用C++和PCL库实现遍历所有叶子节点，并且使所有叶子内点的数量大于100

由于PCL库中没有树结构的类，下面示范使用Octree对点云数据进行八叉树分割，并遍历所有叶子节点，统计叶子内点的数量。

#include <iostream>
#include <pcl/point_cloud.h>
#include <pcl/octree/octree.h>

int main(int argc, char** argv)
{
    // 构造点云
    pcl::PointCloud<pcl::PointXYZ>::Ptr cloud(new pcl::PointCloud<pcl::PointXYZ>);
    cloud->width = 10000;
    cloud->height = 1;
    cloud->points.resize(cloud->width * cloud->height);

    for (std::size_t i = 0; i < cloud->size(); ++i)
    {
        (*cloud)[i].x = static_cast<float>(rand()) / RAND_MAX;
        (*cloud)[i].y = static_cast<float>(rand()) / RAND_MAX;
        (*cloud)[i].z = static_cast<float>(rand()) / RAND_MAX;
    }

    // 构造八叉树
    pcl::octree::OctreePointCloud<pcl::PointXYZ> octree(0.01f);
    octree.setInputCloud(cloud);
    octree.addPointsFromInputCloud();

    // 遍历所有叶子节点，统计叶子内点的数量
    std::vector<int> leaf_node_indices;
    octree.getLeafIndices(leaf_node_indices);

    int leaf_point_num_threshold = 100;
    int leaf_node_count = 0;
    for (const auto& index : leaf_node_indices)
    {
        std::vector<pcl::PointXYZ, Eigen::aligned_allocator<pcl::PointXYZ>> leaf_node_pointcloud;
        octree.getVoxelPoints(index, leaf_node_pointcloud);

        if (leaf_node_pointcloud.size() > leaf_point_num_threshold)
        {
            ++leaf_node_count;
        }
    }

    std::cout << "There are " << leaf_node_count << " leaves with more than " << leaf_point_num_threshold << " points." << std::endl;

    return 0;
}

其中，构造点云部分使用随机数生成点云，仅作为示例。构造八叉树部分使用了点云分辨率为0.01f作为参数。遍历所有叶子节点后，通过getVoxelPoints()获取叶子内的点云，并根据点云的数量判断是否满足条件。

运行该程序，即可得到输出：

There are 94 leaves with more than 100 points.