Point cloud meshing] [Win10 + VS2015

Disclaimer: This article is a blogger original article, shall not be reproduced without the bloggers allowed. https://blog.csdn.net/YunLaowang/article/details/87714576

This article is a reference [computer vision] public life number of series: to learn from scratch with SLAM | point cloud to the evolution of the grid , grid point cloud implementation process.

Gridding process

1. + Downsampling statistical filtering
   sampling point cloud data by reducing the capacity, improve the processing speed; using statistical analysis techniques to remove noise point cloud data set, outliers;

Comparison before and after filtering

2. Resampling, the smoothing processing
   for smoothing the surface of the object and bug fixes by resampling

Before and after the point cloud smoothing

3. Computing surface normals point cloud
   computing cloud point normals, position and orientation point cloud and, color, normal information is merged together, constructed with the point cloud.

Point cloud surface normal

4. Gridding
   using projection greedy triangulation algorithm have to triangulate point clouds, achieve grid sparse point cloud.

Example grid point cloud

Code

/****************************
 * 给定稠密的点云，进行如下操作：
 * 		下采样和滤波、重采样平滑、法线计算，贪心投影网格化。
****************************/
#include <pcl/point_types.h>
#include <pcl/io/io.h>
#include <pcl/io/pcd_io.h>
#include <pcl/visualization/cloud_viewer.h>
#include <pcl/filters/radius_outlier_removal.h>
#include <pcl/filters/voxel_grid.h>
#include <pcl/filters/statistical_outlier_removal.h>
#include <pcl/surface/mls.h>
#include <pcl/features/normal_3d.h>
#include <pcl/surface/gp3.h>
#include <pcl/surface/poisson.h>
typedef pcl::PointXYZ PointT;

int main(int argc, char** argv)
{
	// Load input file
	pcl::PointCloud<PointT>::Ptr cloud(new pcl::PointCloud<PointT>);
    pcl::PointCloud<PointT>::Ptr cloud_downSampled(new pcl::PointCloud<PointT>);
	pcl::PointCloud<PointT>::Ptr cloud_filtered(new pcl::PointCloud<PointT>);
	pcl::PointCloud<PointT>::Ptr cloud_smoothed(new pcl::PointCloud<PointT>);
	if (pcl::io::loadPCDFile("fusedCloud.pcd", *cloud) == -1)
    {
        cout << "could not load the file..." << endl;
    }

    std::cout << "Orginal points number: " << cloud->points.size() << std::endl;

	// 1.下采样，同时保持点云形状特征
	pcl::VoxelGrid<PointT> downSampled;				// 下采样对象
	downSampled.setInputCloud(cloud);
	downSampled.setLeafSize(0.01f, 0.01f, 0.01f);	// 栅格叶的尺寸
	downSampled.filter(*cloud_downSampled);

	// 2.统计滤波
	pcl::StatisticalOutlierRemoval<PointT> sor;		// 滤波对象
	sor.setInputCloud(cloud_downSampled);
	sor.setMeanK(50);
	sor.setStddevMulThresh(1.0);					// 设置判定为离群点的阈值
	sor.filter(*cloud_filtered);

	// 3.对点云重采样,进行平滑
	pcl::search::KdTree<PointT>::Ptr treeSampling(new pcl::search::KdTree<PointT>); // 创建用于最近邻搜索的KD-Tree
	pcl::MovingLeastSquares<PointT, PointT> mls;	// 定义最小二乘实现的对象mls
	mls.setComputeNormals(false);					// 设置在最小二乘计算中是否需要存储计算的法线
	mls.setInputCloud(cloud_filtered);				// 设置待处理点云
	mls.setPolynomialOrder(2);						// 拟合2阶多项式拟合
	mls.setPolynomialFit(false);					// 设置为false可以 加速 smooth
	mls.setSearchMethod(treeSampling);				// 设置KD-Tree作为搜索方法
	mls.setSearchRadius(0.05);						// 单位m.设置用于拟合的K近邻半径
	mls.process(*cloud_smoothed);					// 输出

	// 4.法线估计
	pcl::NormalEstimation<PointT, pcl::Normal> normalEstimation;                    // 创建法线估计的对象
	normalEstimation.setInputCloud(cloud_smoothed);                                 // 输入点云
	pcl::search::KdTree<PointT>::Ptr tree(new pcl::search::KdTree<PointT>);         // 创建用于最近邻搜索的KD-Tree
	normalEstimation.setSearchMethod(tree);
	pcl::PointCloud<pcl::Normal>::Ptr normals(new pcl::PointCloud<pcl::Normal>);    // 定义输出的点云法线

	// K近邻确定方法，使用k个最近点，或者确定一个以r为半径的圆内的点集来确定都可以，两者选1即可
	normalEstimation.setKSearch(10);                    // 使用当前点周围最近的10个点														
	//normalEstimation.setRadiusSearch(0.03);           // 对于每一个点都用半径为3cm的近邻搜索方式

	normalEstimation.compute(*normals); 				// 计算法线

	// 5.将点云位姿、颜色、法线信息连接到一起
	pcl::PointCloud<pcl::PointNormal>::Ptr cloud_with_normals(new pcl::PointCloud<pcl::PointNormal>);
	pcl::concatenateFields(*cloud_smoothed, *normals, *cloud_with_normals);

	// 6.贪心投影三角化

	//定义搜索树对象
	pcl::search::KdTree<pcl::PointNormal>::Ptr tree2(new pcl::search::KdTree<pcl::PointNormal>);
	tree2->setInputCloud(cloud_with_normals);

	// 三角化
	pcl::GreedyProjectionTriangulation<pcl::PointNormal> gp3;   // 定义三角化对象
	pcl::PolygonMesh triangles;									// 存储最终三角化的网络模型

	// 设置三角化参数
	gp3.setSearchRadius(0.1);				// 设置搜索时的半径，也就是KNN的球半径
	gp3.setMu(2.5);							// 设置样本点搜索其近邻点的最远距离为2.5倍（典型值2.5-3），这样使得算法自适应点云密度的变化
	gp3.setMaximumNearestNeighbors(100);    // 设置样本点最多可搜索的邻域个数，典型值是50-100

	gp3.setMinimumAngle(M_PI / 18);			// 设置三角化后得到的三角形内角的最小的角度为10°
	gp3.setMaximumAngle(2 * M_PI / 3);		// 设置三角化后得到的三角形内角的最大角度为120°

	gp3.setMaximumSurfaceAngle(M_PI / 4);	// 设置某点法线方向偏离样本点法线的最大角度45°，如果超过，连接时不考虑该点
	gp3.setNormalConsistency(false);		// 设置该参数为true保证法线朝向一致，设置为false的话不会进行法线一致性检查

	gp3.setInputCloud(cloud_with_normals);  // 设置输入点云为有向点云
	gp3.setSearchMethod(tree2);				// 设置搜索方式
	gp3.reconstruct(triangles);				// 重建提取三角化
	
    // 7.显示网格化结果
    boost::shared_ptr<pcl::visualization::PCLVisualizer> viewer(new pcl::visualization::PCLVisualizer("3D Viewer"));
    viewer->setBackgroundColor(0, 0, 0);  		// 设置背景 
	viewer->addPolygonMesh(triangles, "mesh");  // 网格化点云添加到视窗
    while (!viewer->wasStopped())
    {
    	viewer->spinOnce(100);
    	boost::this_thread::sleep(boost::posix_time::microseconds(100000));
    }
    return 1;
}

result

• Note:
  the desired point cloud number provided the source file [public]: h? ttps: // pan? .baidu.com / s / 1avSGdi4IG3ry3wNCI_jDLQ mention? take? Code: cxjy (Delete to access the "?")