3D rigid transform with 6 degrees of freedom. Transforms point x from a to b as: x_in_b = R * x_in_a + t.

Track class stores all observations of a 3D point.

Float32 kernel with exact predicates for memory efficiency and speed. Uses filtered predicates wrapping Simple_cartesian<float> to maintain robustness

Wrapper class around bitmaps.

3D similarity transform with 7 degrees of freedom. Transforms point x from a to b as: x_in_b = scale * R * x_in_a + t.

Feature descriptors with associated extractor type metadata.

Two-view geometry.

Wrapper for Rotation3d that can hold either owned data or a view into existing memory (e.g., Rigid3d.params). This enables zero-copy access to rotatio

Forward declaration for conversion methods.

Rigs represent a collection of rigidly mounted sensors and the associated sensor from rig transformations. The reference sensor is defined to have ide

Configuration container to setup bundle adjustment problems.

Solver-agnostic bundle adjustment options.

EPNP solver for the PNP (Perspective-N-Point) problem. The solver needs a minimum of 4 2D-3D correspondences. The algorithm is based on the following

Solver for the Generalized P3P problem.

Class that triangulates points during the incremental reconstruction. It holds the state and provides all functionality for triangulation.

Minimal solver for 6-DOF pose and focal length.

Abstract base class for sampling methods.

Triangulation estimator to estimate 3D point from multiple observations. The triangulation must satisfy the following constraints: - Sufficient triang

Worker-side helper that publishes the current phase for a scene to a shared dict. No-op when status_dict is None.

Exception that aggregates multiple exceptions from parallel tasks. Thrown by ThreadPool::Wait() when one or more tasks threw exceptions.

Transform ellipsoidal GPS coordinates to Cartesian coordinate systems and vice versa.

Options for CreateGeometricVerifier

Gravity prior for an image.

Class that holds information about an image. An image is the product of one camera exposure at a certain location (parameterized as the pose). An imag

Use color for images with specific words in their name.

Class that provides all functionality for the incremental reconstruction procedure. Example usage: IncrementalMapper mapper(&database_cache); mapper.

Class that controls the incremental mapping procedure by iteratively initializing reconstructions from the same scene graph.

Least absolute deviations (LAD) fitting via ADMM by solving the problem: min || A x - b ||_1 The solution is returned in the vector x and the iterat