openMVG/openMVG

What it solves

OpenMVG provides a comprehensive framework for 3D reconstruction from images, also known as photogrammetry. It simplifies the process of turning a set of 2D images into a 3D scene by implementing state-of-the-art and classic algorithms for multiple view geometry.

How it works

The project is a C++ framework organized into three main components:

• Libraries: Provide core functionality for image manipulation, feature description and matching, feature tracking, camera models, and structure-from-motion (SfM) algorithms.
• Binaries: Standalone tools that perform specific unit tasks, such as scene initialization, feature detection, and SfM reconstruction.
• Pipelines: Sequences of binaries that chain together to compute image matching relations and solve the Structure from Motion problem (including reconstruction, triangulation, and localization).

Who it’s for

It is designed for researchers and developers who need a reproducible and maintainable way to perform 3D reconstruction and multiple view geometry, as well as those building larger trusted systems based on these capabilities.

Highlights

• End-to-end pipeline: Covers the entire process from raw images to 3D reconstruction.
• Cross-platform: Runs on Android, iOS, Linux, macOS, and Windows.
• Test-driven: Developed with strict test-driven development to ensure reliability.
• Export capabilities: Can export reconstructed scenes to other Multiple-View-Stereovision frameworks for dense point clouds or textured meshes.