\brief Get a pointcloud containing the voxel centroids * \note Only voxels containing a sufficient number of points are used. * \return

\brief Get the number of neighbors used when computing covariances as set by * the user */

\brief Get the voxel covariance. * \return covariance matrix */

\brief Get the minimum allowable ratio between eigenvalues to prevent singular covariance matrices. * \return the minimum allowable ratio betw

\brief Get the eigen values of the voxel covariance. * \note Order corresponds with \ref getEvecs * \return vector of eigen values

\brief Get the eigen vectors of the voxel covariance. * \note Order corresponds with \ref getEvals * \return matrix whose columns

\brief Get the number of iterations required to calculate alignment. * \return final number of iterations */

\brief Get the voxel containing point p. * \param[in] index the index of the leaf structure node * \return const pointer to leaf structu

\return maximum number of iterations at the optimization step

\brief Get the minimum number of points required for a cell to be used. * \return the minimum number of points for required for a voxel to be

\brief Get the point cloud outlier ratio. * \return outlier ratio */

\brief Get the number of points contained by this voxel. * \return number of points */

\brief Get voxel grid resolution. * \return side length of voxels */

\brief Get the rotation epsilon (maximum allowable difference between two * consecutive rotations) as set by the user. */

\brief Get the newton line search maximum step length. * \return maximum step length */

\brief Get the registration alignment probability. * \return transformation probability */

calculate residual v

* \brief Estimated the approximate overlapping ratio for Cloud1 considering Cloud2 * \param[in] Cloud1 : A pointer of the Point Cloud used for over

* \brief Use PCL to accomplish the Principle Component Analysis (PCA) * of one point and its neighborhood * \param[in] in_cloud is the input Point

Brief: calculate the Jacobi Matrix of the imaginary part of a quaternion (q1,q2,q3) with regard to its corresponding euler angle (raw,pitch,yaw) for c

Get Bound of Subsets of a Point Cloud

standard

Intuition: three part, near , medium , far near used to control translation far used to control rotation medium used to control both TODO: change the

By huber loss function

general function for m-estimation

simple implement of bi-threshold ground segmentation without further processing and downsampling

General implement of icp registration algorithm in pcl (baseline method) (with different distance metrics, correspondence estimation, transformation e

\brief Initiate covariance voxel structure. */

inverse and return

this function is for speed up the registration process when the point number is a bit too big

This is for the comparison between LeGO-LOAM , which used the two step LM. We'd like to try the two step LLS for the same task

For OXTS Ground Truth pose format: index timestamp x y z q1 q2 q3 q4

KITTI ground truth pose format mat(0-11) split with space example: 1.000000e+00 9.043680e-12 2.326809e-11 5.551115e-17 9.043683e-12 1.000000e+00 2.392

Note: Set these parameters in the config file , or the default values are used. ----------------------------------------------------------------------

map based active object removal for each non-boundary unground points of current frame, check its nearest neighbor in the local map, if the distance t

filter (0, near_dist_thre) U (dynamic_dist_thre_min, dynamic_dist_thre_max) keep (near_dist_thre, dynamic_dist_thre_min) U (dynamic_dist_thre_max, +in

\return trace of mat1^t . mat2 * \param mat1 matrix of dimension nxm * \param mat2 matrix of dimension nxp */

brief: entrance to mulls transformation estimation

Normal Space Downsampling (with normal already estimated)

unit (km/h)