PoseLib/PoseLib

Health v2 · maintenance only

Adoption not part of the score

1203 stars · 153 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

PoseLib is a C++ library of minimal solvers for calibrated camera pose estimation, covering absolute and relative pose from point and line correspondences, homography, and generalized relative pose. It includes robust LO-RANSAC estimators with non-linear refinement and a consistent interface with minimal dependencies (only Eigen), plus Python bindings.

Use cases

• estimate camera pose from 2D-3D point correspondences
• compute essential or fundamental matrix from image matches
• run RANSAC-based robust pose estimation
• estimate homography between two images
• relative pose from points and monocular depth estimates
• integrate pose solvers into SfM or SLAM pipelines

When to choose

• you need fast, state-of-the-art minimal solvers for calibrated camera geometry
• you want robust estimators that work out-of-the-box with minimal dependencies
• you need a consistent API across many pose solvers or want to extract a single solver
• you work in C++ or Python on computer vision or robotics projects

When to avoid

• you need uncalibrated camera pose estimation or full SfM/SLAM systems
• you require GPU-accelerated or deep-learning-based pose estimation
• you need a high-level vision framework rather than a focused solver library

Facets

library · maturity active

computer-vision math sdk computer-vision robotics graphics cpp python cross-platform windows camera-pose-estimation ransac minimal-solvers epipolar-geometry sfm eigen bundle-adjustment algorithms linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="PoseLib/PoseLib")