About AliceVision
AliceVision is an open source photogrammetric computer vision framework for 3D reconstruction and camera tracking. It infers the geometry of a scene from a set of unordered photographs or video frames by chaining Structure from Motion and Multi View Stereo pipelines into a dense 3D model. The project also covers HDR panoramas, photometric stereo and RAW conversion.
Most people use it through Meshroom, the project's nodal desktop application. Meshroom uses GPUs for quality and speed, scales past 1,000 images, caches results in its node editor, and can extend a solved reconstruction with new images. Meshes can be decimated and retextured after retopology in Blender, ZBrush or Maya, using custom or generated UVs. The HDR panorama workflow fuses bracketed exposures, aligns panorama images and supports fisheye lenses and motorized head files. Jobs can be submitted to a render farm with per node rules for CPU, RAM and GPU, and custom pipelines can be saved as templates. Blender plugins move scenes between the two tools, and a Houdini plugin from SideFX runs AliceVision reconstruction inside Houdini. Companion libraries include CCTag for sub pixel marker detection and PopSIFT, a GPU implementation of SIFT. Use cases shown include film, cultural heritage and entomology.
AliceVision and Meshroom are free open source software with no paid plans. The ALICEVISION Association asks institutions and companies that use them to become corporate sponsors, and the team invites contact for specific pipeline customizations.
Meshroom has its own listing; this entry covers the underlying framework and its libraries.