This guide describes how to install the dependencies and compile PointCloudTools on Windows OS in a quick and easy way. It is assumed that you already have Visual Studio 2019/2022 with the Desktop development with C++ workload installed.
Download the official prebuilt Boost binaries from the following repository:
https://sourceforge.net/projects/boost/files/boost-binaries/
Select a recent version (minimum requirement is 1.58), compatible with your MSVC version (Visual C++ 2022 is MSVC 14.3).
64 bit version is preferred to be installed, but the 32 bit version can also be installed if required.
Download the unofficial prebuilt GDAL binaries for Windows from GISInternals:
https://www.gisinternals.com/release.php
Select a recent version matching your Visual Studio / MSVC version.
64 bit version is preferred to be installed, but the 32 bit version can also be installed if required.
Note: There is two separate download links. You have to download both Compiled binaries AND Compiled libraries and headers, then extract them to the same folder.
Download the official prebuilt OpenCV binaries for Windows from their website:
https://opencv.org/releases/
Select a recent version (minimum requirement is 4.2).
The downloadable executable is a self-extracting archive, which you can extract at any preferred location.
Note: for more recent versions, only the 64 bit version of OpenCV is shipped in this self-extracting archive, but that should be preferred anyway.
Download the official prebuilt binaries for Microsoft's implementation of the MPI library from the follwoing website:
https://learn.microsoft.com/en-us/message-passing-interface/microsoft-mpi
For development, you especially need to install the SDK, execute the msmpisdk.msi
file.
Set up the following environment variables with the appropriate paths used in your system.
Name | Value |
---|---|
BOOST_ROOT |
C:\Programs\Boost\boost_1_81_0 |
GDAL_ROOT |
C:\Programs\GDAL\3.6.2 |
GDAL_BIN |
C:\Programs\GDAL\3.6.2\bin;C:\Programs\GDAL\3.6.2\bin\gdal\apps |
GDAL_DATA |
C:\Programs\GDAL\3.6.2\bin\gdal-data |
GDAL_DRIVER_PATH |
C:\Programs\GDAL\3.6.2\bin\gdal\plugins |
PROJ_LIB |
C:\Programs\GDAL\3.6.2\bin\proj6\share |
OpenCV_DIR |
C:\Programs\OpenCV\build |
MPI_HOME |
C:\Program Files (x86)\Microsoft SDKs\MPI |
Note: for newer GDAL versions 3.8+, there is a proj9
folder instead of proj6
, which should be used.
Also add the GDAL and the OpenCV binary directories to the PATH
environment variable, as these will be linked dynamically:
- add the paths in the
GDAL_BIN
variable; - add the OpenCV binary directory (e.g.
C:\Programs\OpenCV\build\x64\vc16\bin
).
Launch a Developer Command Prompt for VS and generate the build environment with CMake:
mkdir build
cd build
cmake .. -DCMAKE_INSTALL_PREFIX=..\install
Build the solution and optionally install the binaries to the specified install
destination.
msbuild CloudTools.sln
msbuild INSTALL.vcxproj
As a preliminary steps (has to be done only once),
create a copy of the CMakeSettings.json.sample
file with the
name CMakeSettings.json
.
This file contains the prepared build configurations which will be loaded by Visual Studio for the Ninja generator used by CMake.
Then simply open the folder of the project in VS, as Visual Studio has integrated CMake support. Build steps in Visual Studio:
- Select the
x64-Debug
orx64-Release
build configuration in the top menu bar.
Remark: You may use the x86 versions if you have installed the 32 bit dependencies. Note however, that 32 bit processes have limited memory access around 3GB, which can be an issue when processing larger files with PointCloudTools. - CMake should be executed automatically, but you can enforce to run it with the Project -> Delete Cache and Reconfigure option.
- Build the project with the Build -> Build All menu option.
- Install the project to the default
install
folder with the Build -> Install CloudTools option.