From raw points.
To ground truth.
Ingest LiDAR and drone survey data at scale, classify every return in under three hours per billion points, and export canopy height models, building footprints, and volumetric reports — all via OGC API Processes.
- 500B+
- Points processed
- < 3 hrs
- Classification time per 1B points
- ± 2 cm
- Absolute vertical accuracy
- LAZ + COG
- Lossless compressed output
Every return. Every class.
Six automated pipelines turn raw LiDAR and photogrammetric point clouds into analysis-ready products with no manual preprocessing.
Ground Classification
Full ASPRS standard class scheme: Ground (2), Low Vegetation (3), Medium Vegetation (4), High Vegetation (5), Building (6), and Water (9). Progressive Morphological Filter (PMF) and Cloth Simulation Filter (CSF) run in ensemble to handle complex terrain including bridges, viaducts, and dense urban canyons.
Canopy Height Model
CHM derived by subtracting the Digital Terrain Model (DTM) from the Digital Surface Model (DSM) at up to 0.25 m resolution. Automated individual tree crown delineation uses watershed segmentation with configurable minimum crown area and height thresholds.
Building Footprint Extraction
3D building footprints derived from point density analysis on classified building points. Multi-return geometry reconstruction produces LOD-1 and LOD-2 roof forms. Outputs to IFC 2x3/IFC4, CityGML, and GeoJSON for direct ingestion into BIM workflows and planning systems.
Volumetric Analysis
Stockpile volume measurement using triangulated base-plane and convex hull methods, accurate to 0.1% of total volume. Earthwork cut-and-fill computation compares pre- and post-construction surveys to generate balance sheets and haul schedules directly from point cloud differencing.
Noise Filtering
Statistical outlier removal (SOR) identifies points whose mean distance to k-nearest neighbours exceeds a configurable standard-deviation threshold. Isolated-point detection eliminates single-return artefacts from water surfaces, glass facades, and atmospheric scatter before classification.
Format Support
Ingests LAZ, LAS 1.4, E57, PLY, and XYZ ASCII point clouds. Output to Cloud-Optimised GeoTIFF (COG) for derived rasters, GeoPackage for vector results, and Entwine Point Tiles (EPT) for streaming to web viewers. Maximum single-job input is 2 TB uncompressed.
Built for production-scale survey data
| Parameter | Value |
|---|---|
| Point density | Up to 500 pts/m² input |
| Classification | ASPRS standard class codes 1–18 |
| Output formats | LAZ, COG-DEM, GeoPackage, IFC |
| Vertical accuracy | ±2 cm absolute with certified control |
| API standard | OGC API Processes, REST |
| Coordinate systems | All EPSG, ICSM GDA2020 native |
| Processing engine | Distributed PDAL + ENTWINE pipeline |
| Storage | Cloud-optimised EPT (Entwine Point Tiles) |
Ready to classify a billion points?
Book a personalised demo and we will walk through ground classification, canopy height modelling, and volumetric workflows for your survey project.