openlidarviewer

What the viewer does

OpenLiDARViewer is a browser native LiDAR and point-cloud viewer for fast local inspection, 3D navigation, measurement, and terrain analysis. The README describes it as local-first, meaning files are read and rendered on the user's device with nothing uploaded to a server. A user opens the hosted page at lidar.aurtech.mx, then drags a .las, .laz, or .copc.laz file onto the page, or pastes a remote COPC or ept.json URL, and begins navigating the cloud. The project frames itself as a viewer and inspection tool rather than a GIS, photogrammetry, or survey-grade processing suite. Every result it produces discloses its coverage, method, and uncertainty, and terrain and contour exports are evidence-gated so they refuse to over-claim. Supported import formats include LAS, LAZ, E57, PLY, OBJ, GLB, GLTF, XYZ, CSV, PCD, PTX, and PTS, while exports cover LAS, PLY, OBJ, XYZ, CSV, and PNG snapshots. The README stresses that the tool does not claim survey-grade measurement or support for every LiDAR format, and it documents its limitations openly. This honest framing is a recurring theme of the project, and it is backed by a fair-software badge and a DOI of 10.5281/zenodo.21544619 that make the work citable for research use. The documentation site at aurtechmx.github.io/openlidarviewer hosts the format matrix and the scientific validation record for readers who want the deeper detail behind the evidence-gated exports described in the README.

Rendering, navigation and measurement

The viewer renders through WebGPU with an automatic WebGL2 fallback, so it runs in modern Chromium browsers with hardware acceleration and degrades gracefully elsewhere. Navigation is game-like, with Orbit, Walk, Fly, and Pan modes driven by WASD and mouse-look, plus Top, Iso, Oblique, and Planar camera presets. The rendering stack includes Eye Dome Lighting for depth shading, a hillshade relief overlay, soft splat rendering, and several colour modes such as height, intensity, classification, RGB, and surface normal. Measurement covers distance, polyline, area, height, angle, slope, and cross-section profile, with editable points and JSON session export and import. A confidence-aware DTM and contour pipeline performs ground classification, builds a gridded DTM with hold-out RMSE, and produces surface models, with evidence-gated contour and DEM export. Point inspection reveals coordinates, intensity, classification, GPS time, and colour, and a Scan Intelligence panel reports point count, dimensions, density, and spacing. The tool also supports annotations, multi-page PDF reports, and a command palette reachable with Cmd-K. A triangular nav widget surfaces the active mode, and saved camera views let an inspector return to a known framing for repeatable review of the same area across sessions. A curated catalog of 12 hand-vetted public COPC and EPT datasets ships with the viewer, so a new user can explore real scans without first sourcing their own files.

Streaming, accessibility and project standing

For large datasets the viewer streams COPC and EPT progressively, node by node, with bounded memory so a survey never requires a full-file load. A curated catalog of 12 hand-vetted public COPC and EPT datasets is included, and a 3D Tiles or .pnts path is reported as detected but not yet mounted as a streaming layer. Accessibility features include a colourblind-safe Okabe-Ito classification palette, a command palette for keyboard-first access, a searchable shortcut sheet, an onboarding tour, and a mobile touch model with twist, pinch, and pan decomposition. The repository metadata records 21 stars, 2 forks, and a last update of 2026-08-24, and the README shows CI, security audit, coverage, and mutation score badges alongside the DOI already mentioned. System requirements call for a modern Chromium browser with WebGL2 and hardware acceleration, with Firefox and Safari supported through the WebGL2 fallback. The project's stated goal is to make opening a point cloud about as easy as opening an image while preserving the depth and inspection tools LiDAR work needs, and it ships a workflow recorder that captures camera moves and tool actions as replayable files without ever storing the scan data itself. The project carries a fair-software badge and a DOI, which makes the work citable for research use and signals that the authors treat the release as a maintained scientific tool.

Editorial conclusion

The project is written in TypeScript, carries the MIT license, and the metadata records 21 stars with a last update of 2026-08-24.