在点云库 (PCL） 中使用自定义点类型的问题

我在尝试使用 PCL 注册两个点云时遇到了问题。我没有使用提供的 PointT，而是使用了自定义点类型并遵循了教程(http://www.pointclouds.org/documentation/tutorials/adding_custom_ptype.php#adding-custom-ptype(。但是，当我在VS2019中编写代码时，它失败了。我也尝试了这篇文章(PCL 实例化所有函数的新点类型(。它也没有用。

这是要点的定义

#ifndef PCL_NO_PRECOMPILE
#define PCL_NO_PRECOMPILE
#endif
#include <iostream>
#include <pcl/pcl_macros.h>
#include <pcl/point_types.h> 
#include <pcl/impl/point_types.hpp>
#include <pcl/impl/instantiate.hpp>
#include <pcl/point_cloud.h>
#include <pcl/impl/point_types.hpp>
#include <pcl/common/transforms.h>
#include <pcl/common/impl/transforms.hpp>
#include <pcl/registration/icp.h>
#include <pcl/registration/impl/icp.hpp>
#include <pcl/features/fpfh.h>
#include <pcl/features/impl/fpfh.hpp>
#include <pcl/features/pfh.h>
#include <pcl/features/impl/pfh.hpp>
#include <pcl/registration/ia_ransac.h>
#include <pcl/registration/impl/ia_ransac.hpp>
#include <pcl/features/normal_3d.h>
#include <pcl/features/impl/normal_3d.hpp>
#include <pcl/kdtree/kdtree_flann.h>
#include <pcl/kdtree/impl/kdtree_flann.hpp>
#include <pcl/filters/voxel_grid.h>
#include <pcl/filters/impl/voxel_grid.hpp>
#include <pcl/filters/approximate_voxel_grid.h>
#include <pcl/filters/impl/approximate_voxel_grid.hpp>
#include <Eigen/Core>
struct dPointXYZ {
double x;
double y;
double z;
};
POINT_CLOUD_REGISTER_POINT_STRUCT(dPointXYZ,
(double, x, x)
(double, y, y)
(double, z, z)
)
PCL_INSTANTIATE(VoxelGrid, dPointXYZ);
PCL_INSTANTIATE(NormalEstimation, dPointXYZ);
PCL_INSTANTIATE(KdTree, dPointXYZ);
PCL_INSTANTIATE(FPFHEstimation,dPointXYZ);
PCL_INSTANTIATE(SampleConsensusInitialAlignment, dPointXYZ)

以下是自定义点类型的使用方式

void voxelFilteringPoints(pcl::PointCloud<pcl::dPointXYZ>::Ptr& ProjPoint, pcl::PointCloud<pcl::dPointXYZ>::Ptr& filteredPoints, float dGridSize)
{
pcl::VoxelGrid<pcl::dPointXYZ> ApproximateVoxelGrid;
ApproximateVoxelGrid.setLeafSize(dGridSize, dGridSize, dGridSize);
ApproximateVoxelGrid.setInputCloud(ProjPoint);
ApproximateVoxelGrid.filter(*filteredPoints);
//std::cout<<"Filtered points size: "<<*filteredPoints<<std::endl;
}
pcl::PointCloud<dNormal>::Ptr computeNormals(pcl::PointCloud<dPointXYZ>::Ptr& ProjPoint, double dNormalsRadius)
{
pcl::PointCloud<dNormal>::Ptr NormalsPtr(new pcl::PointCloud<dNormal>());
pcl::NormalEstimation<dPointXYZ, dNormal> NormEstimate;
NormEstimate.setInputCloud(ProjPoint);
NormEstimate.setRadiusSearch(dNormalsRadius);
pcl::search::KdTree<dPointXYZ>::Ptr PtrSearchMethod(new pcl::search::KdTree<dPointXYZ>);
NormEstimate.setSearchMethod(PtrSearchMethod);
NormEstimate.compute(*NormalsPtr);
//std::cout<<"Normals size: "<<*NormalsPtr<<std::endl;
return NormalsPtr;
}
pcl::PointCloud<pcl::FPFHSignature33>::Ptr getFPFHFeatures(pcl::PointCloud<dPointXYZ>::Ptr& ProjCloud, pcl::PointCloud<dNormal>::Ptr& ProjNormals, double dFeatureRadius)
{
pcl::PointCloud<pcl::FPFHSignature33>::Ptr FPFHFeatures(new pcl::PointCloud<pcl::FPFHSignature33>());
//std::cout<<"FPFH size: "<<FPFHFeatures<<std::endl;
pcl::search::KdTree<dPointXYZ>::Ptr PtrSearchMethod(new pcl::search::KdTree<dPointXYZ>);
pcl::FPFHEstimation<dPointXYZ, dNormal, pcl::FPFHSignature33> FPFHEstimate;
FPFHEstimate.setInputCloud(ProjCloud);
FPFHEstimate.setInputNormals(ProjNormals);
FPFHEstimate.setSearchMethod(PtrSearchMethod);
FPFHEstimate.setRadiusSearch(dFeatureRadius);
FPFHEstimate.compute(*FPFHFeatures);
return FPFHFeatures;
}
pcl::SampleConsensusInitialAlignment<dPointXYZ,dPointXYZ, pcl::FPFHSignature33> alignFPFH(pcl::PointCloud<dPointXYZ>::Ptr& ProjPoint1,
pcl::PointCloud<dPointXYZ>::Ptr& ProjPoint2, pcl::PointCloud<pcl::FPFHSignature33>::Ptr& FPFHProjFeatures1,
pcl::PointCloud<pcl::FPFHSignature33>::Ptr& FPFHProjFeatures2, int iMaxIterations, double dMinDist, double dMaxDist)
{
pcl::SampleConsensusInitialAlignment<dPointXYZ, dPointXYZ, pcl::FPFHSignature33> SCIA;
Eigen::Matrix4f transMatrix;
SCIA.setInputCloud(ProjPoint2);
SCIA.setSourceFeatures(FPFHProjFeatures2);
SCIA.setInputTarget(ProjPoint1);
SCIA.setTargetFeatures(FPFHProjFeatures1);
SCIA.setMaximumIterations(iMaxIterations);
SCIA.setMinSampleDistance(dMinDist);
SCIA.setMaxCorrespondenceDistance(dMaxDist);
pcl::PointCloud<dPointXYZ> registeredCloud;
SCIA.align(registeredCloud);
SCIA.getCorrespondenceRandomness();
return SCIA;
}

编译器指示已定义类模板。 错误 C2953 'pcl：:d etail：：转换器'：已定义类模板 C：\Program 文件\PCL_1.9.1\include\pcl-1.9\pcl\common\impl\transforms.hpp 89

错误 C2995 'void pcl：：transformPointCloud(const pcl：:P ointCloud &，pcl：:P ointCloud &，const Eigen：：Transform &，bool('：函数模板已定义 C：\Program Files\PCL_1.9.1\include\pcl-1.9\pcl\common\impl\transforms.hpp 215

错误 C2995 'void pcl：：transformPointCloud(const pcl：:P ointCloud &，const std：：vector> &，pcl：:P ointCloud &，const Eigen：：Transform &，bool('：函数模板已定义

我知道我可以通过规范化点来简单地使用浮点类型，但我想尝试自定义点类型。你有什么想法如何解决这个问题吗？