pcl基于最小切割的点云分割

在本教程中，我们将学习如何使用pcl::MinCutSegmentation类中实现的基于最小切的分割算法。该算法对给定的输入云进行二进制分割。根据对象中心及其半径，该算法将点云划分为两组：前景点和背景点（属于对象的点和不属于对象的点）。

源码：

 创建 min_cut_segmentation.cpp 文件

 1#include <iostream>
 2#include <vector>
 3#include <pcl/io/pcd_io.h>
 4#include <pcl/point_types.h>
 5#include <pcl/visualization/cloud_viewer.h>
 6#include <pcl/filters/filter_indices.h> // for pcl::removeNaNFromPointCloud
 7#include <pcl/segmentation/min_cut_segmentation.h>
 8
 9int main ()
10{
11  pcl::PointCloud <pcl::PointXYZ>::Ptr cloud (new pcl::PointCloud <pcl::PointXYZ>);
12  if ( pcl::io::loadPCDFile <pcl::PointXYZ> ("min_cut_segmentation_tutorial.pcd", *cloud) == -1 )
13  {
14    std::cout << "Cloud reading failed." << std::endl;
15    return (-1);
16  }
17
18  pcl::IndicesPtr indices (new std::vector <int>);
19  pcl::removeNaNFromPointCloud(*cloud, *indices);
20
21  pcl::MinCutSegmentation<pcl::PointXYZ> seg;
22  seg.setInputCloud (cloud);
23  seg.setIndices (indices);
24
25  pcl::PointCloud<pcl::PointXYZ>::Ptr foreground_points(new pcl::PointCloud<pcl::PointXYZ> ());
26  pcl::PointXYZ point;
27  point.x = 68.97;
28  point.y = -18.55;
29  point.z = 0.57;
30  foreground_points->points.push_back(point);
31  seg.setForegroundPoints (foreground_points);
32
33  seg.setSigma (0.25);
34  seg.setRadius (3.0433856);
35  seg.setNumberOfNeighbours (14);
36  seg.setSourceWeight (0.8);
37
38  std::vector <pcl::PointIndices> clusters;
39  seg.extract (clusters);
40
41  std::cout << "Maximum flow is " << seg.getMaxFlow () << std::endl;
42
43  pcl::PointCloud <pcl::PointXYZRGB>::Ptr colored_cloud = seg.getColoredCloud ();
44  pcl::visualization::CloudViewer viewer ("Cluster viewer");
45  viewer.showCloud(colored_cloud);
46  while (!viewer.wasStopped ())
47  {
48  }
49
50  return (0);
51}

说明：

1、相关头文件

 1#include <iostream>
 2#include <vector>
 3#include <pcl/io/pcd_io.h>
 4#include <pcl/point_types.h>
 5#include <pcl/visualization/cloud_viewer.h>
 6#include <pcl/filters/filter_indices.h> // for pcl::removeNaNFromPointCloud
 7#include <pcl/segmentation/min_cut_segmentation.h>

2、从 .pcd 文件加载点云。

  pcl::PointCloud <pcl::PointXYZ>::Ptr cloud (new pcl::PointCloud <pcl::PointXYZ>);
  if ( pcl::io::loadPCDFile <pcl::PointXYZ> ("min_cut_segmentation_tutorial.pcd", *cloud) == -1 )
  {
    std::cout << "Cloud reading failed." << std::endl;
    return (-1);
  }

3、表明pcl::MinCutSegmentation类可以使用索引。这里，只选择有效点进行分割。

  pcl::IndicesPtr indices (new std::vector <int>);
  pcl::removeNaNFromPointCloud(*cloud, *indices);

4、pcl::MinCutSegmentation类实例化。模板类只有一个参数 - PointT - 说明将使用哪种类型的点。

  pcl::MinCutSegmentation<pcl::PointXYZ> seg;

5、提供了必须分割的点云和索引。

  seg.setInputCloud (cloud);
  seg.setIndices (indices);

6、算法需要已知为对象中心的点。这些行提供了它。

  pcl::PointCloud<pcl::PointXYZ>::Ptr foreground_points(new pcl::PointCloud<pcl::PointXYZ> ());
  pcl::PointXYZ point;
  point.x = 68.97;
  point.y = -18.55;
  point.z = 0.57;
  foreground_points->points.push_back(point);
  seg.setForegroundPoints (foreground_points);

7、设置平滑成本计算所需的和对象半径。

  seg.setSigma (0.25);
  seg.setRadius (3.0433856);

8、表示在构造图时要找到多少邻居。设置的邻居越多，它包含的边数就越多。

  seg.setNumberOfNeighbours (14);

9、设置前景惩罚

  seg.setSourceWeight (0.8);

10、启动算法，分割后的簇将包含结果

  std::vector <pcl::PointIndices> clusters;
  seg.extract (clusters);

11、访问在图形切割期间计算的流量值

  std::cout << "Maximum flow is " << seg.getMaxFlow () << std::endl;

12、创建 CloudViewer 类的实例以实现结果可视化。

  pcl::PointCloud <pcl::PointXYZRGB>::Ptr colored_cloud = seg.getColoredCloud ();
  pcl::visualization::CloudViewer viewer ("Cluster viewer");
  viewer.showCloud(colored_cloud);
  while (!viewer.wasStopped ())
  {
  }

编译和运行

1、在 CMakeLists.txt 文件中添加以下代码行:

 1cmake_minimum_required(VERSION 3.5 FATAL_ERROR)
 2
 3project(min_cut_segmentation)
 4
 5find_package(PCL 1.5 REQUIRED)
 6
 7include_directories(${PCL_INCLUDE_DIRS})
 8link_directories(${PCL_LIBRARY_DIRS})
 9add_definitions(${PCL_DEFINITIONS})
10
11add_executable (min_cut_segmentation min_cut_segmentation.cpp)
12target_link_libraries (min_cut_segmentation ${PCL_LIBRARIES})

2、运行

$ ./min_cut_segmentation

3、输出