如何学习、上手点云算法(三)：用VsCode、Visual Studio来debug基于PCL、Open3D的代码

写在前面

• 本文内容
  以PCL 1.14.0，Open3D0.14.1为例，对基于PCL、Open3D开发的代码进行源码debug；
  如何学习、上手点云算法系列：
  如何学习、上手点云算法(一)：点云基础
  如何学习、上手点云算法(二)：点云处理相关开源算法库、软件、工具
  如何学习、上手点云算法(三)：用VsCode、Visual Studio来debug基于PCL、Open3D的代码
  更多点云基础、算法相关内容请关注专栏：
  点云处理基础
  点云配准(PointCloud Registration)
  Open3D点云处理
  PCL点云处理
  点云算法

• 平台/环境
  Windows10, CMake, Open3D, PCL

• 转载请注明出处：
  https://blog.csdn.net/qq_41102371/article/details/136504094

PCL

准备

安装PCL1.14.0: PCL1.14.0安装、使用教程
VsCode配置PCL: VsCode配置PCL、Open3D自动补全
在此基础上，下载debug需要的pdb文件：https://github.com/PointCloudLibrary/pcl/releases

打开压缩包

将pdb文件复制到之前装PCL的bin路径下，我这里是
D:\carlos\install\PCL 1.14.0\bin

注：PDB文件的作用见vs2019配置opencv+contrib-440 + PCL1.10.0 + 源码单步调试

编译debug版本

在VsCode配置PCL、Open3D自动补全和的基础上，新建一个compile_debug.bat:

cmake -DCMAKE_BUILD_TYPE=Debug ^
-DPCL_DIR="D:/carlos/install/PCL 1.14.0/cmake" ^
-S ./ -B ./build_debugcmake --build ./build_debug --config Debug --target ALL_BUILD

然后使用compile_debug.bat进行编译，完了会自动生成build_debug

配置launch.json

创建launch

添加配置

修改配置
"program"就是我们debug的程序对象，"environment"就是为当前程序添加PCL的环境变量，让其能找到PCL、VTK、FLANN等的动态库(.dll)
示例：

{// Use IntelliSense to learn about possible attributes.// Hover to view descriptions of existing attributes.// For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387"version": "0.2.0","configurations": [{"name": "(Windows) Launch","type": "cppvsdbg","request": "launch","program": "${workspaceFolder}/build_debug/Debug/statistical_removal.exe","args": [],"stopAtEntry": false,"cwd": "${fileDirname}","environment": [{"name": "PATH","value": "D:/carlos/install/PCL 1.14.0/bin;D:/carlos/install/PCL 1.14.0/3rdParty/FLANN/bin;D:/carlos/install/PCL 1.14.0/3rdParty/VTK/bin;D:/carlos/install/PCL 1.14.0/3rdParty/Qhull/bin;D:/carlos/install/PCL 1.14.0/3rdParty/OpenNI2/Tools;$(PATH)"}],"console": "externalTerminal"}]
}

在statistical_removal.cpp中随便加一个断点，然后使用F5或者右上角的Debug C/C++ File， 开始debug

然后就可以看到程序运行到了断点处，左侧有变量状态，自己添加监控，线程的显示

Visual Studio

打开文件夹，找到.sln文件双击打开

鼠标右键，将statistical_removal设置为启动项目

再点开最下面的属性，Debug–调试–环境–编辑，输入

PATH=;D:\carlos\install\PCL 1.14.0\bin;
D:\carlos\install\PCL 1.14.0\3rdParty\FLANN\bin;
D:\carlos\install\PCL 1.14.0\3rdParty\VTK\bin;
D:\carlos\install\PCL 1.14.0\3rdParty\Qhull\bin;
D:\carlos\install\PCL 1.14.0\3rdParty\OpenNI2\Tools;
$(PATH)

确定

Open3D

准备

Open3D没有像PCL提供Debug版的lib和对应的pdb文件，需要自己编译，在Open3D0.14.1编译、安装、demo使用教程中我们已经编译好了debug版本，现在需要去build_debug/lib/Debug中找到pdb文件
然后将pdb文件复制到install的目录中

添加open3d测试

在上面PCL示例project中，我们添加一个open3d的测试，在编译脚本中添加open3d的debug版的路径

compile_debug.bat:

cmake -DCMAKE_BUILD_TYPE=Debug ^
-DPCL_DIR="D:/carlos/install/PCL 1.14.0/cmake" ^
-DOpen3D_DIR="D:/carlos/install/open3d141_d/CMake" ^
-S ./ -B ./build_debugcmake --build ./build_debug --config Debug --target ALL_BUILD

添加一个测试代码test_open3d.cpp，该代码作用是平面拟合，并把拟合出的平面与剩下点云用不同颜色显示

#include <open3d/Open3D.h>int main()
{std::shared_ptr<open3d::geometry::PointCloud> pcd(new open3d::geometry::PointCloud);open3d::io::ReadPointCloud("./table_scene_lms400.pcd", *pcd);pcd->SegmentPlane();pcd->PaintUniformColor({1, 0, 0});open3d::visualization::DrawGeometries({pcd});return 0;
}

修改CMakeLists.txt

cmake_minimum_required(VERSION 3.18)project(statistical_removal)# PCL
find_package(PCL 1.2 REQUIRED)
include_directories(${PCL_INCLUDE_DIRS})
link_directories(${PCL_LIBRARY_DIRS})
add_definitions(${PCL_DEFINITIONS})
# PCL test
add_executable(statistical_removal statistical_removal.cpp)
target_link_libraries(statistical_removal ${PCL_LIBRARIES})# Open3D
option(STATIC_WINDOWS_RUNTIME "Use static (MT/MTd) Windows runtime" ON)
if(STATIC_WINDOWS_RUNTIME)set(CMAKE_MSVC_RUNTIME_LIBRARY "MultiThreaded$<$<CONFIG:Debug>:Debug>")
else()set(CMAKE_MSVC_RUNTIME_LIBRARY "MultiThreaded$<$<CONFIG:Debug>:Debug>DLL")
endif()
find_package(Open3D REQUIRED)
include_directories(${Open3D_INCLUDE_DIRS})# Open3D test
add_executable(test_open3d test_open3d.cpp)
target_link_libraries(test_open3d ${Open3D_LIBRARIES})

然后使用compile_debug.bat进行编译

调试

修改launch.json

{// Use IntelliSense to learn about possible attributes.// Hover to view descriptions of existing attributes.// For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387"version": "0.2.0","configurations": [{"name": "(Windows) Launch","type": "cppvsdbg","request": "launch",// "program": "${workspaceFolder}/build_debug/Debug/statistical_removal.exe","program": "${workspaceFolder}/build_debug/Debug/test_open3d.exe","args": [],"stopAtEntry": false,"cwd": "${fileDirname}","environment": [{"name": "PATH","value": "D:/carlos/install/PCL 1.14.0/bin;D:/carlos/install/PCL 1.14.0/3rdParty/FLANN/bin;D:/carlos/install/PCL 1.14.0/3rdParty/VTK/bin;D:/carlos/install/PCL 1.14.0/3rdParty/Qhull/bin;D:/carlos/install/PCL 1.14.0/3rdParty/OpenNI2/Tools;$(PATH)"}],"console": "externalTerminal"}]
}

在源码添加断点，开始调试

平面分割结果可视化

用Visual Studio debug参照上面PCL示例，打开.sln，设置启动项

但是Open3D是静态库，所以不用设置环境变量就可以了，另外用Visual Studio调试前把代码中点云读取路径改成绝对路径再重新编译一下，不然会找不到点云

参考

https://code.visualstudio.com/docs/cpp/launch-json-reference
其余文中已列出

完

主要做激光/影像三维重建，配准、分割等常用点云算法，熟悉open3d、pcl等开源点云库，技术交流、咨询可私信