1. Setup

This tutorial guides you through the processes and settings for installing the required software and packages for the ACon Sumo openCV programm.

1.1. Installing the required software

Install the following software on your computer

1.1.1. Microsoft Visual Studio 2017 Community

During the installation process, select Desktop Development with C++ (Desktopentwicklung mit C++)

_images/vs2017.png

After installing, add Visual Studio Compilers to your paths. You may need to restart your computer afterwards or just kill the explorer.exe task in the task manager and execute explorer.exe again.

Tip

On Windows 10, go to Settings, click on System. Then on the left side click About and select System info at the bottom or at the top right hand side. In the new Control Panel window that opens, click Advanced system settings on the left. Now in the new window that comes up, select Environment Variables... at the bottom.

Now select Path and System variables and click Edit and then New. Add the path to the bin folder. For example: C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Tools\MSVC\14.x.x\bin

1.1.2. QT Open Source

During the installation process, make sure you have msvc2017_64 selected. You can uncheck every other component to save space.

_images/qt.png

1.2. Download openCV source files

1. Download openCV master source files. Either download the .zip file from https://github.com/opencv/opencv or checkout using Git https://github.com/opencv/opencv.git

2. Download openCV contrib modules. Either download the .zip file from https://github.com/opencv/opencv_contrib or checkout using Git https://github.com/opencv/opencv_contrib.git

It is recommended to use a folder structure similar to this:

-- opencv
    |- build
    |- opencv-contrib
    |- opencv-master

Unzip the downloaded source files to their respective folder.

1.3. Create your own openCV Build

Start Cmake Gui

_images/cmake1.png
  1. Click Browse source... and locate opencv-master folder
  2. Click Browse build... and locate build folder
  3. Press Configure
_images/cmake2.png
  1. Select the Generator. For example Visual Studio 15 2017 Win64 and select Use default native compilers
  2. Press Finish

Note

This only works if you added the correct Path to your Environment Variables. If you get an error, check the Microsoft Visual Studio 2017 Community Section again.

  1. After configuration is done, it should look similar to this
_images/cmake3.png
  1. Now you need to specify the following values (Use search function)
Name Value
OPENCV_EXTRA_MODULES_PATH path to modules, eg: C:/opencv/opencv-contrib/modules
OPENCV_ENABLE_NONFREE ENABLE
BUILD_DOCS DISABLE (optional, greatly reduces build time)
  1. Press Configure again. New values are displayed in red.
Name Value
BUILD_opencv_aruco ENABLE
BUILD_opencv_ccalib ENABLE
BUILD_opencv_xfeatures2d ENABLE
BUILD_opencv_contrib_world DISABLE
   
WITH_QT ENABLE
  1. You’ll get the following error

Error

Error in configuration process, project files may be invalid.

This is normal. You have to specify the QT5 paths. For this, you have to search inside your QT installation folders. It is assumed that you installed QT in C:/QT/5.9.1. Please set it according to your installation folder:

Name Value
QT_QMAKE_EXECUTABLE C:/QT/5.9.1/msvc2017_64/bin/qmake.exe
Qt5Concurrent_DIR C:/QT/5.9.1/msvc2017_64/lib/cmake/Qt5Concurrent
QT5Core_DIR C:/QT/5.9.1/msvc2017_64/lib/cmake/Qt5Core
QT5Gui_DIR C:/QT/5.9.1/msvc2017_64/lib/cmake/Qt5Gui
QT5Test_DIR C:/QT/5.9.1/msvc2017_64/lib/cmake/Qt5Test
QT5Widgets_DIR C:/QT/5.9.1/msvc2017_64/lib/cmake/Qt5Widgets

Note

If you don’t have the msvc2017_64 folder, check the QT Open Source Section again.

  1. Press Configure again. It’ll show a new error and ask for the QT5OpenGL_DIR
Name Value
QT5OpenGL_DIR C:/QT/5.9.1/msvc2017_64/lib/cmake/Qt5OpenGL
  1. Press Configure again.
  2. Press Generate.
  3. Inside the build folder open the Visual Studio Solution File OpenCV.sln with Microsoft Visual Studio.
_images/vs1.png
  1. Select the right configuration (Debug vs Release).

Tip

It is recommended that you build the Release Configuration because it is a lot faster.

Rightclick on ALL_BUILD and then build. Afterwards rightclick on INSTALL and click build again.

  1. Sit back and drink a coffee. This may take a while.
  2. If there are no errors, there should be a new install folder inside your build folder.

Congratulations! You build your own openCV build with additional contrib modules and QT.

1.4. Build your own ACon Sumo opencv program

1. Download ACon Sumo openCV source files. Either download the .zip file from https://github.com/gfre/sumo_opencv or checkout using Git https://github.com/gfre/sumo_opencv.git

Note

Check if the master branch is up to date. Otherwise checkout the develop branch

  1. Open the Visual Studio Solution file ´´sumo_opencv.sln´´ with Microsoft Visual Studio.
_images/vs2.png
  1. Make sure you select the correct Configuration (Release).
_images/sumo.png
  1. In the Solution Explorer, rightclick on sumo_opencv and select Properties. The go to C/C++ and specify Additional Include Directories
Name Value
Additional Include Directories C:\opencv\build\install\include
_images/sumo1.png
  1. Under Preprocessor, add the following entry
Name Value
Preprocessor definitions CRT_SECURE_NO_WARNINGS
_images/sumo2.png
  1. Under Linker, add the following entry to Additional Library Directories
Name Value
Additional Library Directories C:\opencv\build\install\x64\vc15\lib
_images/sumo3.png

  1. Under Input, add the following entries to Additional Dependencies:

    opencv_arucoXXX.lib
opencv_calib3dXXX.lib
opencv_coreXXX.lib
opencv_features2dXXX.lib
opencv_flannXXX.lib
opencv_highguiXXX.lib
opencv_imgcodecsXXX.lib
opencv_imgprocXXX.lib
opencv_stitchingXXX.lib
opencv_videoXXX.lib
opencv_videoioXXX.lib
opencv_videostabXXX.lib
opencv_xfeatures2dXXX.lib

Note

Change XXX according to your openCV Version. For example: Version 3.2.0 it would be opencv_aruco320.lib etc. Check your C:\opencv\build\install\x64\vc15\lib folder if you are not sure.

_images/sumo4.png

8. Now you’re ready to go and you can build the project by either pressing Ctrl+Shift+B or by rightclicking on sumo_opencv and selection build. You may get an error that soem .dll files cannot be found. In that case navigate to your build folder and copy all the required dll’s to your sumo_opencv project folder

Copy the dll’s:

opencv_arucoXXX.dll
opencv_calib3dXXX.dll
opencv_coreXXX.dll
opencv_features2dXXX.dll
opencv_flannXXX.dll
opencv_highguiXXX.dll
opencv_imgcodecsXXX.dll
opencv_imgprocXXX.dll
opencv_stitchingXXX.dll
opencv_videoXXX.dll
opencv_videoioXXX.dll
opencv_videostabXXX.dll
opencv_xfeatures2dXXX.dll

from

C:\opencv\build\install\x64\vc15\bin

to

sumo_opencv project folder/sumo_opencv

Additionally you need to copy the QT5 dll files:

Qt5Concurrent.dll
Qt5Core.dll
Qt5Gui.dll
Qt5OpenGl.dll
Qt5Test.dll
Qt5Widgets.dll

from

C:/QT/5.9.1/msvc2017_64/bin

to

sumo_opencv project folder/sumo_opencv

Note

It may be possible that you need additional openCV dll files. In that case just copy them too.

Note

(Optional) If you want to run the programm from an executabel file (sumo_opencv.exe), you need to copy the exact same dll files to sumo_opencv project folder/x64/Release or Debug respetively. Also make sure, that the camera calibration .xml file is in the x64/Relese or x64/Debug. Camera Calibration is covered in a later chapter.

  1. After succesfully building the project a command prompt should open and close immediately. You’re now ready to install and connect the camera. Proceed to the next section.

1.5. Troubleshooting

  • TBA -

If you’re experiencing any errors, please contact your supervisor or send me an email.