OpenMVS: multi-view stereo reconstruction library

What it supplies: OpenMVS is a library aimed at the multi-view stereo reconstruction community. The README says it provides dense point-cloud reconstruction, mesh reconstruction, mesh refinement to recover fine detail and mesh texturing, and complements structure-from-motion pipelines such as OpenMVG. A script, MvsScalablePipeline.py, runs the whole chain for scenes too large for memory by splitting them into sub-scenes. Useful for: - Turn posed photographs (from a structure-from-motion tool) into a dense point cloud, a mesh and a textured mesh. - Process very large scenes by splitting them. First action: - Read the building wiki page and build for Windows, Ubuntu or macOS, or take the automatic Windows x64 builds from the artifacts page. - Follow the usage example wiki page. - Run scripts/python/MvsScalablePipeline.py scene.mvs for large scenes. Prerequisites and limits: Input is a scene produced by an SfM tool. The README mentions Windows, Ubuntu and macOS x64 continuous integration and gives no hardware figures in the text read. - The README defers to the wiki for most detail. - It covers the dense stage, not camera pose estimation. The README only says to see the copyright file for the licence, so check the repository's licence and copyright files before use. Reconstructed models carry the rights of the photographed subjects. Written from the README text only. Statements are as the README gives them; nothing was built or run. Source inspected: https://raw.githubusercontent.com/cdcseacave/openMVS/14ad2571aac650bfabf54296d31aabed7a19fdde/README.md ; 2680 bytes ; sha256 836510edb96a1c95099770bab7a8d0f49c4bb003b35eeb7d16509ae7acfba8a0 ; whole README text

{
  "revision": {
    "ref": "commit 14ad2571aac650bfabf54296d31aabed7a19fdde",
    "pinned": false,
    "source_url": "https://raw.githubusercontent.com/cdcseacave/openMVS/14ad2571aac650bfabf54296d31aabed7a19fdde/README.md"
  },
  "checked_at": "2026-11-07T11:23:07.590Z",
  "usefulness": "- Turn posed photographs (from a structure-from-motion tool) into a dense point cloud, a mesh and a textured mesh.\n- Process very large scenes by splitting them.",
  "source_type": "repository",
  "availability": "available",
  "source_owner": "cdcseacave (OpenMVS) (declared repository namespace)",
  "known_limitations": [
    "Input is a scene produced by an SfM tool. The README mentions Windows, Ubuntu and macOS x64 continuous integration and gives no hardware figures in the text read.\n\n- The README defers to the wiki for most detail.\n- It covers the dense stage, not camera pose estimation.\n\nThe README only says to see the copyright file for the licence, so check the repository's licence and copyright files before use. Reconstructed models carry the rights of the photographed subjects.\n\nWritten from the README text only. Statements are as the README gives them; nothing was built or run. ",
    "Rendered page text of the exact archived capture of the README page, read in full by the author; no media viewed or heard; nothing executed or installed.",
    "Source-documented steps; execution not performed. Internal affiliated review is not independent validation."
  ]
}