C++ library implementing triangulation methods described by Hartley & Sturm.
The shared library can be built and installed using CMake.
mkdir build && cd build
cmake ..
make && make install
There are 6 triangulation methods in total:
- Linear-LS
- Linear-Eigen
- Iterative-LS
- Iterative-Eigen
- Poly
- PolyAbs
For each method there is a separate class in Triangulation namespace with constructor that expects two camera projection matrices P0, P1. The points can be triangulated using the triangulate method which expect either two points or two vectors of points.
#include <Triangulation/LinearLS.h>
#include <Triangulation/IterativeLS.h>
#include <Triangulation/LinearEigen.h>
#include <Triangulation/IterativeEigen.h>
#include <Triangulation/Poly.h>
#include <Triangulation/PolyAbs.h>
...
cv::Mat P0, P1; // 3x4 camera projection matrices
cv::Point2d point0, point1; // triangulated image points
cv::Point3d point; // triangulation result
...
point = Triangulation::LinearLS(P0, P1).triangulate(point0, point1);
point = Triangulation::IterativeLS(P0, P1).triangulate(point0, point1);
point = Triangulation::LinearEigen(P0, P1).triangulate(point0, point1);
point = Triangulation::IterativeEigen(P0, P1).triangulate(point0, point1);
point = Triangulation::Poly(P0, P1).triangulate(point0, point1);
point = Triangulation::PolyAbs(P0, P1).triangulate(point0, point1);
...
If you find this library useful, please cite:
HARTLEY, Richard I.; STURM, Peter. Triangulation. Computer vision and image understanding, 1997, 68.2: 146-157.