建立空间索引在点云数据处理中已被广泛应用,常见空间索引一般是自顶向下逐级划分空间的各种空间索引结构,比较有代表性的包括 BSP 树、KD 树、KDB 树、 R树、R+树、CELL 树、四叉树和八叉树等索引结构,而在这些结构中 KD 树和八叉树在 3D点云数据排列中应用较为广泛。 PCL 对八叉树的数据结构建立和索引方法进行了实现,以方便在此基础上对点云进行处理操作 。
八叉树(Octree)的定义是:若不为空树的话,树中任一节点的子节点恰好只会有八个,或零个,也就是子节点不会有0与8以外的数目。那么,这要用来做什么?想象一个立方体,我们最少可以切成多少个相同等分的小立方体?答案就是8个。再想象我们有一个房间,房间里某个角落藏着一枚金币,我们想很快的把金币找出来,聪明的你会怎么做?我们可以把房间当成一个立方体,先切成八个小立方体,然后排除掉没有放任何东西的小立方体,再把有可能藏金币的小立方体继续切八等份….如此下去,平均在(n表示房间内的所有物体数)的时间内就可找到金币。因此,八叉树就是用在3D空间中的场景管理,可以很快地知道物体在3D场景中的位置,或侦测与其它物体是否有碰撞以及是否在可视范围内。
这里引入了一个概念:
Voxel
翻译为体积元素,简称体素。描述了一个预设的最小单位的正方体
pcl的octree库提供了从点云数据创建具有层次的数据结构的方法。这样就可以对点数据集进行空间分区,下采样和搜索操作。每个八叉树节点有八个子节点或没有子节点。根节点描述了一个包围所有点的3维包容盒子。
pcl_octree实现提供了有效的最近邻居搜索(邻域搜索)API,例如“ 体素(Voxel)邻居搜索”,“ K最近邻居搜索”和“半径搜索邻居”。叶子节点类也提供其他功能,例如空间“占用率”和“每个体素(Voxel)的点密度”检查;序列化和反序列化功能可将八叉树结构有效地编码为二进制格式;此外,内存池实现减少了昂贵的内存分配和释放操作,以便快速创建八叉树。
下图说明了最低树级别的八叉树节点的体素边界框。八叉树体素围绕着兔子表面的每个3D点。红点代表点数据。该图像是使用octree_viewer创建的(visualization/tools/octree_viewer
)
方式二:K 近邻搜索
本例中K被设置成10, K 近邻搜索(K nearest neighbor search)方法把搜索结果写到两个分开的向量中, 第一个pointIdxNKNSearch 包含搜索结果〈结果点的索引的向量〉, 第二个pointNKNSquaredDistance 保存相应的搜索点和近邻之间的距离平方。
#include <iostream>
#include <vector>
#include <ctime>
#include <pcl/point_cloud.h>
#include <pcl/octree/octree_search.h>
#include <pcl/visualization/cloud_viewer.h>int main(int argc, char** argv) {srand((unsigned int)time(NULL));pcl::PointCloud<pcl::PointXYZ>::Ptr cloud(new pcl::PointCloud<pcl::PointXYZ>);// Generate pointcloud datacloud->width = 1000;cloud->height = 1;cloud->points.resize(cloud->width * cloud->height);for (size_t i = 0; i < cloud->points.size(); ++i) {cloud->points[i].x = 1024.0f * rand() / (RAND_MAX + 1.0f);cloud->points[i].y = 1024.0f * rand() / (RAND_MAX + 1.0f);cloud->points[i].z = 1024.0f * rand() / (RAND_MAX + 1.0f);}float resolution = 128.0f; // 设置分辨率为128// resolution该参数描述了octree叶子leaf节点的最小体素尺寸。pcl::octree::OctreePointCloudSearch<pcl::PointXYZ> octree(resolution);octree.setInputCloud(cloud); // 设置输入点云octree.addPointsFromInputCloud(); // 通过点云构建octreepcl::PointCloud<pcl::PointXYZ>::Ptr searchCloud(new pcl::PointCloud<pcl::PointXYZ>);pcl::PointXYZ searchPoint;searchPoint.x = 1024.0f * rand() / (RAND_MAX + 1.0f);searchPoint.y = 1024.0f * rand() / (RAND_MAX + 1.0f);searchPoint.z = 1024.0f * rand() / (RAND_MAX + 1.0f);searchCloud->push_back(searchPoint);int K = 10;std::vector<int> pointIdxNKNSearch;std::vector<float> pointNKNSquaredDistance;std::cout << "K nearest neighbor search at (" << searchPoint.x<< " " << searchPoint.y<< " " << searchPoint.z<< ") with K=" << K << std::endl;if (octree.nearestKSearch(searchPoint, K, pointIdxNKNSearch, pointNKNSquaredDistance) > 0) {for (size_t i = 0; i < pointIdxNKNSearch.size(); ++i)std::cout << " " << cloud->points[pointIdxNKNSearch[i]].x<< " " << cloud->points[pointIdxNKNSearch[i]].y<< " " << cloud->points[pointIdxNKNSearch[i]].z<< " (squared distance: " << pointNKNSquaredDistance[i] << ")" << std::endl;}pcl::visualization::PointCloudColorHandlerCustom<pcl::PointXYZ> originColorHandler(cloud, 255, 255, 255);pcl::visualization::PointCloudColorHandlerCustom<pcl::PointXYZ> searchColorHandler(searchCloud, 0, 255, 0);pcl::visualization::PCLVisualizer viewer("PCL Viewer");viewer.setBackgroundColor(0.1176, 0.1176, 0.2353);viewer.addPointCloud<pcl::PointXYZ>(cloud, originColorHandler, "cloud");viewer.addPointCloud<pcl::PointXYZ>(searchCloud, searchColorHandler, "search_cloud");viewer.setPointCloudRenderingProperties(pcl::visualization::PCL_VISUALIZER_POINT_SIZE, 5, "search_cloud");pcl::PointXYZ originPoint(0.0, 0.0, 0.0);viewer.addLine(originPoint, searchPoint, "line"); // 添加从原点到搜索点的线段viewer.setShapeRenderingProperties(pcl::visualization::PCL_VISUALIZER_COLOR, 1, 1, 0.2784, "line");viewer.addCoordinateSystem(200); // 指定坐标轴的长度,单位mmwhile (!viewer.wasStopped()) {viewer.spinOnce();}return 0;
}
octree.nearestKSearch(searchPoint, K, pointIdxNKNSearch, pointNKNSquaredDistance)
pcl::octree::OctreePointCloudSearch<pcl::PointXYZ>::nearestKSearch()
方法返回一个整数值,表示搜索操作是否成功并且找到了足够的最近邻点。如果搜索成功并且找到了至少K个最近邻点,则返回实际找到的最近邻点数量;否则,返回0。
该方法将给定的搜索点作为参数,并在八叉树中搜索最近的K个邻近点。找到的最近邻点的索引将存储在pointIdxNKNSearch
向量中,而对应的欧氏距离平方将存储在pointNKNSquaredDistance
向量中。
方式一:“体素近邻搜索”
体素近邻搜索(Neighbors within voxel search):它把查询点所在的体素中其他点的索引作为查询结果返回,结果以点索引向量的形式保存,因此搜索点和搜索结果之间的距离取决于八叉树的分辨率参数。
#include <iostream>
#include <vector>
#include <ctime>
#include <pcl/point_cloud.h>
#include <pcl/octree/octree_search.h>
#include <pcl/visualization/cloud_viewer.h>int main(int argc, char** argv) {srand((unsigned int)time(NULL));pcl::PointCloud<pcl::PointXYZ>::Ptr cloud(new pcl::PointCloud<pcl::PointXYZ>);// Generate pointcloud datacloud->width = 1000;cloud->height = 1;cloud->points.resize(cloud->width * cloud->height);for (size_t i = 0; i < cloud->points.size(); ++i) {cloud->points[i].x = 1024.0f * rand() / (RAND_MAX + 1.0f);cloud->points[i].y = 1024.0f * rand() / (RAND_MAX + 1.0f);cloud->points[i].z = 1024.0f * rand() / (RAND_MAX + 1.0f);}float resolution = 128.0f; // 设置分辨率为128// resolution该参数描述了octree叶子leaf节点的最小体素尺寸。pcl::octree::OctreePointCloudSearch<pcl::PointXYZ> octree(resolution);octree.setInputCloud(cloud); // 设置输入点云octree.addPointsFromInputCloud(); // 通过点云构建octreepcl::PointCloud<pcl::PointXYZ>::Ptr searchCloud(new pcl::PointCloud<pcl::PointXYZ>);pcl::PointXYZ searchPoint;searchPoint.x = 1024.0f * rand() / (RAND_MAX + 1.0f);searchPoint.y = 1024.0f * rand() / (RAND_MAX + 1.0f);searchPoint.z = 1024.0f * rand() / (RAND_MAX + 1.0f);searchCloud->push_back(searchPoint);std::vector<int> pointIdxVec;if (octree.voxelSearch(searchPoint, pointIdxVec)) {std::cout << "Neighbors within voxel search at (" << searchPoint.x<< " " << searchPoint.y<< " " << searchPoint.z << ")"<< std::endl;for (size_t i = 0; i < pointIdxVec.size(); ++i)std::cout << " " << cloud->points[pointIdxVec[i]].x<< " " << cloud->points[pointIdxVec[i]].y<< " " << cloud->points[pointIdxVec[i]].z << std::endl;}pcl::visualization::PointCloudColorHandlerCustom<pcl::PointXYZ> originColorHandler(cloud, 255, 255, 255);pcl::visualization::PointCloudColorHandlerCustom<pcl::PointXYZ> searchColorHandler(searchCloud, 0, 255, 0);pcl::visualization::PCLVisualizer viewer("PCL Viewer");viewer.setBackgroundColor(0.1176, 0.1176, 0.2353);viewer.addPointCloud<pcl::PointXYZ>(cloud, originColorHandler, "cloud");viewer.addPointCloud<pcl::PointXYZ>(searchCloud, searchColorHandler, "search_cloud");viewer.setPointCloudRenderingProperties(pcl::visualization::PCL_VISUALIZER_POINT_SIZE, 5, "search_cloud");pcl::PointXYZ originPoint(0.0, 0.0, 0.0);viewer.addLine(originPoint, searchPoint, "line"); // 添加从原点到搜索点的线段viewer.setShapeRenderingProperties(pcl::visualization::PCL_VISUALIZER_COLOR, 1, 1, 0.2784, "line");viewer.addCoordinateSystem(200); // 指定坐标轴的长度,单位mmwhile (!viewer.wasStopped()) {viewer.spinOnce();}return 0;
}
octree.voxelSearch(searchPoint, pointIdxVec)
pcl::octree::OctreePointCloudSearch<pcl::PointXYZ>::voxelSearch()
方法返回一个bool
类型的值,表示搜索操作是否成功。如果搜索成功并且找到了匹配的体素,则返回true
;否则,返回false
。
该方法将给定的搜索点作为参数,并在八叉树中搜索包含该点的体素。如果找到匹配的体素,则会将体素内的点索引存储在pointIdxVec
向量中,并返回true
。如果没有找到匹配的体素,则pointIdxVec
向量将为空,并返回false
。
方式三:半径内近邻搜索
半径内近邻搜索(Neighbors within radius search)原理和“K 近邻搜索”类似,它的搜索结果被写入两个分开的向量中, 这两个向量分别存储结果点的索引和对应的距离平方
#include <iostream>
#include <vector>
#include <ctime>
#include <pcl/point_cloud.h>
#include <pcl/octree/octree_search.h>
#include <pcl/visualization/cloud_viewer.h>int main(int argc, char** argv) {srand((unsigned int)time(NULL));pcl::PointCloud<pcl::PointXYZ>::Ptr cloud(new pcl::PointCloud<pcl::PointXYZ>);// Generate pointcloud datacloud->width = 1000;cloud->height = 1;cloud->points.resize(cloud->width * cloud->height);for (size_t i = 0; i < cloud->points.size(); ++i) {cloud->points[i].x = 1024.0f * rand() / (RAND_MAX + 1.0f);cloud->points[i].y = 1024.0f * rand() / (RAND_MAX + 1.0f);cloud->points[i].z = 1024.0f * rand() / (RAND_MAX + 1.0f);}float resolution = 128.0f; // 设置分辨率为128// resolution该参数描述了octree叶子leaf节点的最小体素尺寸。pcl::octree::OctreePointCloudSearch<pcl::PointXYZ> octree(resolution);octree.setInputCloud(cloud); // 设置输入点云octree.addPointsFromInputCloud(); // 通过点云构建octreepcl::PointCloud<pcl::PointXYZ>::Ptr searchCloud(new pcl::PointCloud<pcl::PointXYZ>);pcl::PointXYZ searchPoint;searchPoint.x = 1024.0f * rand() / (RAND_MAX + 1.0f);searchPoint.y = 1024.0f * rand() / (RAND_MAX + 1.0f);searchPoint.z = 1024.0f * rand() / (RAND_MAX + 1.0f);searchCloud->push_back(searchPoint);std::vector<int> pointIdxRadiusSearch;std::vector<float> pointRadiusSquaredDistance;float radius = 256.0f * rand() / (RAND_MAX + 1.0f);std::cout << "Neighbors within radius search at (" << searchPoint.x<< " " << searchPoint.y<< " " << searchPoint.z<< ") with radius=" << radius << std::endl;if (octree.radiusSearch(searchPoint, radius, pointIdxRadiusSearch, pointRadiusSquaredDistance) > 0) {for (size_t i = 0; i < pointIdxRadiusSearch.size(); ++i)std::cout << " " << cloud->points[pointIdxRadiusSearch[i]].x<< " " << cloud->points[pointIdxRadiusSearch[i]].y<< " " << cloud->points[pointIdxRadiusSearch[i]].z<< " (squared distance: " << pointRadiusSquaredDistance[i] << ")" << std::endl;}pcl::visualization::PointCloudColorHandlerCustom<pcl::PointXYZ> originColorHandler(cloud, 255, 255, 255);pcl::visualization::PointCloudColorHandlerCustom<pcl::PointXYZ> searchColorHandler(searchCloud, 0, 255, 0);pcl::visualization::PCLVisualizer viewer("PCL Viewer");viewer.setBackgroundColor(0.1176, 0.1176, 0.2353);viewer.addPointCloud<pcl::PointXYZ>(cloud, originColorHandler, "cloud");viewer.addPointCloud<pcl::PointXYZ>(searchCloud, searchColorHandler, "search_cloud");viewer.setPointCloudRenderingProperties(pcl::visualization::PCL_VISUALIZER_POINT_SIZE, 5, "search_cloud");pcl::PointXYZ originPoint(0.0, 0.0, 0.0);viewer.addLine(originPoint, searchPoint, "line"); // 添加从原点到搜索点的线段viewer.setShapeRenderingProperties(pcl::visualization::PCL_VISUALIZER_COLOR, 1, 1, 0.2784, "line");viewer.addCoordinateSystem(200); // 指定坐标轴的长度,单位mmwhile (!viewer.wasStopped()) {viewer.spinOnce();}return 0;
}
octree.radiusSearch(searchPoint, radius, pointIdxRadiusSearch, pointRadiusSquaredDistance)
pcl::octree::OctreePointCloudSearch<pcl::PointXYZ>::radiusSearch()
方法是用于在八叉树中执行半径搜索的函数。它根据给定的搜索点和半径,在八叉树中找到落在指定半径范围内的所有点。
以下是该方法的参数说明:
searchPoint
:要进行半径搜索的点,类型为pcl::PointXYZ
。即你希望以这个点为中心进行半径搜索。radius
:指定的搜索半径,类型为double
。它定义了从搜索点出发的球体的半径,只有落在该球体内的点才会被搜索到。pointIdxRadiusSearch
:存储在搜索半径范围内的点的索引的向量,类型为std::vector<int>
。pointRadiusSquaredDistance
:存储在搜索半径范围内的点与搜索点之间的欧氏距离平方的向量,类型为std::vector<float>
。
使用radiusSearch()
方法时,传递一个搜索点和半径作为输入参数,并提供两个向量作为输出参数。方法将填充pointIdxRadiusSearch
向量以保存搜索半径范围内的点的索引,同时填充pointRadiusSquaredDistance
向量以保存每个点与搜索点之间的欧氏距离平方。