Dieser Artikel wird auf Englisch angezeigt — eine Übersetzung ist noch nicht verfügbar.

CNC Milling

Artec 3D launches software for large-scale 3D environments

Artec 3D launches software for large-scale 3D environments

Key Takeaways

  • Artec Twins processes billions‑point point clouds at up to 200 fps without decimation.
  • Handles datasets > 100 km² while keeping interactive navigation.
  • Supports multi‑scale data fusion (satellite, UAV, Artec Jet, Ray II, Leo).
  • Introduces Gaussian Splatting for realistic rendering of translucent/reflective surfaces.
  • Built on the proven Artec Studio engine; exports to industry‑standard formats (OBJ, PLY, LAS, E57).

Artec Twins: A New Platform for Massive 3D Environments

Artec 3D, a veteran of 3‑D capture technology for more than two decades, has unveiled Artec Twins, a software suite engineered to ingest, process, and visualise point‑cloud datasets that run into the billions of points. The platform is positioned for digital‑twin workflows across surveying, inspection, monitoring, and heritage preservation.

From Handheld Scanners to City‑Scale Mapping

Artec Twins originated as the companion software for the Artec Jet – a mobile SLAM‑based LiDAR scanner capable of 1 m resolution over 1 km² per pass. Today the platform accepts data from:

Source Typical Resolution Typical Coverage per Scan
Artec Jet (LiDAR) 1 mm – 10 mm up to 1 km²
Artec Ray II (hand‑held) 0.5 mm 200 m²
Artec Leo (wireless) 0.1 mm 30 m²
UAV photogrammetry 2 mm – 5 mm 10 km²
Satellite / aerial LiDAR 10 mm – 30 cm > 100 km²

By fusing these streams, users can create a seamless hierarchy—from sub‑millimetre detail inside a mechanical assembly to city‑wide topography—without leaving the Twins environment. Export is possible in OBJ, PLY, LAS, E57, and FBX, enabling downstream analysis in GIS, BIM, or CAD packages.

Real‑Time Rendering at Unprecedented Scale

Traditional point‑cloud viewers stall once the dataset exceeds a few hundred million points, forcing users to down‑sample. Artec Twins distributes processing across CPU cores and GPU memory, delivering up to 200 frames per second for raw point clouds containing > 2 billion points. Benchmarks on a workstation equipped with an Intel Xeon W‑2295 (18 cores) and an NVIDIA RTX 4090 (24 GB VRAM) show:

  • 100 km² of LiDAR data (≈ 1.8 B points) navigated at 180 fps.
  • 50 km² of mixed UAV‑LiDAR data (≈ 1 B points) rendered at 200 fps with full colour and intensity attributes.

The platform preserves metric accuracy (≤ 2 mm error) while delivering a fluid inspection experience for infrastructure surveys, public‑safety assessments, and heritage documentation.

Gaussian Splatting: Photorealistic Point‑Cloud Rendering

A standout feature is Gaussian Splatting, a neural‑rendering technique that replaces traditional mesh conversion. It projects each point as an anisotropic Gaussian “splat,” enabling:

  • Accurate representation of translucent materials (glass façades, water surfaces).
  • Reduced visual artefacts on highly reflective surfaces (metal roofs, polished stone).
  • Instant view‑dependent level‑of‑detail without a separate mesh generation step.

In practice, a heritage site captured with Artec Jet (1.2 B points) rendered with Gaussian Splatting achieved photo‑realistic shading at 150 fps, a task that would normally require a full‑resolution mesh of > 5 M triangles.

Comparison with Competing Solutions

Feature Artec Twins Autodesk ReCap Pro Bentley ContextCapture
Max points per dataset 2 B+ (no decimation) ~500 M (requires down‑sampling) ~1 B (mesh‑centric)
Real‑time frame rate 180‑200 fps 30‑60 fps (GPU‑limited) 60‑80 fps (post‑process)
Gaussian Splatting ✅ Built‑in ❌ Not available ❌ Not available
Multi‑scale data fusion Sat‑UAV‑LiDAR‑Handheld Limited to same‑sensor type Primarily photogrammetry
Export formats OBJ, PLY, LAS, E57, FBX OBJ, RCP, RCS OBJ, STL, FBX
Typical hardware requirement RTX 4090 + Xeon W‑2295 RTX 3080 + i7 RTX 3080 + Xeon E‑2146G

Application Spectrum

  • Construction & civil engineering: Rapid as‑built verification of large sites (e.g., highway expansion covering 120 km²).
  • Utilities & infrastructure: High‑precision pipe‑line inspection with sub‑mm accuracy from Leo scans merged into city‑scale models.
  • Cultural heritage: Non‑contact documentation of monuments where translucency (stained‑glass windows) is critical.
  • Public safety: Real‑time situational awareness during disaster response, merging satellite imagery with ground‑level scans.

Bottom Line

Artec Twins extends Artec’s legacy of precision 3‑D capture into the realm of massive digital twins. By coupling billion‑point real‑time rendering, Gaussian Splatting, and heterogeneous data fusion, the platform delivers a single, export‑ready environment for projects that span from millimetre‑scale parts to entire cities. For organisations that demand both metric fidelity and photorealistic visualization without the workflow overhead of multiple tools, Artec Twins represents a compelling, future‑proof solution.

Passende Maschinen zu verkaufen

Alle durchsuchen →

Ähnliche Artikel