Parallax CPP

This library uses a novel approach at estimating the essential matrix. The essential matrix is then used to determine which 3D world points are actually moving. For more information see the wiki.

Install

This library is build to be installed onto your machine. To install it, clone the repository, and navigate to the directory.

mkdir build
cd build
cmake ..
make
sudo make install

If you have multiple versions of OpenCV on your system, you can specify which version to use by changing the OpenCV_DIR tag. For example

mkdir build
cd build
cmake -DOpenCV_DIR=/usr/local/share/OpenCV ..
make
sudo make install

How to use in your own project

// Include the parallax detector library
#include "parallax_detector/parallax_detector.h"

// Declare a Parallax Detector Object
gnsac::ParallaxDetector  parallax_detec_;

// Class object that stores the essential matrix, rotation matrix, translation vector and others. See below for more information
common::EHypothesis results_

// Initialize the Parallax Detector Object 
// @param gnsac_solver_filename The absolute file path to the file with all of the gnsac parameters. If 
//                              the file name is empty, default parameters will be used.
parallax_initiated_ = parallax_detec_.Init(gnsac_solver_filename);

// Run the parallax compensation algorithm which computes and returns essential matrix, rotation matrix, and translation vector.
// @param ud_prev_matched_ Calibrated image points in the previous frame
// @param ud_curr_matched_ Corresponding calibrated image points in the current frame
// @param moving_parallax_ Output vector of bools corresponding to each matched points. If moving_parallax_[i] is true, then point
//                         ud_curr_matched_[i] is considered moving perpendicular to the epipolar lines and has velocity in the world frame.
results_ parallax_detec_.ParallaxCompensation(ud_prev_matched_, ud_curr_matched_, moving_parallax_);

common::EHypothesis

class EHypothesis
{
public:
	EHypothesis();
	EHypothesis(Eigen::Matrix3d& E);
	EHypothesis(Eigen::Matrix3d& E, Eigen::Matrix3d& R, Eigen::Vector3d& t);
	Eigen::Matrix3d E;            // Essential Matrix
	Eigen::Matrix3d R;            // Rotation Matrix
	Eigen::Vector3d t;            // Translation Vector
	double cost;                  // Resulting Error from the optimization algorithm used to compute the essential matrix
	bool has_RT;                  // If true, R and t have been calculated
};

Parameters

See the wiki