las数据转pcd并显示

不念不忘少年蓝@ 2022-08-30 04:12 275阅读 0赞

成功的方法：

可以转成ply bin pcd格式，转换流程如下所示：  
1、打开CloudCompare软件，加载点云数据(.las或者是.laz等等)。  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2RheXVoYWl0YW5nMQ_size_16_color_FFFFFF_t_70]  
2、选择要进行转换的文件，此时屏幕上会出现一个立方体。  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2RheXVoYWl0YW5nMQ_size_16_color_FFFFFF_t_70 1]  
3、点击【File】中的【Save】，选择你要保存的格式即可。  
![在这里插入图片描述][20200110095235679.png]

laspy的方法没成功，c++的方法还没试。

pip install laspy==1.5.1

不能安装2.0.0以后版本，2.0.1会报错。

[https://github.com/grantbrown/laspy/releases][https_github.com_grantbrown_laspy_releases]

from laspy.file import File
    import numpy as np
    
    def las2pcd():
        inFile = File(r'F:\biadu_down\las-20200813-110417_1\data-20200813-110417_1.las', mode='r')
    
        I = inFile.Classification == 2
    
        outFile = File('aaaa.pcd', mode='w', header=inFile.header)
        outFile.points = inFile.points
        # outFile.points = inFile.points[I]
        outFile.close()
    
    if __name__ == '__main__':
        import numpy as np
        las2pcd()
    
        import open3d as o3d
    
        pcd1 = o3d.io.read_point_cloud(r'aaaa.pcd')
        # pcd1 = o3d.io.read_point_cloud(r'F:\biadu_down\las-20200813-110417_1\data-20200813-110417_1.las')
        pcd1.paint_uniform_color([0, 0, 0])
    
        o3d.visualization.draw_geometries([pcd1])

转换之后，open3d直接读las，出来是白屏。

c++:

[https://github.com/yoshiyama/las2pcd\_gui/blob/master/las2pcd/las2pcd.cpp][https_github.com_yoshiyama_las2pcd_gui_blob_master_las2pcd_las2pcd.cpp]

#include <iostream>
    #include <cstdlib>
    #include <liblas/liblas.hpp>
    #include <pcl/io/io.h>
    #include <pcl/io/pcd_io.h>
    #include <pcl/point_types.h>
    
    using namespace std;
    
    int main (int argc, char *argv[])
    {
    	string filePath;
    	cout << "===================================================================" << endl;
    	cout << "LAS2PCD - Converts .las point clouds into PCL-friendly format .pcd" << endl;
    	cout << "ver 0.3 - 16 March 2018" << endl;
    	cout << "(c) Arnadi Murtiyoso" << endl;
    	cout << "PAGE Group, ICube Laboratory UMR 7357 INSA Strasbourg" << endl;
    	cout << "contact: arnadi.murtiyoso@insa-strasbourg.fr" << endl;
    	cout << "https://github.com/murtiad" << endl;
    	cout << "Ubuntu tweaks by Jonathan Greenberg, jgreenberg@unr.edu" << endl;
    	cout << "Ubuntu version: https://github.com/gearslaboratory/las2pcd" << endl;
    	cout << "===================================================================" << endl;
    	cout << endl;
    
    	// cout << "Enter full .las file path: (or you can also drag the file here)" << endl;
        
    	// getline(cin, filePath);
    	// Edited to make fully command-line:
    	filePath=argv[1];
    
        std::cerr << "INFO : Loading : " << filePath << std::endl;
        
        // instancing a new PCL pointcloud object
        pcl::PointCloud<pcl::PointXYZRGB> cloud;
    
        // Opening  the las file
        std::ifstream ifs(filePath.c_str(), std::ios::in | std::ios::binary);
    
    	// Safeguard against opening failure
        if(ifs.fail()) 
    	{
            std::cerr << "ERROR : Impossible to open the file : " << filePath <<std::endl;
            return 1;
        }
    
        liblas::ReaderFactory f;
        liblas::Reader reader = f.CreateWithStream(ifs); // reading las file
        unsigned long int nbPoints=reader.GetHeader().GetPointRecordsCount();
    
    	// Fill in the cloud data
    	cloud.width    = nbPoints;				// This means that the point cloud is "unorganized"
    	cloud.height   = 1;						// (i.e. not a depth map)
    	cloud.is_dense = false;
    	cloud.points.resize (cloud.width * cloud.height);
    
    	cout << "INFO : " << cloud.points.size () << " points detected in " << filePath << endl;
    
    	int i=0;				// counter
    	uint16_t r1, g1, b1;	// RGB variables for .las (16-bit coded)
    	int r2, g2, b2;			// RGB variables for converted values (see below)
    	uint32_t rgb;			// "packed" RGB value for .pcd
    
    	while(reader.ReadNextPoint()) 
    	{
    		// get XYZ information
    		cloud.points[i].x = (reader.GetPoint().GetX());
    	    	cloud.points[i].y = (reader.GetPoint().GetY());
    	    	cloud.points[i].z = (reader.GetPoint().GetZ());
    		
    		// get intensity information
    		//cloud.points[i].intensity = (reader.GetPoint().GetIntensity());
    				
    		// get RGB information. Note: in liblas, the "Color" class can be accessed from within the "Point" class, thus the triple gets
    		r1 = (reader.GetPoint().GetColor().GetRed());
    		g1 = (reader.GetPoint().GetColor().GetGreen());
    		b1 = (reader.GetPoint().GetColor().GetBlue()); 
    
    		// .las stores RGB color in 16-bit (0-65535) while .pcd demands an 8-bit value (0-255). Let's convert them!
    		r2 = ceil(((float)r1/65535)*(float)255);
    		g2 = ceil(((float)g1/65535)*(float)255);
    		b2 = ceil(((float)b1/65535)*(float)255);
    
    		// PCL particularity: must "pack" the RGB into one single integer and then reinterpret them as float
    		rgb = ((int)r2) << 16 | ((int)g2) << 8 | ((int)b2);
    
    		cloud.points[i].rgb = *reinterpret_cast<float*>(&rgb);
    					
    		i++; // ...moving on
    	}
      
    	// Allows output file to be set:
    	//pcl::io::savePCDFileASCII (argv[2], cloud);
    	pcl::io::savePCDFileBinaryCompressed (argv[2], cloud);
      
    	std::cerr << "Saved " << cloud.points.size () << " data points to pointcloud.pcd." << std::endl;
    
    	return (0);
    }

[https://www.freesion.com/article/2519238749/][https_www.freesion.com_article_2519238749]

[https://blog.csdn.net/qq\_42570058/article/details/83068759][https_blog.csdn.net_qq_42570058_article_details_83068759]

#include <liblas\liblas.hpp>
    #include <fstream>
    #include <iostream>
    #include <string>
    #include<vector>
    #include<pcl\io\pcd_io.h>
    #include<pcl\point_types.h>
    #include<time.h>
    int main()
    {
    	clock_t sT, tT,eT;
    	//打开文件
    	ifstream fp;
    	fp.open("1.las",ios::in|ios::binary);
    	sT = clock();
    	//读取
    	liblas::ReaderFactory readerFactory;
    	liblas::Reader reader = readerFactory.CreateWithStream(fp);
    	tT = clock();
    	cout << "读取" << reader.GetHeader().GetPointRecordsCount() <<"个点花费"<<(double)(tT-sT)/CLOCKS_PER_SEC<<"s"<<endl;
    	//点云类型
    	pcl::PointCloud<pcl::PointXYZ> cloudOutput;
    	cloudOutput.clear();
    	//转换xyz
    	while (reader.ReadNextPoint())
    	{
    		double x = reader.GetPoint().GetX();
    		double y = reader.GetPoint().GetY();
    		double z = reader.GetPoint().GetZ();
    		pcl::PointXYZ thePt(x,y,z);
    		cloudOutput.push_back(thePt);
    	}
    	cloudOutput.width = cloudOutput.size();
    	cloudOutput.height = 1;
    	cloudOutput.is_dense = false;
    	cloudOutput.resize(cloudOutput.width*cloudOutput.height);
    	pcl::io::savePCDFileASCII("1.pcd",cloudOutput);
    	cloudOutput.clear();
    	eT = clock();
    	cout << "存入" << reader.GetHeader().GetPointRecordsCount() << "个点花费" << (double)(eT - tT) / CLOCKS_PER_SEC << "s" << endl;
    	return 0;
    }

[https://www.dtmao.cc/news\_show\_15044.shtml][https_www.dtmao.cc_news_show_15044.shtml]

.las数据转.pcd格式是参考这里点击打开链接，很容易能够得到正确的.pcd格式文件；

但是，当用PCL库读取并显示的时候，却看不到图形，我是用的PCL的第一种显示方法CloudViewer显示的，也是用的官网示例代码

#include <pcl/visualization/cloud_viewer.h>
    #include <iostream>
    #include <pcl/io/io.h>
    #include <pcl/io/pcd_io.h>
        
    int user_data;
        
    void 
    viewerOneOff (pcl::visualization::PCLVisualizer& viewer)
    {
        viewer.setBackgroundColor (1.0, 0.5, 1.0);
        pcl::PointXYZ o;
        o.x = 1.0;
        o.y = 0;
        o.z = 0;
        viewer.addSphere (o, 0.25, "sphere", 0);
        std::cout << "i only run once" << std::endl;
        
    }
        
    void 
    viewerPsycho (pcl::visualization::PCLVisualizer& viewer)
    {
        static unsigned count = 0;
        std::stringstream ss;
        ss << "Once per viewer loop: " << count++;
        viewer.removeShape ("text", 0);
        viewer.addText (ss.str(), 200, 300, "text", 0);
        
        //FIXME: possible race condition here:
        user_data++;
    }
        
    int 
    main ()
    {
        pcl::PointCloud<pcl::PointXYZRGBA>::Ptr cloud (new pcl::PointCloud<pcl::PointXYZRGBA>);
        pcl::io::loadPCDFile ("my_point_cloud.pcd", *cloud);
        
        pcl::visualization::CloudViewer viewer("Cloud Viewer");
        
        //blocks until the cloud is actually rendered
        viewer.showCloud(cloud);
        
        //use the following functions to get access to the underlying more advanced/powerful
        //PCLVisualizer
        
        //This will only get called once
        viewer.runOnVisualizationThreadOnce (viewerOneOff);
        
        //This will get called once per visualization iteration
        viewer.runOnVisualizationThread (viewerPsycho);
        while (!viewer.wasStopped ())
        {
        //you can also do cool processing here
        //FIXME: Note that this is running in a separate thread from viewerPsycho
        //and you should guard against race conditions yourself...
        user_data++;
        }
        return 0;
    }
        于是查找原因，准备换一种显示方法，就是比较复杂的PCLVisualizer,但是代码有点多，也是第一次学习使用PCL,就没有立刻写代码，而是分析了一下源代码中的示例数据，bunny.pcd;这只兔子的数据都是小于1的，猜测应该是归一化之后得到的坐标，所以，网上找相关内容，但没有得到肯定的答案，因为也担心归一化的效率问题，想看看有没有更好的办法啊，最终，没有找到更好的办法来保留原始坐标系统，通过修改.las转.pcd格式的代码，也就是增加了一个归一化步骤，之后得到数据，居然能够正常显示了，也算是有点收获；
    
    // LASlib_3DlasTo2D.cpp: 定义控制台应用程序的入口点。
    //
     
    #include "stdafx.h"
    #include "lasreader.hpp"
    #include "laswriter.hpp"
    #include <pcl/visualization/cloud_viewer.h>
    #include <iostream>
    #include <pcl/io/io.h>
    #include <pcl/io/pcd_io.h>
    #include <GL/glut.h>  
    #include "stdafx.h"
    #include <pcl/io/pcd_io.h>
    #include <pcl/point_types.h>
    #include <string>
    #include <vector>
    #include <fstream>
    #include <ios>
    #include <iostream>
     
    //#include <glut.h>
    #include <vtkAutoInit.h>
    VTK_MODULE_INIT(vtkRenderingOpenGL);
    VTK_MODULE_INIT(vtkInteractionStyle);
     
    /*
    将'/'格式的文件路径转换成 '\\';
    */
    static std::string convertFilePath(const std::string& file)
    {
        int i = 0;
        std::string s(file);
        for (i = 0; i < s.size(); ++i)
        {
            if (s[i] == '/')
                s[i] = '\\';
        }
        return s;
    }
    int main()
    {
     
    /*
        LASreadOpener lasreadopener;
        lasreadopener.set_file_name("1.las");
        LASreader* lasreader = lasreadopener.open();
        LASwriteOpener laswriteopener;
        laswriteopener.set_file_name("2.laz");
        LASwriter* laswriter = laswriteopener.open(&lasreader->header);
        while (lasreader->read_point())
            laswriter->write_point(&lasreader->point);
        laswriter->close();
        delete laswriter;
        lasreader->close();
        delete lasreader;
    */
        
        const char* your_las_file_path = "D:\\任务：点云和影像配准\\data\\LIDAR\\4test_60km.las";
            const char* your_pcd_out_file_path = "XX.pcd";
            //laslib只允许'\\'格式的文件路径。
            std::string lasFile = convertFilePath(your_las_file_path);
     
     
            //打开las文件
            LASreadOpener lasreadopener;
            lasreadopener.set_file_name(lasFile.c_str());
            LASreader* lasreader = lasreadopener.open();
            size_t count = lasreader->header.number_of_point_records;
            pcl::PointCloud<pcl::PointXYZ>::Ptr pointCloudPtr(new pcl::PointCloud<pcl::PointXYZ>);
            pointCloudPtr->resize(count);
            pointCloudPtr->width = 1;
            pointCloudPtr->height = count;
            pointCloudPtr->is_dense = false;
            size_t i = 0;
            int maxX = -9999, minX = 9999999999, MaxY = -999999999, minY = 9999999999, maxZ = -999999999, minZ=999999999;
            /*int x, y, z;*/
            
            maxX = lasreader->get_max_x();
            minX = lasreader->get_min_x();
            MaxY = lasreader->get_max_y();
            minY = lasreader->get_min_y();
            maxZ = lasreader->get_max_z();
            minZ = lasreader->get_min_z();
     
            double dx = maxX - minX;
            double dy = MaxY - minY;
            double dz = maxZ - minZ;
            double maxd = dx > dy ? (dx > dz ? dx : dz) : (dy > dz ? dy : dz);
            /*double maxd,mind,Dd;
            dx > dy ? (dx > dz ? (maxd = maxX, mind = minX, Dd = dx) : (maxd = maxZ, mind = minZ, Dd = dz)) : (dy > dz ? (maxd = MaxY, mind = minY, Dd = dy) : (maxd = maxZ, mind = minZ, Dd = dz));
    */
     
            printf( "%f",maxd);
     
     
            size_t j = 0;
            
            while (lasreader->read_point() && j < count)
            {
                pointCloudPtr->points[j].x = (lasreader->point.get_x() - minX)/maxd;
                pointCloudPtr->points[j].y = (lasreader->point.get_y() - minY)/maxd;
                pointCloudPtr->points[j].z = (lasreader->point.get_z() - minZ)/maxd;
                /*pointCloudPtr->points[j].x = (lasreader->point.get_x()) / 1;
                pointCloudPtr->points[j].y = (lasreader->point.get_y()) / 1;
                pointCloudPtr->points[j].z = (lasreader->point.get_z()) / 1;*/
                ++j;
            }
            pcl::io::savePCDFileASCII(your_pcd_out_file_path, *pointCloudPtr);
        
     
            getchar();
        return 0;
    }

这是CloudViewer显示结果，没有添加渲染效果；

这是原始.las数据，ENVILiDar软件显示：  
————————————————  
版权声明：本文为CSDN博主「啊吼！」的原创文章，遵循CC 4.0 BY-SA版权协议，转载请附上原文出处链接及本声明。  
原文链接：https://blog.csdn.net/xujie126/article/details/80824924

[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2RheXVoYWl0YW5nMQ_size_16_color_FFFFFF_t_70]: /images/20220829/5762c1ee951542d2839ca1a578c5b2f8.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2RheXVoYWl0YW5nMQ_size_16_color_FFFFFF_t_70 1]: /images/20220829/aca278e736c1495bb89ca2939372573b.png
[20200110095235679.png]: /images/20220829/ea19879852984c13bff2b9d0f71cfbfd.png
[https_github.com_grantbrown_laspy_releases]: https://github.com/grantbrown/laspy/releases
[https_github.com_yoshiyama_las2pcd_gui_blob_master_las2pcd_las2pcd.cpp]: https://github.com/yoshiyama/las2pcd_gui/blob/master/las2pcd/las2pcd.cpp
[https_www.freesion.com_article_2519238749]: https://www.freesion.com/article/2519238749/
[https_blog.csdn.net_qq_42570058_article_details_83068759]: https://blog.csdn.net/qq_42570058/article/details/83068759
[https_www.dtmao.cc_news_show_15044.shtml]: https://www.dtmao.cc/news_show_15044.shtml