Attention

You are viewing an older version of the documentation. The latest version is 2.2.

Estimating Surface Normals in a PointCloud

In this tutorial we will learn how to get the surface normals of each point in the cloud.

Note

This tutorial can be run both inside and outside a Docker* image. We assume that the pcl-oneapi-tutorial Deb package has been installed, and the user has copied the tutorial directory from /opt/intel/pcl/oneapi/tutorials/ to a user-writable directory.

1. Prepare the environment:

   cd <path-to-oneapi-tutorials>/normal_estimation
   

   

2. oneapi_normal_estimation.cpp should be in the directory with following content:

   #include <pcl/io/pcd_io.h>
   #include <pcl/oneapi/features/normal_3d.h>
   #include <pcl/oneapi/kdtree/kdtree_flann.h>
   #include <pcl/oneapi/point_cloud.h>
   
   int main (int argc, char** argv)
   {
     int k = 10;
     float radius = 0.01;
   
     std::cout << "Running on device: " << dpct::get_default_queue().get_device().get_info<sycl::info::device::name>() << "\n";
   
     // load point cloud
     pcl::PointCloud<pcl::PointXYZ>::Ptr cloud_ptr( new pcl::PointCloud<pcl::PointXYZ>() );
   
     int result = pcl::io::loadPCDFile("bun0.pcd", *cloud_ptr);
     if (result != 0)
     {
       pcl::console::print_info ("Load pcd file failed.\n");
       return result;
     }
   
     // estimate normals with knn search
     pcl::oneapi::NormalEstimation<pcl::PointXYZ, pcl::Normal> ne;
     ne.setSearchMethod (pcl::oneapi::KdTreeFLANN<pcl::PointXYZ>::Ptr (new pcl::oneapi::KdTreeFLANN<pcl::PointXYZ>));
     ne.setInputCloud(cloud_ptr);
     ne.setKSearch(k);
   
     // save normal estimation to CPU memory point cloud
     pcl::PointCloud<pcl::Normal>::Ptr normals_knn(new pcl::PointCloud<pcl::Normal>);
     ne.compute(*normals_knn);
   
     std::cout << "normals_knn.size (): " << normals_knn->size () << std::endl;
   
     // estimate normals with radius search
     ne.setSearchMethod (pcl::oneapi::KdTreeFLANN<pcl::PointXYZ>::Ptr (new pcl::oneapi::KdTreeFLANN<pcl::PointXYZ>));
     ne.setInputCloud(cloud_ptr);
     ne.setRadiusSearch(radius);
     ne.setKSearch(0);
   
     // save normal estimation output to device shared memory point cloud
     pcl::oneapi::PointCloudDev<pcl::Normal>::Ptr normals_radius(new pcl::oneapi::PointCloudDev<pcl::Normal>) ;
     ne.compute(*normals_radius);
   
     std::cout << "normals_radius.size (): " << normals_radius->size () << std::endl;
   
     return 0;
   }
   

   

3. Source the Intel® oneAPI Base Toolkit environment:

   source /opt/intel/oneapi/setvars.sh
   

   

4. (Optional) Setup proxy setting to download test data:

   export http_proxy="http://<http_proxy>:port"
   export https_proxy="http://<https_proxy>:port"
   

   

5. Build the code:

   mkdir build && cd build
   cmake ../
   make -j
   

   

6. Run the binary:

   ./oneapi_normal_estimation
   

   

7. Expected results example:

   normals_knn.size (): 397
   normals_radius.size (): 397
   

   

Code Explanation

The example PCD is initially loaded into a PointCloud<PointXYZ>.

  // load point cloud
  pcl::PointCloud<pcl::PointXYZ>::Ptr cloud_ptr( new pcl::PointCloud<pcl::PointXYZ>() );

  int result = pcl::io::loadPCDFile("bun0.pcd", *cloud_ptr);
  if (result != 0)
  {
    pcl::console::print_info ("Load pcd file failed.\n");
    return result;
  }

This tutorial includes two normal estimation processes: KNN search and radius search. The initial step involves adjusting the parameters for normal estimation in the KNN search.

  // estimate normals with knn search
  pcl::oneapi::NormalEstimation<pcl::PointXYZ, pcl::Normal> ne;
  ne.setSearchMethod (pcl::oneapi::KdTreeFLANN<pcl::PointXYZ>::Ptr (new pcl::oneapi::KdTreeFLANN<pcl::PointXYZ>));
  ne.setInputCloud(cloud_ptr);
  ne.setKSearch(k);

Normal estimation is then executed.

  pcl::PointCloud<pcl::Normal>::Ptr normals_knn(new pcl::PointCloud<pcl::Normal>);
  ne.compute(*normals_knn);

The parameters for normal estimation are modified for the radius search.

  // estimate normals with radius search
  ne.setSearchMethod (pcl::oneapi::KdTreeFLANN<pcl::PointXYZ>::Ptr (new pcl::oneapi::KdTreeFLANN<pcl::PointXYZ>));
  ne.setInputCloud(cloud_ptr);
  ne.setRadiusSearch(radius);
  ne.setKSearch(0);

Normal estimation is performed once more.

  // save normal estimation output to device shared memory point cloud
  pcl::oneapi::PointCloudDev<pcl::Normal>::Ptr normals_radius(new pcl::oneapi::PointCloudDev<pcl::Normal>) ;
  ne.compute(*normals_radius);