struct mp2p_icp_filters::FilterTransformPointCloud::Parametersļƒ

Overviewļƒ

#include <FilterTransformPointCloud.h>

struct Parameters
{
    // fields

    std::string input_pointcloud_layer;
    std::string output_pointcloud_layer;
    mrpt::math::TPose3D pose;
    bool invert_pose = false;

    // methods

    void load_from_yaml(
        const mrpt::containers::yaml& c,
        FilterTransformPointCloud& parent
        );
};

Fieldsļƒ

mrpt::math::TPose3D pose

In the context of a mola_lidar_odometry pipeline, this is typically:

pose: [robot_x, robot_y, robot_z, robot_yaw, robot_pitch, robot_roll]

Each of the 6 YAML sequence entries is bound as-is into the corresponding mrpt::math::TPose3D field, so angles must already be in RADIANS (TPose3Dā€™s own convention) e.g. the robot_yaw / robot_pitch / robot_roll dynamic variables above are, since they come from mrpt::poses::CPose3D::yaw()/pitch()/roll().

bool invert_pose = false

If false (default), each point is replaced by pose \oplus p_i. If true, by pose^{-1} \oplus p_i instead.