, long-distance facades, or large infrastructure projects where targets exceed 120 meters. The extended range maintains point density at distance better than the 120-meter variants.

Shared Specifications Across All Configurations:

• Dual 48-megapixel panoramic cameras (360° horizontal, 300° vertical coverage)
• 6-axis IMU for motion tracking
• Optional RTK module (multi-constellation GNSS: GPS, GLONASS, Galileo, BeiDou)
• WiFi 6 and Bluetooth 5.0 connectivity
• 256GB internal storage
• 4-6 hour battery life (varies with configuration and RTK usage)
• IP54 dust and splash resistance
• 1.8kg weight (without RTK module)

Lixel K1 Specifications

The K1 serves as the entry-level device with simplified hardware at reduced cost:

LiDAR System:
Single solid-state LiDAR module with 120-meter range, scanning at 320,000 points per second. The solid-state design (no rotating parts) improves reliability but produces lower point density than the L2 Pro's rotating systems.

Camera System:
Single 12-megapixel camera with 180° field of view. Captures adequate color information for visualization but with lower resolution and coverage than the L2 Pro's dual panoramic setup.

Positioning:
6-axis IMU for motion tracking. No RTK capability—georeferencing requires control point workflow exclusively.

Physical Specifications:

• 128GB internal storage
• 3-4 hour battery life
• 1.2kg weight
• IP54 dust and splash resistance

Use Case Fit:
The K1 suits interior documentation, small construction sites, and applications where RTK georeferencing is not required. The lower point density and single camera limit its effectiveness for large outdoor areas or projects requiring survey-grade deliverables.

Georeferencing Methods

Three methods provide absolute coordinate output, each with distinct accuracy, workflow, and cost implications:

RTK (Real-Time Kinematic)
RTK provides real-time absolute positioning during scanning by receiving corrections from a base station or NTRIP service. The RTK module communicates with GNSS satellites and correction sources simultaneously, computing centimeter-level position throughout the scan session.

Advantages: No control points required, immediate absolute coordinates, fastest workflow, ideal for large sites or corridors where control point placement is difficult.

Limitations: Requires continuous satellite visibility (minimum 5 satellites with good geometry), correction signal must remain available (cellular connection for NTRIP or radio link to base station), accuracy degrades during satellite blockage, requires RTK-capable hardware (L2 Pro only).

Accuracy: 3cm horizontal/vertical RMSE with continuous RTK fix. Accuracy degrades to 10-15cm in sections with RTK float or loss, though SLAM bridging maintains relative accuracy. Maximum recommended gap without RTK fix is 100 meters.

PPK (Post-Processed Kinematic)
PPK records GNSS observations during scanning and processes them afterward with base station data. This method tolerates temporary satellite blockage better than RTK because post-processing can resolve ambiguities that real-time processing cannot.

Advantages: More reliable than RTK in partially obstructed environments, no real-time correction signal needed, can achieve similar accuracy to RTK with proper processing, works with recorded base station data from permanent networks.

Limitations: Requires post-processing step before final coordinates are known, needs base station observation files (RINEX format), requires RTK-capable hardware even though processing is post, operator cannot verify absolute positioning during collection.

Accuracy: 3cm horizontal/vertical RMSE with successful PPK solution. Processing may fail in heavily obstructed areas where RTK would also fail.

GCP (Ground Control Points)
GCP workflow uses surveyed markers placed throughout the scan area. During scanning, the operator marks these points using the control point base and app function. Processing software aligns the scan to control point coordinates through least-squares optimization.

Advantages: Works without GNSS hardware, effective in GNSS-denied environments (urban canyons, underground, indoor), works with both L2 Pro and K1 hardware, control point accuracy is independent of scan duration.

Limitations: Requires pre-placed surveyed markers, control point marking during scanning adds time, requires control point visibility from scan path, coordinate file preparation and name matching required, minimum 3-4 points needed (more recommended for large areas).

Accuracy: 3cm horizontal/vertical RMSE achievable with proper control point network. Accuracy depends on control point spacing (recommended maximum 100 meters between points), marking precision (careful placement on control point base), and network geometry (points should surround scan area, not linear arrangement).

Method Selection Guidelines:

• Large outdoor sites with good sky visibility → RTK
• Partially obstructed outdoor sites with base station access → PPK
• Indoor environments or GNSS-denied areas → GCP
• Projects requiring maximum flexibility and reliability → GCP with RTK/PPK backup
• Budget constraints or K1 hardware → GCP only option

Complete Workflow Overview

The Lixel workflow consists of four phases: Planning, Collection, Processing, and Deliverable Preparation. Understanding the complete sequence prevents common errors that require recollection or reduce deliverable quality.

Planning Phase

Effective planning prevents collection failures and ensures deliverables meet specifications. This phase typically requires 30 minutes to 2 hours depending on project complexity.

Site reconnaissance identifies critical factors affecting data collection. Walking the site reveals obstacles requiring workarounds, areas needing special attention, GNSS obstruction zones affecting RTK reliability, and hazards requiring safety procedures. This reconnaissance should occur before mobilizing equipment when possible, though on-demand projects may combine reconnaissance with collection.

Route planning determines the scanning path through the site. Effective routes create loops that return to starting areas rather than linear paths, prioritize coverage of critical features while minimizing redundant passes, identify initialization locations with good feature visibility and stability, plan control point or RTK check point locations (if using absolute positioning), and estimate scan duration based on area and detail requirements.

Control point establishment (if using GCP method) requires surveying marker locations before scanning. The control point network should surround the scan area rather than forming linear arrangements, maintain spacing appropriate to site size (typically 30-100 meters between points), use markers that won't move during scanning (permanent bolts, painted corners, or steel bases), and be visible from the planned scan path without obstructions. The coordinate file must use a consistent format with point names exactly matching the names that will be entered in LixelGO during scanning.

RTK configuration (if using RTK/PPK method) establishes connection parameters before arriving on site. This includes identifying the NTRIP caster URL and credentials or base station IP and port, confirming the coordinate system and datum for output (usually project-specific), testing the connection from the planned work area to verify signal strength, and ensuring the iPad has adequate cellular data or hotspot connection.

Equipment verification confirms all hardware and software are ready for collection. This includes charging all batteries (scanner, iPad, RTK module if separate, backup batteries), updating firmware and apps to latest versions, clearing storage on scanner and iPad for new data, loading any required coordinate files or configuration profiles, and testing the complete system end-to-end with a brief practice scan.

Collection Phase

The collection phase captures raw sensor data through mobile scanning. This phase typically requires 30 minutes to several hours depending on site size and complexity, though the entire collection day including setup and teardown often spans 4-8 hours for large sites.

Device initialization establishes the scanner's starting position and orientation, providing the reference for all subsequent measurements. Proper initialization requires:

• Placing the device on a flat, stable surface (using the steel control point base)
• Ensuring 360-degree clear view of surroundings
• Aiming toward feature-rich areas rather than blank walls
• Maintaining complete device stability for the full 15-second countdown
• Waiting an additional 15 seconds after countdown before moving

Poor initialization causes issues throughout the entire scan that cannot be fully corrected in processing. If the device moves during initialization or the initialization location lacks sufficient features, restart the session with better positioning.

Active scanning involves walking the planned route while the device captures data continuously. Best practices include:

• Holding the device upright (vertical) with tilt not exceeding 20 degrees during normal walking
• Maintaining smooth consistent walking speed without rapid rotations or sudden stops
• Keeping the device more than 1 meter from surrounding objects to prevent LiDAR occlusion
• Creating loops that return to previously scanned areas rather than following linear paths
• Slowing significantly at transitions like doorways or stairs to ensure adequate feature overlap

The real-time preview in LixelGO provides feedback on coverage quality. Green points indicate successful measurements, while sparse areas suggest moving slower or making additional passes. The preview mesh helps identify gaps requiring additional coverage.

Control point marking (if using GCP georeferencing) occurs during active scanning when the route passes control point locations:

1. Approach the point slowly
2. Align the sharp corner of the control point base with the physical marker
3. Lower the device smoothly to the ground ensuring stability
4. Trigger the marking function in LixelGO (records timestamp and scanner position)
5. Wait 5 seconds keeping the device completely stable
6. Slowly pick up the device and circle the control point 1-2 times for adequate coverage
7. Continue scanning

The control point names entered in the app must exactly match the names in the coordinate file prepared during planning. Case-sensitive matching is required—"CP1" does not match "cp1".

RTK monitoring (if using RTK georeferencing) requires continuous attention to the RTK module status light and app display:

• Indicator light should remain green (fixed solution) throughout scanning
• Blue light (float solution) or red light (no RTK) requires pausing until fix reacquires or noting the affected section length
• Each section without RTK fix should remain under 100 meters (L2 Pro) or 50 meters (K1) to maintain accuracy specifications
• Ensure the antenna remains vertical by holding the device upright
• Avoid areas where satellite visibility is completely blocked

If RTK fix is lost in a critical area, consider rescanning that section when satellite visibility improves or supplementing with control points in the affected zone.

Session completion requires proper shutdown to prevent data corruption:

1. Wait for the device indicator light to change from slow-flashing green (scanning active) through fast-flashing green (saving data) to solid green (save complete)
2. Never power off during the fast-flashing phase—this corrupts data and may render the entire session unrecoverable
3. After reaching solid green, the session data resides safely on internal storage awaiting download

Processing Phase

The processing phase transforms raw sensor streams into usable point clouds and photorealistic models through computational optimization and reconstruction. This phase typically requires 2-24 hours of computer time depending on scan duration and selected quality settings.

LixelStudio project processing begins by importing the project folder copied from the scanner. Configuration includes:

Coordinate transformation (if absolute coordinates are required):

• RTK method - Uses data embedded during scanning, no additional files needed
• PPK method - Imports base station observation files in RINEX format
• GCP method - Imports control point coordinate file and matches scanned points to surveyed locations

Camera coloring configuration:

• Built-in dual camera coloring - Uses the scanner's integrated panoramic cameras (if device has dual cameras)
• External panoramic camera coloring - Uses Insta360 or similar video captured during scanning (requires synchronization)

Advanced settings:

• Filter strength - Adjusts noise removal aggressiveness (higher removes more outliers but may remove valid detail)
• Loop closure - Enable if the scan returned to its starting point (improves global accuracy)
• SLAM modes - Special algorithms for challenging environments (underground, corridor, outdoor open area)
• Point cloud enhancement - Generates maximum density (up to 1,000,000 points/m²) by incorporating visual features

The software processes data through several stages:

1. Load sensor data and perform preliminary alignment of LiDAR and camera data
2. SLAM optimization solves for scanner position and orientation across the entire trajectory
3. Map construction builds the point cloud from optimized poses and LiDAR measurements
4. Coloring projects panoramic imagery onto geometry when configured
5. Coordinate transformation applies RTK/PPK/GCP data if configured
6. Export results in selected formats

Processing outputs:

• Optimized point cloud in LAS format with attributes (coordinates, intensity, RGB color if enabled)
• Trajectory file containing scanner position and orientation at each timestamp
• Panoramic images if panoramic output was enabled
• Accuracy report if coordinate transformation was configured with check points
• Processing logs for troubleshooting if issues occurred

LCC CyberColor Studio reconstruction (if photorealistic visualization is required) imports the same scanner project folder used in LixelStudio. Configuration includes:

Quality level:

• Fast - Lower splat count, faster processing, suitable for preliminary review
• Standard - Balanced quality and processing time for most deliverables
• Slow - Maximum splat count and quality for critical presentations

Advanced settings:

• Maximum Gaussian splat count - Constrains VRAM usage during viewing
• Portability settings - Optimizes for web/mobile versus desktop viewing
• HD enhancement - Incorporates additional DSLR photos for critical detail areas

The reconstruction process differs fundamentally from traditional photogrammetry. Rather than creating explicit mesh geometry with texture maps, the optimization fits millions of Gaussian primitives to reproduce the appearance of the scene from any viewpoint. This approach handles complex geometry, partial transparency, and view-dependent effects more effectively than traditional mesh reconstruction while requiring less processing time.

Reconstruction outputs:

• LCC model file in proprietary compressed format
• Optional PLY export in standard 3DGS format
• Trajectory and panoramic image files if configured
• Reconstruction report documenting settings and statistics

Deliverable Preparation Phase

The deliverable preparation phase refines processed data and generates client-ready outputs in appropriate formats. This phase typically requires 1-4 hours of user interaction depending on deliverable complexity.

Point cloud deliverable preparation uses LixelStudio tools to perform:

Quality checks:

• Review coverage to identify gaps
• Identify noise or outliers requiring removal
• Verify accuracy at check points if available

Cleanup operations:

• Denoising to remove outliers
• Resampling to reduce file size if needed (while maintaining required density)
• Clipping to remove unwanted areas

Coordinate verification:

• Confirm output coordinates match specifications
• Test integration with client software (AutoCAD, Revit, GIS platforms)

Format export:

• LAS/LAZ - General point cloud use, widely supported
• E57 - Structured archival format with metadata
• RCP - Autodesk Recap/AutoCAD native format

LCC model deliverable preparation uses LCC Studio editor to perform:

Quality review:

• Navigate scene to verify completeness
• Check color quality and lighting consistency

Model refinement:

• Crop unwanted areas using selection tools
• Color grading to adjust brightness/contrast/saturation
• Add annotations for critical features

Measurement extraction:

• Document dimensions for verification
• Create measurement reports if required

Export:

• LCC - XGRIDS ecosystem compatibility, smallest file size
• PLY - Open-source viewer compatibility, game engine integration
• USDZ - NVIDIA Omniverse workflows
• 3D Tiles - Web GIS platforms (Cesium, ArcGIS)

2D drawing deliverable creation (if required) uses LixelStudio drawing tools to extract vectors from point clouds:

AI automatic extraction:

• Configure parameters (ground sampling distance, minimum segment length)
• Review automatic results
• Manual refinement of vectors

Manual digitizing:

• Trace features directly from point cloud
• Use snap modes for accuracy
• Create floor plans or elevations
• Add dimensions and annotations
• Organize into layers

Export:

• DXF format for AutoCAD and other CAD platforms

Volume calculation deliverables (if required) use LixelStudio volume tools:

1. Define stockpile boundaries
2. Select base plane (horizontal reference or custom three-point plane)
3. Configure meshing parameters
4. Calculate volumes with automatic report generation
5. Export reports with volume values, visualizations, and methodology documentation

Web publishing (for LCC models) uses LCC Studio publishing function:

1. Upload models to XGRIDS cloud
2. Configure access controls (public link, password protection, or third-party sharing)
3. Add model descriptions and metadata
4. Generate shareable URLs

Published models load in standard web browsers without requiring software installation or file downloads, making them ideal for client presentations and stakeholder reviews.

The workflow completes when all deliverables have been verified against project requirements, organized per client specifications, and transferred through agreed methods (portable drive, cloud upload, or direct access to published models).