pcl_visualizer | Visualizer for 3D point cloud using PCL Library | Image Editing library

First build and install vtk:

Build

• tar xf VTK-9.0.1.tar.gz
• cd VTK-9.0.1
• mkdir build && cd build
• cmake ..
• make -j

Install

(if you have a different preferred place for installations than ~/.local, just change insdir below)

• insdir=$(echo ~/.local)
• mkdir $insdir
• cmake --install . --prefix $insdir

Then add the include path (-I $insdir/include/vtk-9.0), the library path (-L $insdir/lib64 or possibly $insdir/lib on a 32 bit machine) and vtk libraries (-llibname1 ... -llibnameX) to your compilation command.

Example:

• allvtklibs=$(ls $insdir/lib64/libvtk*.so | sed -E 's,^.*/lib(.*)\.so$,-l\1,')
• g++ -std=c++14 d2pc.cpp -o d2pc -I $insdir/include/vtk-9.0 -L $insdir/lib64 $(pkg-config --cflags --libs opencv pcl_io-1.11 pcl_visualization-1.11) -Wl,-rpath=$insdir/lib64 $allvtklibs -lboost_system

1. Looks like you're not linking the PCL visualization lib file and probably others. Do this in the project properties > Linker > Input > Additional Dependencies. The path to your PCL lib directory should be specified in Project properties > VC++ Directories > Library Directories. At the same time the path to your PCL include directory should be specified in Project properties > VC++ Directories > Include Directories.
2. PCL functions and types are templated on the point type. So you will use pcl::PointXYZ as the point type for all functions.
3. No. Yours will be more similar to the Simple Cloud Visualization example in the CloudViewer tutorial.
4. You don't copy the PCL .h and .cpp files into your solution. You #include the PCL headers that remain in their installed location, and you link in the PCL lib files, (see 1 above.)
5. None. If the PCL bin/ directory is not in your $PATH then place the PCL .dlls alongside your .exe.

Make sure you match release libraries with a release build and debug libraries with a debug build. Make sure you're building either for Win32 or x64 depending on which pre-installed binary you installed.

The CMakeLists.txt doesn't need to be so verbose, fragile. Moreover, you are not linking the libraries, just specifying the paths to search for. Related

Docs for link_directories As it says in the docs, this is bad practice since find_package is supposed to do all this for you.

You need to specify a target_link_libraries in order to find the symbols.

I use the following (link to repo):

Welcome to Stack Overflow. What the documentation says is:

Given a geometric surface, it’s usually trivial to infer the direction of the normal at a certain point on the surface as the vector perpendicular to the surface in that point.

From what I gather from what you say is that you already have the surfaces for which you can easily calculate surface normals. Normal estimation is used because 3D point cloud data is basically a bunch of sample points from the real world. You do not have surface information in this kind of data. What you do is you estimate a surface around a pixel using Planar Fitting(2D Regression). Then, you obtain surface normal. You cannot compare these two methods. They essentially gave different purposes.

For question two: Yes. Refer to this SO answer.

You can close an open viewer at any time with the method

For anyone still digging into this problem, PCL Visualizer is not prepared to handle different color spaces other than RGB (and intensity types) for display purposes. Make use of pcl::PointXYZHSVtoXYZRGB to handle the point cloud conversion

Order matters when linking... you need to make sure that libraries are listed before their dependencies

Looking at your command you've listed libpcl_visualization AFTER your vtk library - meaning that it will not be able to use it.

You can use getFPS(). The issue was fixed in https://github.com/PointCloudLibrary/pcl/pull/1974

Then you can do something like

From the pcl visualizer example: