XGRIDS Pro Guide™ / Module 3: Field Technique

Route Planning Visual

Interactive diagram showing backbone-first scan routing, target placement, and RTK strategy applied to a large data center.

How to Read This Plan

This diagram uses 2 pods of a data center as the example. Any multi-zone facility such as warehouses, hospitals, campuses, and commercial buildings all follow similar route planning principals. Scan the backbone first as a loop covering the main corridor and center cross-aisle. Then capture each zone as its own session with overlap. Where possible, start and end every session outdoors where RTK has satellite coverage.

This page is a visual companion to 3.2 Route Planning and 3.5 Data Centers, which cover the full rationale and technique details. Read those sections first. Use this plan as a field reference.

Data Center (2 Pod Example)

OUTSIDE OUTSIDE POD A POD B C1 H1 C2 H2 C3 H3 C4 H4 SERVER RACKS C5 H5 C6 H6 C7 H7 C8 H8 SERVER RACKS ADDITIONAL ZONES: SAME METHOD 2 START + END 1 START + END 3 START + END RTK RE-FIX MAIN ACCESS CORRIDOR 1 2 3 MAX SPEED: 0.5 m/s (1.5 ft/s) ~100 ft

Target Types

OUTDOORINDOORGCP 1GCP 2WALK THROUGHWALL

Threshold GCP Pair

1 sticker outdoors and 1 indoors at each building entry. Mark both in LixelGO while RTK carry-in is active. These anchor the coordinate frame at the transition.

REFLECTIVE STICKERraised floor tile

Floor GCP Sticker

Circle once, set scanner on target, hold stationary for a few seconds while pointing toward the area with the most features, then mark in LixelGO. Maximum 100 ft apart for L2 Pro in data center aisles. 50 ft for K1.

rack endMAGNETIC BASE

Magnetic Checkerboard Plate

Attaches to rack uprights, steel columns, containment framing. Face toward aisle centerline. Keep away from hot-aisle vents.

100 MM SPHEREpermanent tribrach mount

Reference Sphere (Permanent Option)

Mount can stay installed with the spheres kept with the scanner kit.

GCPs can be made permanent. Both indoor and outdoor control points can use permanently installed markers. For facilities that will be rescanned on a recurring schedule, leaving GCP stickers, magnetic plates, and sphere mounts in place between campaigns eliminates repositioning error and reduces setup time on every return visit.

Pre-Scan Assessment and Preparation

Preparation determines whether the scan succeeds before the scanner powers on. Complete these items before entering the data hall.

Standardize Lighting

Confirm all aisles are fully lit. The SLAM system's visual tracking component relies on consistent lighting to detect and match features between frames. Partially lit aisles, especially cold aisles with overhead containment that blocks ceiling fixtures, degrade visual feature quality and increase drift risk. If facility lighting is uneven, coordinate with operations to bring all zones to full illumination before scanning begins.

Coordinate Containment and Door State

Open all doors that will be traversed during the scan and decide on a fixed state for rack enclosures before recording starts. Any element that moves between the time you scan it and the time you return past it creates "ghosting," duplicate geometry at different positions in the point cloud. This includes containment doors, rack enclosure panels, fire doors on hold-open magnets, and rollup barriers between zones. Set the state once and leave it for the full session.

Mark Vertical Landmark Locations

Walk the site before scanning and identify areas with high-density overhead infrastructure: main busway runs, primary cable tray intersections, and overhead PDU distribution. Mark these on your scan plan as locations where the scanner needs to be lifted above rack height using the XGRIDS extension pole accessory (available in 2 m and 3 m lengths). Planning these lifts in advance prevents backtracking mid-session to capture overhead geometry you missed on the first pass.

Place Visual Anchors in Repetitive Aisles

Data center aisles with identical rack faces on both sides provide minimal visual texture for the SLAM camera to track. Place temporary visual anchors at regular intervals along aisles that lack distinguishing features. High-contrast printed targets, QR codes, cardboard boxes, or even unique equipment labels placed on the floor or on rack doors give the visual SLAM component unambiguous features to lock onto. These also serve as secondary GCP candidates if positioned at known locations. Remove or leave them after scanning depending on facility policy.

Visual anchors are not optional in visually repetitive environments. Without them, the SLAM camera component has no way to distinguish one position from another in a row of identical black rack fronts. LiDAR geometry alone may not be sufficient to prevent drift in these conditions.

Technique Notes for Data Centers

These techniques address the specific challenges of scanning repetitive, feature-sparse environments. They supplement the general technique guidance in Module 3 and the detailed rationale in 3.5 Data Centers.

360° Junction Rotations

At every corridor intersection, aisle entrance, and pod boundary crossing, perform a slow 360° rotation with the scanner before continuing. This eliminates occlusions behind pillars, containment framing, and rack end caps that a straight walk-through would miss. The rotation captures geometry from all directions at that junction and gives SLAM a strong registration anchor at the transition point. Every intersection is a junction rotation opportunity. Skipping rotations at intersections is the most common source of registration gaps in multi-aisle scans.

Vertical Scanning for Overhead Infrastructure

Cable trays, busways, and overhead power distribution are frequently the primary deliverable for data center renovation projects. Standard hand-height scanning captures rack faces well but misses overhead geometry above the top of the racks. At planned intervals (identified during the pre-scan walkthrough), lift the scanner above rack height using the XGRIDS extension pole accessory. Hold the elevated position steady for 3 to 5 seconds, then lower and continue. This captures overhead cable infrastructure, tray routing, and clearances that are invisible from aisle level.

Extension pole accessories: XGRIDS offers 2 m and 3 m extension poles for both the L2 Pro and K1. The 2 m pole covers most standard data center overhead capture. The 3 m pole is appropriate for tall-ceiling industrial or mezzanine environments.

GCP Collection Technique

When marking a ground control point during the scan, approach the target, circle it once to capture surrounding geometry, then set the scanner down on or adjacent to the target. Hold stationary for a few seconds with the scanner pointed toward the direction with the most visible features (not toward a blank wall or rack face). Mark the point in LixelGO, then circle the point 1 to 2 more times before continuing. The stationary dwell and feature-facing orientation give LixelStudio stronger geometry to match the control point against during processing.

Execution of the Nested Loop

This workflow applies the backbone-first routing shown in the diagram above to a live scan. Each step builds on the previous session's coverage.

Primary Circuit (Backbone)

Walk the perimeter of the main corridor and cross-aisle, forming the backbone loop. Start and end outdoors where RTK has satellite coverage. If the facility has 2 accessible entries, exit at the far end to re-establish RTK Fixed status, then return along the same corridor to close the loop from the opposite direction. The backbone session captures corridor geometry that every subsequent pod session will reference during registration.

Secondary Aisle Circuits (Pod Sessions)

Enter each pod from the corridor and scan aisles in serpentine U-turn loops, weaving between rack rows as shown in the diagram. Return to the pod entrance between each aisle pair to re-verify the local reference frame against the corridor geometry already captured by the backbone. Each return to the corridor creates a loop closure that constrains drift accumulated in the aisle. Perform a 360° rotation at each aisle entrance before and after each U-turn pair.

Transition Verifications Between Pods

When moving between pods through the shared corridor, pause at the transition point and scan backward briefly toward the pod you just completed. This "look-back" links the 2 spaces by giving SLAM overlapping geometry from both sessions at the boundary. Then continue into the next pod session.

Quality Control and Export

Review Drift in Real Time

After completing each pod session, use LixelGO's mobile viewer to check for visible slippage in long aisles before concluding. Look for doubled walls, stepped floor lines, or rack geometry that does not align at the overlap zone. If drift is visible, rescan the affected aisles while the setup is still in place. Discovering drift after leaving the site means a return trip.

Target Verification

Confirm that at least 3 georeferenced targets are clearly visible in the scan data and that each was captured from multiple angles. Targets seen from only 1 direction may not register reliably during processing. Cross-check the target count in LixelGO against your scan plan before powering down.

Post-Processing Optimization

Apply global bundle adjustment during LixelStudio processing to tighten registration across all sessions. Enable moving object removal to eliminate personnel, carts, or equipment that moved during capture. The result is a clean deliverable with consistent geometry across the full facility extent.

GPS-Denied Facilities

The scan plan above assumes outdoor access at building entries for RTK initialization. Many modern data centers are sealed facilities with no exterior-accessible entry points. Colocation facilities, underground vaults, and interior suites within larger buildings may have no path to open sky at any entry.

If starting or ending a session outdoors is impossible, the framework still works, but absolute georeferencing relies entirely on the internal target network. Pre-survey all GCP locations with a total station referenced to an exterior control network before scanning begins. The surveyed coordinates provide the same absolute positioning that RTK would deliver at the threshold, without requiring satellite coverage at any point during the scan.

In a GPS-denied facility, the GCP network is the only source of absolute coordinates. There is no fallback. Plan target density at the tighter end of the spacing guidelines (50 ft for K1, 100 ft for L2 Pro) and ensure targets are distributed in 3 dimensions, not just on the floor. Magnetic plates at varying rack heights and reference spheres in corridors provide the vertical distribution that floor-only stickers cannot.

Legend and Key Specifications

Scan Routes

Session 1: Backbone cross
Session 2: Pod A
Session 3: Pod B
Direction of travel
Loop closure (line intersection)

Target Network

Threshold GCP (entry pair)
Floor GCP sticker
Magnetic checkerboard
Reference sphere

Zones

Exterior / sky view
Corridor anchor zone
Session overlap zone

System Specifications

GCP spacing (L2 Pro, data center)≤ 100 ft
GCP spacing (K1)≤ 50 ft
Session overlap minimum≥ 50 ft
RTK carry-in range~330 ft from entry
Walking speed (indoor/aisle)0.5 m/s max
Rows per session2 to 4
Map Fusion segment limit10 segments
Map Fusion total time limit200 minutes
Map Fusion per-segment limit20 minutes
RTK initialization route33 x 33 ft L-shape minimum

On-Site QA/QC Checklist

Check
Tolerance
What Failure Means
Rack position and aisle width
3/16 to 3/8 in
Future rack or containment clash in coordination model
Overhead clearance
Project defined, typically 3/8 in
New tray or busway routes conflict with existing structure
Control point residuals
3/4 in or less for BIM-grade
Target placement error or trajectory instability
Aisle linearity
No visible kinks or steps
SLAM drift or mis-registration between sessions
Pod-to-pod alignment
Continuous ceiling lines across boundaries
Aliasing or insufficient overlap at corridor zone

Return to the full Data Centers guide for technique details, target placement rationale, and advanced trajectory design.

3.5 Data Centers →

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