Reality Capture, Delivered.

Interior + Complex Environment Scanning

High-density, colored point clouds captured at walking pace through occupied buildings, mechanical rooms, and active renovation sites. We capture with the Lixel L3, a handheld SLAM LiDAR scanner that moves through your space without setting up tripods or interrupting operations. The deliverable is registered, georeferenced, and ready for design and coordination workflows with no additional processing on your end.

Data centers, MEP-dense mechanical rooms, occupied offices, renovation projects, warehouse environments, and multi-floor facilities. Capture continues through tight corridors, stairwells, and complex layouts without interruption.

You receive: Registered, georeferenced point clouds in E57, LAS, LAZ, or RCP/RCS. RTK-grade absolute positioning available. Compatible with Revit, Navisworks, AutoCAD, Civil 3D, and GIS platforms.

Aerial Mapping: LiDAR + Photogrammetry

Design-ready site data: terrain geometry with the true ground surface under vegetation, photorealistic 3D reconstructions of buildings and vertical structures, and sharp true-color imagery of the entire site. Everything is georeferenced to your project coordinate system by an on-site base station, so the data drops into your design, engineering, and GIS workflows without adjustment.

Oblique photogrammetry documents every facade and rooftop plane for detailed reconstruction and inspection. Multi-return LiDAR (up to 16 returns per pulse) reaches the terrain beneath tree cover, so your surface models reflect the ground you will build on, not the canopy above it. Remote sites without cell coverage map at the same accuracy as urban ones, and a smaller drone platform handles tight, access-constrained sites. The fleet behind it: DJI Matrice 400, Zenmuse L3 LiDAR, Zenmuse P1 full-frame photogrammetry camera, and the D-RTK 3 base station, processed in-house in DJI Terra.

You receive: Aerial point clouds (LAS), orthomosaics (GeoTIFF), DSMs, DTMs, contour maps, classified terrain data, photogrammetric meshes (OBJ), and high-resolution still photography.

BIM and CAD Documentation

Point cloud to BIM and CAD deliverables at the level of detail your project requires. Architectural, MEPF, structural, and 2D scan-to-CAD work delivered through Alpine's certified production partner. We manage the full pipeline from scan data through delivered model. Scope, tolerances, and LOD are defined before modeling begins.

Capabilities span architectural BIM, MEPF (mechanical, electrical, plumbing, fire), structural modeling, federated coordination models, and 2D scan-to-CAD drawings. Outputs comply with US codes (ADA, IBC, NFPA). Models include FM-ready asset tagging and COBie fields when requested.

You receive: Architectural and structural Revit (RVT) models, MEPF Revit models, 2D documentation (DWG, PDF), Navisworks federated coordination models (NWC, NWD), IFC exports, and FM-ready models with asset tagging.

As-Built Digital Twins

Reality capture data converted into a navigable, mesh-based digital twin. The twin supports clash detection against actual site conditions and virtual layout studies, helping teams validate engineering decisions before committing to physical changes.

Data center operators maintaining current-state models of live facilities. Brownfield retrofits where new design must be validated against actual conditions. Layout planning where equipment placement and clearances are studied against real geometry.

You receive: Mesh-based digital twin hosted on a cloud platform with role-based access. Suitable for design review, clash analysis, and stakeholder collaboration.

Simulation-Ready 3D Environments

Real-world environments captured with the handheld Lixel L3 as LiDAR-fused 3D Gaussian Splats for robotics simulation, embodied AI training, world model development, and real-time visualization. One capture produces two aligned data layers: a photorealistic 3DGS model for perception, and a georeferenced LiDAR point cloud for geometry, physics, and measurement. Both share one coordinate system, so they load together without manual alignment. Where camera-only capture breaks down (glass, blank walls, repetitive textures, foliage), the LiDAR framework holds the geometry while the splat carries the lighting, transparency, and specular detail.

Three markets run on this data. Robotics and embodied AI teams import OpenUSD environments into NVIDIA Isaac Sim and Omniverse to train navigation, manipulation, and perception in photoreal replicas of real facilities before deploying to hardware. Industrial operators use them for remote inspection and hazardous-site training in spaces like pump rooms, substations, and active construction. Virtual production teams bring real locations onto LED walls and real-time sets, including the surfaces traditional photogrammetry cannot handle.

Homes, offices, restaurants, warehouses, industrial floors, stadiums, and data centers. We capture the real operating environments AI systems need to train on and that AI teams cannot easily access themselves. Available as capture-for-hire on your sites or under data licensing agreements.

You receive: Photorealistic 3D Gaussian Splat models in PLY, OpenUSD (.usd) for NVIDIA Isaac Sim and Omniverse pipelines, and 3D Tiles for browser-based streaming via Cesium; registered georeferenced point clouds (E57, LAS, LAZ); capture metadata (sensor, positioning method, coordinate system, capture date); and web-shareable review links. Mesh geometry and additional formats scoped per project.