PCL学习笔记——自定义pcl点云类型的创建

在进行多传感器融合后，往往会给点云数据赋予更多的信息，类似xyzrgb、法向量、强度等等点云数据，此时就需要我们对pcl库中的点云类型进行自定义操作，以此来满足我们点云数据处理时对点云数据结构的需求。

#define PCL_NO_PRECOMPILE   //如果想使用PCL提供的模板类，比如体素下采样，需要添加此行代码
#include <iostream>
#include <pcl/io/pcd_io.h>
using namespace std;


//自定义点云类型
struct PointXYZRGBNormalI
{
    PCL_ADD_POINT4D;  //添加XYZ，此时相当于 PointXYZ
    PCL_ADD_RGB;      //添加RGB，此时相当于 PointXYZRGB
    PCL_ADD_NORMAL4D; //添加NORMAL，此时相当于 PointXYZRGBNormal
    float intensity;  //添加intensity，此时相当于 PointXYZRGBNormalintensity
    PCL_MAKE_ALIGNED_OPERATOR_NEW   //确保new操作符对齐操作
}EIGEN_ALIGN16;   //强制SSE内存对齐


//注册点类型宏(要包含所有添加的字段，否则无法正常使用未注册的字段)
POINT_CLOUD_REGISTER_POINT_STRUCT(PointXYZRGBNormalI,
    (float,x,x)
    (float,y,y)
    (float,z,z)
    (uint8_t,b,b)
    (uint8_t,g,g)
    (uint8_t,r,r)
    (float,rgb,rgb)
    (float,normal_x,normal_x)
    (float,normal_y,normal_y)
    (float,normal_z,normal_z)
    (float,intensity,intensity)
);



int main() {
    pcl::PointCloud<PointXYZRGBNormalI> cloud; //定义一个自定义的点云数据结构
    cloud.width = 2;
    cloud.height = 1; //定义该点云为无序点云
    cloud.points.resize(cloud.width*cloud.height);
    cloud[0].x=cloud[0].y=cloud[0].z = 1;
    cloud[1].x=cloud[1].y=cloud[1].z = 2;
    cloud[0].r = cloud[0].g =cloud[0].b = 2;
    cloud[1].r = cloud[1].g =cloud[1].b = 3;
    cloud[0].intensity =cloud[1].intensity = 2;
    cloud[0].normal_x= cloud[1].normal_x =cloud[0].normal_y= cloud[1].normal_y=cloud[0].normal_z= cloud[1].normal_z =1;
    pcl::io::savePCDFile("saved_points.pcd",cloud);  //保存自定义的点云数据文件
    cout<<"the size of saved pcd:"<<cloud.points.size()<<endl;
    for (size_t i = 0; i < cloud.points.size(); ++i) //逐个输出该点云中的点
    {
        std::cout << cloud.points[i].x << "   "
                  << cloud.points[i].y << "   " << cloud.points[i].z<<"   "<<int(cloud.points[i].r)<<"   "<<int(cloud.points[i].g)<<"   "<<int(cloud.points[i].b)<<"   "<<cloud.points[i].normal_x<<"   "<<cloud.points[i].normal_y<<"   "<<cloud.points[i].normal_z<<"   "<<cloud.points[i].intensity<<std::endl;
    }

    return 0;
}