2d雷达的原始数据生成点云地图pcd

sensor_msgs::LaserScan转换为sensor_msgs::PointCloud2
主要用到了transformLaserScanToPointCloud()这个函数，
订阅消息结构为sensor_msgs::LaserScan的话题scan，
发布消息结构为sensor_msgs::PointCloud2的话题cloud2
完整的功能包上传至
https://download.csdn.net/download/zhaohaowu/21227408
或者

1、创建功能包

catkin_create_pkg LaserScan2PointCloud2 roscpp rospy tf sensor_msgs laser_geometry

2、CMakeLists.txt

cmake_minimum_required(VERSION 3.0.2)
project(LaserScan2PointCloud2)

find_package(catkin REQUIRED COMPONENTS
  laser_geometry
  roscpp
  rospy
  sensor_msgs
  tf
)

catkin_package(
   CATKIN_DEPENDS laser_geometry roscpp rospy sensor_msgs tf
)

include_directories(
   ${catkin_INCLUDE_DIRS}

)

add_executable(LaserScan2PointCloud2 src/LaserScan2PointCloud2.cpp)

target_link_libraries(LaserScan2PointCloud2
  ${catkin_LIBRARIES}
)

3、src写两个文件
LaserScan2PointCloud2.cpp

#include "LaserScan2PointCloud2.h"

LaserScan2PointCloud2::LaserScan2PointCloud2(){
    sub = n.subscribe<sensor_msgs::LaserScan>("/scan", 100, &LaserScan2PointCloud2::callback, this);
    pub = n.advertise<sensor_msgs::PointCloud2>("/cloud2", 100);
    listener.setExtrapolationLimit(ros::Duration(0.1));
}

void LaserScan2PointCloud2::callback(const sensor_msgs::LaserScan::ConstPtr& scan){
    sensor_msgs::PointCloud2 cloud2;
    transformer.transformLaserScanToPointCloud("laser", *scan, cloud2, listener);
    pub.publish(cloud2);
}

int main(int argc, char** argv){
    ros::init(argc, argv, "LaserScan2PointCloud2");
    LaserScan2PointCloud2 LaserScan2PointCloud2;
    ros::spin();
    return 0;
}

LaserScan2PointCloud2.h

#include <ros/ros.h>
#include <tf/transform_listener.h>
#include <laser_geometry/laser_geometry.h>
#include <sensor_msgs/PointCloud2.h>
#include <sensor_msgs/LaserScan.h>

class LaserScan2PointCloud2{
    private:
        ros::NodeHandle n;
        laser_geometry::LaserProjection transformer;
        tf::TransformListener listener;
        ros::Publisher pub;
        ros::Subscriber sub;
    public:
        LaserScan2PointCloud2();
        void callback(const sensor_msgs::LaserScan::ConstPtr& scan);
};

4、消息结构为sensor_msgs::PointCloud2的话题cloud2（a.bag）转换一下话题名字points_raw（b.bag）

rosparam set use_sim_time true
rosbag record /points_raw
rosbag play --clock a.bag /cloud2:=/points_raw

这个步骤可以不要，在上面代码里将cloud2的话题直接改为points_raw
5、安装autoware
https://blog.csdn.net/zhaohaowu/article/details/119839147
6、消息结构为sensor_msgs::PointCloud2的话题points_raw以b.bag格式转换为pcd（a.pcd）

roslaunch runtime_manager runtime_manager.launch

setup里面点击TF和Vehicle Model
Map里面TF处点击Ref，加载

~/autoware.ai/src/autoware/documentation/autoware_quickstart_examples/launch/tf_local.launch

，然后点击TF
Computing里面点击lidar_localizer的ndt_mapping
Simulation里面点击Ref加载b.bag，点击play，再点击Pause，点击rviz
选择Fixed Frame为map，添加话题/ndt_map/PointCloud2，size修改为0.1
回到Runtime Manager，点击Pause，等待Playing结束
点击Computing中的lidar_localizer的ndt_mapping的app，再点击Ref选择pcd保存位置，点击PCD OUTPUT，生成a.pcd