XGRIDS Pro Guide™ / Module 9: LCC Studio

9.4 Aerial-Ground Fusion Processing

Processing requirements, data organization, and step-by-step procedure for combining ground-level scan data with drone aerial imagery in LCC Studio.

What Aerial-Ground Fusion Produces

Aerial-Ground Map Fusion combines ground-level scan data with drone aerial imagery to produce a unified 3DGS model covering both the interior ground-level geometry and the building exterior, roof, and surrounding site. The ground data provides precise interior geometry. The drone imagery provides aerial coverage of areas inaccessible from ground level.

This mode is appropriate for campus facilities, construction sites, heritage documentation, and any project where roof or aerial context is required alongside the interior scan. It is not necessary for interior-only projects regardless of how large they are.

Processing Requirements

Aerial-Ground Fusion has significantly higher hardware requirements than single-scene or map fusion processing. Attempting it on hardware below the minimum produces either a processing failure or an unusable result.

Requirement
Specification
CPU
16-core minimum
RAM
96 to 128 GB
GPU
NVIDIA RTX 4090
Ground data input
Raw scanning data only. The output of a completed Map Fusion project cannot be used as ground data input. Only unprocessed scan segments are accepted.
Ground segment limits
Maximum 10 segments, total duration under 200 minutes. These are the same constraints as standard Map Fusion.
Drone imagery
100 to 10,000 images (JPG/JPEG). Resolution must be 1024×768 or higher, consistent across all images.
Drone image overlap
≥85% overlap between adjacent frames. Use a tic-tac-toe flight pattern for best results.
RTK requirement
Both the drone and the Lixel scanner must have RTK enabled during data collection. Coordinate system must be set to WGS84. Weak or intermittent RTK signal during either ground or aerial collection degrades or prevents fusion.
Processing time
24 to 48 hours or more for large sites
App and firmware version
LCC Scan v1.2.0_p1 or higher; firmware V3.2.3-20251104.144651 or higher (for PortalCam ground collection). Verify before the site visit.

Ground Data Constraints

  • Requirement Aerial-Ground Fusion accepts raw scanning data only as ground input. You cannot process a Map Fusion project first and then feed the fused output into an Aerial-Ground Fusion project. The ground data slots accept raw, unprocessed scan segments directly from the device. If your project requires both ground-level Map Fusion and aerial coverage, plan the ground segments to fit within the 10-segment/200-minute limit of a single Aerial-Ground Fusion job.
  • Warning Monitor the Estimated RAM bar after importing all ground and aerial data. If the bar turns red, the combined dataset exceeds available system capacity and reconstruction will likely fail. Reduce the number of ground segments, reduce drone image count, or increase system resources before starting.

Processing Procedure

  1. Organize your data before opening LCC Studio. Keep ground scan segments in clearly labeled individual folders. Keep drone imagery in a separate folder labeled with a _Drone or _Aerial suffix at the same directory level. Each ground segment must be a raw, unprocessed scan folder transferred directly from the device.
  2. Click Create and select Aerial-Ground Map Fusion. This mode is distinct from Map Fusion Reconstruction and Single Scene.
  3. Add the ground scan data under Ground Data. Add the drone image folder under Aerial Data. LCC Studio reads the fusion point markers embedded in the ground scan to align the aerial imagery. Verify the ground segment count does not exceed 10 and the combined duration stays under 200 minutes.
  4. Check the Estimated RAM bar. After importing all data, confirm the bar has not turned red. A red bar indicates the project exceeds available capacity. Reduce scope or increase resources before proceeding.
  5. Verify fusion points are detected. LCC Studio displays which fusion points from the ground scan are recognized. A minimum of 2 to 3 fusion points must be present and distributed across the site. Points clustered at one end of a large site do not provide adequate geometric constraint across the full model.
  6. Configure reconstruction settings. The same quality, Max Gaussian Splats, portability, and SLAM mode settings apply. Max Gaussian Splats in this mode applies per chunk with auto-adjustment. Enable Exposure Optimization if the ground scan includes both indoor and outdoor environments.
  7. Start processing and allow the full estimated duration. A 24 to 48 hour run is normal for large sites. Do not interrupt. Do not run LixelStudio processing simultaneously on the same machine. Schedule this run to complete before any delivery deadline.
  8. Review boundary zones between ground and aerial coverage in the viewer. The transition between ground-level interior geometry and aerial exterior geometry should be seamless at the building envelope. Discontinuities indicate fusion point placement issues from the field collection phase.

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