AprilTag detect (C++): feasible with OpenCV 3.4.16

TODO:

configure the camera parameters according to your own devices

Pre-requisitions

• cmake
• OpenCV >= 2.3
• GLUT or freeglut (optional, used for test program)
• Cairo (optional, used to print tags)
• CGAL (optional, used for new quad detection algorithm)

OpenCV 3.4.16

• Install OpenCV following this guide; Only by the 2nd step we choose 3.4.16 instead of 4.6.0

• after all the percedures, check your opencv version with this command:

  pkg-config --modversion opencv

  it should return the version number 3.4.16.
  IF NOT:
  e.g. returns a message like this:

  Package opencv was not found in the pkg-config search path.
  Perhaps you should add the directory containing `opencv.pc'
  to the PKG_CONFIG_PATH environment variable
  No package 'opencv' found

  It might because your opencv not yet added to the path correctly. follow the solotion here

  • cd to /usr/local/lib/pkgconfig, open in terminal

  • touch opencv.pc
    sudo gedit opencv.pc  # here only whith sudo you can save your changes!

  • the contents of opencv.pc is:

    # Package Information for pkg-config
    
    prefix=/usr/local
    exec_prefix=${prefix}
    libdir=${exec_prefix}/lib
    includedir_old=${prefix}/include/opencv
    includedir_new=${prefix}/include
    
    Name: OpenCV
    Description: Open Source Computer Vision Library
    Version: 3.4.16
    Libs: -L${exec_prefix}/lib -lopencv_dnn -lopencv_highgui -lopencv_ml -lopencv_objdetect -lopencv_shape -lopencv_stitching -lopencv_superres -lopencv_videostab -lopencv_calib3d -lopencv_videoio -lopencv_imgcodecs -lopencv_features2d -lopencv_video -lopencv_photo -lopencv_imgproc -lopencv_flann -lopencv_viz -lopencv_core
    Libs.private: -ldl -lm -lpthread -lrt
    Cflags: -I${includedir_old} -I${includedir_new}
    

  • check again using command pkg-config --modversion opencv

• Now you should have finished with the installation and setup for OpenCV!

canberra-gtk-module

sudo apt-get install libcanberra-gtk-module

Install and Build

1) Clone the repo

clone this git repo to local
and thanks @swatbotics for their awesome work!

2) Modify the CMakelists.txt in the /src folder:

This part IS ALREADY DONE.
Just in case you run into any similar errors and you can fix them referring to the following information.

ATTENTION: don't mix with the CMakeLists.txt in the root folder

add target_link_libraries(<main_name> ${OpenCV_LIBS}) for EVERY executable's build task.
(or for convience you can simply copy the following and substitute all the contents of your original CMakeLists.txt in /src)

add_library(apriltags
  CameraUtil.cpp
  DebugImage.cpp
  Geometry.cpp
  GrayModel.cpp
  MathUtil.cpp
  Refine.cpp
  TagDetector.cpp
  TagFamily.cpp
  TagFamilies.cpp
  UnionFindSimple.cpp
)

find_package(OpenCV REQUIRED)

include_directories(${OpenCV_INCLUDE_DIRS}) # not needed for opencv>=4.0

set(AT_LIBS apriltags ${OPENCV_LDFLAGS})

target_link_libraries(apriltags ${OpenCV_LIBS}) # add a target link for opencv here
target_link_libraries(apriltags ${AT_LIBS})

add_executable(camtest camtest.cpp)
target_link_libraries(camtest ${OpenCV_LIBS}) # add a target link for opencv here
target_link_libraries(camtest ${AT_LIBS})

add_executable(tagtest tagtest.cpp)
target_link_libraries(tagtest ${OpenCV_LIBS}) # add a target link for opencv here
target_link_libraries(tagtest ${AT_LIBS})

add_executable(quadtest quadtest.cpp)
target_link_libraries(quadtest ${OpenCV_LIBS}) # add a target link for opencv here
target_link_libraries(quadtest ${AT_LIBS})

if(GLUT_LIBRARY)
    add_executable(gltest gltest.cpp)
    target_link_libraries(gltest ${GLUT_LIBRARY} ${OPENGL_LIBRARY} ${AT_LIBS})
endif()

if(CAIRO_FOUND)
    add_executable(maketags maketags.cpp)
    target_link_libraries(maketags ${CAIRO_LIBRARIES} ${AT_LIBS} ${CAIRO_LIBS})
endif()

3) Compile the code

cd to the root folder /apriltags-cpp

mkdir build
cd build
cmake .. -DCMAKE_BUILD_TYPE=Release
make

Possible issues

• missing some packages

  install them

• errors like "xxx was not declared in this scope"

  like this:

  /home/jun/apriltag_ros2/apriltags-cpp/src/gltest.cpp:490:18: error: ‘CV_CAP_PROP_FRAME_WIDTH’ was not declared in this scope
  490 |     capture->set(CV_CAP_PROP_FRAME_WIDTH, opts.frame_width);
      |                  ^~~~~~~~~~~~~~~~~~~~~~~
  /home/jun/apriltag_ros2/apriltags-cpp/src/gltest.cpp:491:18: error: ‘CV_CAP_PROP_FRAME_HEIGHT’ was not declared in this scope
  491 |     capture->set(CV_CAP_PROP_FRAME_HEIGHT, opts.frame_height);
      |                  ^~~~~~~~~~~~~~~~~~~~~~~~

  This is because of the mismatch of different OpenCV versions. The variable names in this project's scripts are no longer consistent with them of our installed OpenCV library.
  You can modify them one by one according to the error messages (and that is exactly how I managed to build this ... x_x)
  Some common substitutions:

  CV_AA -> cv::LINE_AA
  CV_CAP_PROP_FRAME_HEIGHT -> cv::CAP_PROP_FRAME_HEIGHT
  CV_CAP_PROP_FRAME_WIDTH -> cv::CAP_PROP_FRAME_WIDTH
  CV_INTER_AREA-> cv::INTER_AREA
  CV_RGB2GRAY -> cv::COLOR_RGB2GRAY
  ...

  You can search them on Google if you don't know the variable name in new OpneCV version
  Theoretically you won't run into errors like this if you use OpenCV 3.4.16, but in case you can refer to these tips.

Run

Demo/Utility programs

The APRIL tags library is intended to be used as a library, from C++, but there are also five demo/utility programs included in this distribution:

• tagtest- Demonstrate tag recognition and time profiling.
  • There are some test images in the /images directory, that may be useful to use with the tagtest program.
• camtest- Demonstrate 3D tag locations using OpenCV to visualize, with an attached camera.
• gltest- Demonstrate 3D tag locations using OpenGL to visualize, with an attached camera.
• quadtest- Demonstrate/test tag position refinement using a template tracking approach.
• maketags- Create PDF files for printing tags.

---

1) tagtest using local pictures of AprilTags

# stay in /apriltags-cpp/build
./tagtest ../images/iphonecam/IMG_0612.jpg

2) camtest with our external SONY a7r4

Setup SONY connection: SONY a7r4 setup for Ubuntu

• Terminal A: connect SONY camera
  connect the camera to computer, turn on the camera and setup the correct USB connection options on camera, then:

  cd /path/to/v4l2loopback
  sudo modprobe v4l2loopback
  ls -1 /sys/devices/virtual/video4linux # check lookback devices, make sure "video2" in returned message
  sudo modprobe v4l2loopback exclusive_caps=1 max_buffers=2
  gphoto2 --stdout --capture-movie | ffmpeg -i - -vcodec rawvideo -pix_fmt yuv420p -threads 8 -f v4l2 /dev/video2

• Terminal B: launch camtest

  # cd to /apriltags-cpp/build
  ./camtest

  You can modify the device id in the source code /src/camtest.cpp

  typedef struct CamTestOptions {
    CamTestOptions() :
        params(),
        family_str(DEFAULT_TAG_FAMILY),
        error_fraction(1),
        //device_num(0),  // 0 is the laptop's built-in webcam
        device_num(2),  // here 2 is our external SONY a7r4
        focal_length(500),
        tag_size(0.1905),
        frame_width(0),
        frame_height(0),
        mirror_display(true)
    {
    }
    TagDetectorParams params;
    std::string family_str;
    double error_fraction;
    int device_num;
    double focal_length;
    double tag_size;
    int frame_width;
    int frame_height;
    bool mirror_display;
  } CamTestOptions;

  and then recompile:

  cd apriltags-cpp/build
  make

  The result of cam test: