XGRIDS Pro Guide™ / Module 6: Advanced L3, L2 Pro, and K2

Module 6 Quick Field Guide

The numbers, limits and sequences for advanced L2 Pro and K2 work, on one page. Segment planning, camera coloring, drone mode, aerial-ground fusion, and the mistakes that cost a job.

Numbers to Remember

Plan segments at 12 to 15 minutes and keep every capture under 20 minutes. That satisfies every limit in the table on any supported workstation and in either pipeline.

Figure
Value
What it governs
Device capture ceiling
90 minutes, L2 Pro and K2
Firmware hard stop on one recording session. Not a scan target.
Battery runtime
About 90 minutes; about 72 below 50°F (10°C)
The L2 Pro uses a 14.4 V, 46.8 Wh battery and the K2 a 1,900 mAh battery. A separate fact that shares the number 90.
Alpine session ceiling
About 70 minutes per charge
Alpine practice: end the scan at 20 percent remaining, leaving reserve for the overlap walk and the write process.
LixelStudio Map Fusion
10 maps; 20 minutes per segment; 200 minutes combined
Point cloud pipeline. The per-segment limit is LixelStudio's own; the combined total is the same as LCC Studio's.
LCC Studio Map Fusion
10 segments; 200 minutes combined
3DGS pipeline. No per-segment time limit applies.
LCC Studio reconstruction memory
64 GB to 30 min, risk past 45. 128 GB to 60 min, risk past 90.
The limit that actually fails jobs. LixelStudio publishes no memory-to-duration figure.
Reconstruction time
About 20 minutes per minute of capture
A 45-minute scene is most of a day of compute before you know whether it worked.
Segment overlap
50 ft (15 m) floor; 65 to 100 ft (20 to 30 m) target
Shared scanned ground between neighbouring segments. Cannot be added later.
Control point interval
L2 Pro 328 ft (100 m); K2 164 ft (50 m)
Maximum spacing. Distribute evenly, never in a straight line.
Post-stop write
30 to 60 seconds, longer after a long scan
Fast-flashing green means writing. Solid green means safe.
Colorization minimum
L2 Pro 2 minutes; K2 3 minutes
Continuous movement. A shorter scan cannot complete the coloring step.
Initialization
About 30 to 45 seconds per segment
Hold still throughout. Movement corrupts the reference frame for the whole segment.
LAS filename for conversion
20 characters or fewer
LAS to E57 and LAS to RCP conversion. A longer name makes the conversion fail, hang or crash.

The two 90-minute figures are unrelated, and neither is a scan target. One is how long the firmware keeps recording; the other is how long the battery lasts. XGRIDS recommends against capturing a full 90-minute dataset in one session on any hardware, because a single fault inside a long capture, whether drift, interference or a battery problem, can cost the entire dataset. Split large scenes and combine them with Map Fusion. Full reasoning on 6.2 Scan Splitting.

Battery and Segment Planning

Battery Discipline
  • End the scan at 20 percent remaining. That is roughly 70 minutes of scanning on a full charge, not 90. The reserve covers walking the overlap zone, reaching a clean stop position, and the write process.
  • Plan cold-weather battery counts on 72 minutes, not 90. Below 50°F (10°C) runtime drops roughly 20 percent.
  • Add initialization time at each segment start, plus travel between areas and setup.
  • End the segment as soon as the app warns that the battery is low. End immediately rather than trying to finish the room.
  • Change batteries between segments, never during one. Removing the battery cuts power and ends the recording, and neither device resumes a scan from a breakpoint.
Splitting a Scan
  • Split at feature-rich geometry: a corridor intersection, equipment cluster, or lobby with columns. Never mid-corridor or against a blank wall.
  • Scan past the split point by at least 50 ft (15 m) to build the overlap zone. Target 65 to 100 ft (20 to 30 m).
  • Mark the connection before stopping. Valid RTK in every segment, or a shared control point with an identical, case-matched name in both segments.
  • Start the next segment where the last one ended, without moving the device, then walk back through at least 50 ft (15 m) of the previous coverage before new ground. On a control point seam, mark the same name after initialization and wait 10 seconds before lifting.
  • Vary your line through the overlap zone. Retracing the exact path gives the aligner only one viewpoint.
  • Never let one segment sit entirely inside another. With no unique coverage of its own, the aligner has nothing to solve against.

An L2 Pro or K2 segment with neither RTK nor a shared control point cannot be fused, and the link cannot be added after capture. Overlap and visual features alone do not qualify a segment. An uncontrolled segment has to be recaptured, so place fallback control points at every seam even on jobs running RTK. The L3 does not require control points for Map Fusion: its sessions are matched by region selection.

Camera Coloring

The L3, L2 Pro, and K2 all produce color from onboard cameras during scanning. No external camera and no pre-scan camera configuration is required for any of them. The same raw camera data serves both pipelines; the pipeline decides only how it is applied.

Device
Onboard cameras
Coloring timing
L3
1 forward (100° × 75°) plus 2 panoramic (200° × 200° each), 4096 × 3072 each
Real-time, on device
L2 Pro
2 × 48 MP panoramic plus 1 MP positioning camera
Applied during processing
K2
3 cameras: 2 fisheye plus 1 forward, 200° × 200° coverage
Real-time true color in RGB Mode; applied during processing in Guided Mode

K2 Guided Mode and RGB Mode are a pre-scan choice you cannot reverse. RGB Mode gives real-time true color but removes the Scan Assist System and the Scan Quality view, which are the only in-field checks on coverage density. Guided Mode keeps those checks but shows elevation or quality coloring instead of true color. If nobody is reviewing coverage on site, choose Guided Mode and check Scan Quality before you leave. Choose RGB Mode when a client is watching the screen or true-color confirmation matters more than a coverage check.

Field Practices for Reliable Coloring
  • Keep every lens clear. A sleeve, strap, or hand across a lens for any sustained period leaves a coloring gap in that part of the scan, and it is not fixable afterward.
  • Do not point cameras into the sun or any bright source. Direct overexposure corrupts color for the affected area, not just the bright frames.
  • Meet the minimum duration: 2 minutes on the L2 Pro, 3 minutes on the K2, with continuous movement. A shorter scan does not carry enough camera data for the coloring step to complete. Pad a short scan with extra walk time before stopping.
  • Hold a steady pace and avoid sharp rotations. Rapid angular movement degrades frame quality and produces blurred coloring.
  • Light the space evenly. Bring supplemental lighting for utility tunnels, deep basements, and shadowed mechanical rooms.

Drone Mode

Drone mode is an L2 Pro capability only. A drone mount for the L3 is in development for the DJI Matrice 400. The K2 is handheld and is not drone-mountable. Compatible aircraft are the DJI M300 RTK and M350 RTK, which supply RTK positioning from their own GNSS.

L2 Pro Setup Sequence
  • Remove the RTK module from the L2 Pro before mounting. The aircraft provides RTK in drone mode, and leaving the module fitted causes a positioning conflict.
  • Mount the L2 Pro in the bracket and engage the green safety lock.
  • Connect the L2 Pro to the aircraft with the Type-C power cable.
  • Power the drone first, then the L2 Pro. Powered through the Type-C cable, the L2 Pro initializes in drone mode. Powered from its own battery first, it initializes handheld and drone positioning will not work correctly.
  • Configure the scan in the field app before takeoff and confirm drone mode is active.
  • The aircraft will show a camera obstruction warning in flight. That is expected. The L2 Pro body partially covers a downward obstacle-avoidance camera. Do not abort for it.
L2 Pro parameter
Recommended
Maximum
Flight speed
3 m/s
5 m/s. Data quality degrades noticeably above this.
Flight altitude
98 ft (30 m)
164 ft (50 m). Higher altitude reduces point density.
Flight distance
No restriction
No restriction

Aerial-Ground Fusion

Aerial-ground fusion combines drone imagery with ground scan data into one model covering roof-level and interior geometry. LCC Studio supports the L3, the L2 Pro, and the K2 as the ground scanner, using RTK to tie the ground scan to the drone data.

Hard Requirements Before Starting
  • RTK enabled on the scanner and on the drone, in the same coordinate system (WGS84). The L3, L2 Pro and K2 mark no aerial-ground fusion point: that button belongs to the PortalCam in LCC Scan. Update the field app and the device firmware before the site visit, not during it. Exact versions are on 6.4 Aerial-Ground Fusion.
  • Processing hardware: 16 or more cores, such as AMD Ryzen 9 9950X, 64 GB of DDR5 memory (128 GB for larger-scale data), NVIDIA RTX 3090 or better. This is a higher bar than single-scene reconstruction.
  • Aerial input: more than 100 and fewer than 10,000 images. Outside that range the job will not run.
  • Budget 24 to 48 hours or more of processing for a large site.
  • Ground segments: 10 segments, 200 minutes combined. LCC Studio sets no aerial-ground per-segment cap. Plan 20 to 30 minutes per segment, sized to the workstation that will process the job.
  • All three inputs are mandatory: ground scan data, drone imagery, and photos at each takeoff and landing point. A missing input may reduce aerial-ground fusion quality or cause the reconstruction to fail.
Fusion Point Planning
  • Plan at least 3 to 4 takeoff and landing points per scene, spread across it. The layout XGRIDS gives for a small site: on a site of about 100 by 100 m or less, one point at each corner and one at the center. Points clustered at one end do not constrain the far end of the model.
  • On a larger site, plan at least one takeoff and landing point every 164 to 328 ft.
  • Each point must be visible both from overhead and from the ground scan path. Put them in open areas you can reach during the walk.
  • On the L3, L2 Pro and K2, route the ground scan past each takeoff and landing point with RTK Fixed. There is no point to mark. The PortalCam method, a fusion point at the takeoff spot facing the drone camera direction with a full loop around it, is on 7.2 PortalCam Advanced Scanning.

Before You Open the Processing Software

Data Preparation Checklist
  • Keep the original data on the device until processing is complete and the output is verified. Do not delete device files after copying. The K2 internal storage is not field-swappable, so plan around end-of-day transfers rather than mid-session swaps.
  • Check the workstation against the segment length before you start. For LCC Studio reconstruction, 64 GB handles 30 minutes of capture data reliably and 128 GB handles 60 minutes of capture data. Full requirements on 10.7 Device Specifications.
  • Give every segment its own named folder before importing. Inconsistent folder structure on a large project causes segments to reference the wrong files and adds hours to deliverable prep.
  • Keep paths free of spaces, using letters, digits, hyphens, underscores and periods only. Import the top-level project folder, not a sub-folder.
  • Keep drone imagery in a clearly labelled separate folder at the same level as the ground segment folders, with a _Drone suffix.
  • Process from a local internal disk. External drives, USB flash drives and network paths are a documented cause of post-processing crashes.

Critical Warnings

  • Data loss Stopping the scan does not end the job. Wait out the write process. After you tap Stop, the scanner runs a sequence of processes and writes the project file to internal storage. That takes 30 to 60 seconds, and longer after a long capture. The green indicator flashes rapidly the whole time and goes solid green when it is finished. Do not power off, remove the battery, or disconnect until it is solid green. An interrupted write corrupts the project file and the capture is not recoverable.
  • No recovery On the L2 Pro and K2, fusion connections are made in the field only. A segment captured without RTK and without a shared control point cannot be admitted to a fusion, and the link cannot be added afterward. That segment must be recaptured.
  • Warning Control point names are case-sensitive across segments. A point named fp01 in one segment and FP01 in another reads as two different points. The connection fails silently, with no error, and the segment drops out of the result.
  • Warning Do not move during initialization. Movement corrupts the reference frame and the error is carried by the entire segment. The only fix is to restart the segment.
  • Warning Onboard cameras must stay unobstructed. A blocked lens leaves a coloring gap for that portion of the scan that cannot be repaired in processing.
  • Warning Meet the colorization minimum: 2 minutes on the L2 Pro, 3 on the K2. A shorter scan cannot complete the coloring step.
  • Warning Remove the RTK module before mounting the L2 Pro on a drone, and power the aircraft before the scanner. Either mistake produces a positioning conflict or a handheld-mode initialization.
  • Warning Flying the L2 Pro above 5 m/s or above 164 ft (50 m) degrades point density. Both are soft limits. The scanner keeps working, but the gaps and reduced density cannot be corrected after collection.
  • Warning The per-segment rule differs between the pipelines. The combined total does not. The per-session limit of 20 minutes per segment belongs to LixelStudio Map Fusion, and LCC Studio Map Fusion sets no per-segment cap. Both stop at 200 minutes combined.

©2026 Alpine Reality Capture LLC  •  XGRIDS Pro Guide™  •  Site Disclaimer