這篇文章將說明在 Ubuntu Linux 系統上搭建 OpenCV 開發環境的過程，以支持 Python 和 C++ 編程語言開發基於 OpenCV 的應用，或學習 OpenCV。本文說明的過程在 Ubuntu 20.04 版本的系統上經過測試驗證。

在 Ubuntu 系統上安裝 OpenCV-Python

在 Ubuntu 系統中安裝 OpenCV-Python 有兩種方式：

從 Ubuntu 倉庫安裝可用的預編譯二進制文件
由源碼編譯。

這裏這兩種方式都會介紹。

另一個非常重要的問題是必須的附加庫。OpenCV-Python 只需要 Numpy。但在這個教程中，我們還使用 Matplotlib 進行一些簡單而漂亮的繪圖。Matplotlib 是可選的，但強烈建議安裝它。類似地，我們也將看到 IPython，一個交互式 Python 終端，同樣強烈建議安裝它。

從預編譯二進制文件安裝 OpenCV-Python

在僅僅想要編程和開發 OpenCV 應用程序時，這個方法最好。

在終端中通過如下命令安裝 python3-opencv 包（以 root 用戶）：

$ sudo apt-get install python3-opencv
正在讀取軟件包列表... 完成
正在分析軟件包的依賴關係樹       
正在讀取狀態信息... 完成       
下列軟件包是自動安裝的並且現在不需要了：
  ibus-data python3-ibus-1.0
使用'sudo apt autoremove'來卸載它(它們)。
將會同時安裝下列軟件：
  cpp-8 gcc-8 gcc-8-base gdal-data gfortran gfortran-8 gfortran-9 ibverbs-providers libaec0 libarmadillo9 libarpack2 libcaf-openmpi-3
  libcfitsio8 libcharls2 libcoarrays-dev libcoarrays-openmpi-dev libdap25 libdapclient6v5 libepsilon1 libevent-core-2.1-7 libevent-dev
  libevent-extra-2.1-7 libevent-openssl-2.1-7 libevent-pthreads-2.1-7 libfabric1 libfreexl1 libfyba0 libgcc-8-dev libgdal26 libgdcm3.0
  libgeos-3.8.0 libgeos-c1v5 libgeotiff5 libgfortran-8-dev libgfortran-9-dev libgl2ps1.4 libhdf4-0-alt libhdf5-103 libhdf5-openmpi-103
  libhwloc-dev libhwloc-plugins libhwloc15 libibverbs-dev libibverbs1 libilmbase24 libkmlbase1 libkmldom1 libkmlengine1 liblept5 libminizip1
  libmpx2 libnetcdf-c++4 libnetcdf15 libnl-3-dev libnl-route-3-dev libnuma-dev libodbc1 libogdi4.1 libopencv-calib3d4.2 libopencv-contrib4.2
  libopencv-core4.2 libopencv-dnn4.2 libopencv-features2d4.2 libopencv-flann4.2 libopencv-highgui4.2 libopencv-imgcodecs4.2
  libopencv-imgproc4.2 libopencv-ml4.2 libopencv-objdetect4.2 libopencv-photo4.2 libopencv-shape4.2 libopencv-stitching4.2
  libopencv-superres4.2 libopencv-video4.2 libopencv-videoio4.2 libopencv-videostab4.2 libopencv-viz4.2 libopenexr24 libopenmpi-dev libopenmpi3
  libpmix2 libpq5 libproj15 libpsm-infinipath1 libpsm2-2 libqhull7 librdmacm1 libsocket++1 libspatialite7 libsuperlu5 libsz2 libtbb2
  libtesseract4 liburiparser1 libvtk6.3 libxerces-c3.2 libxnvctrl0 odbcinst odbcinst1debian2 openmpi-bin openmpi-common proj-bin proj-data
  python3-numpy
建議安裝：
  gcc-8-locales gcc-8-multilib gcc-8-doc gfortran-multilib gfortran-doc gfortran-8-multilib gfortran-8-doc gfortran-9-multilib gfortran-9-doc
  geotiff-bin gdal-bin libgeotiff-epsg libhdf4-doc libhdf4-alt-dev hdf4-tools libhwloc-contrib-plugins libmyodbc odbc-postgresql tdsodbc
  unixodbc-bin ogdi-bin openmpi-doc mpi-default-bin python-numpy-doc python3-pytest python3-numpy-dbg
下列【新】軟件包將被安裝：
  cpp-8 gcc-8 gcc-8-base gdal-data gfortran gfortran-8 gfortran-9 ibverbs-providers libaec0 libarmadillo9 libarpack2 libcaf-openmpi-3
  libcfitsio8 libcharls2 libcoarrays-dev libcoarrays-openmpi-dev libdap25 libdapclient6v5 libepsilon1 libevent-core-2.1-7 libevent-dev
  libevent-extra-2.1-7 libevent-openssl-2.1-7 libevent-pthreads-2.1-7 libfabric1 libfreexl1 libfyba0 libgcc-8-dev libgdal26 libgdcm3.0
  libgeos-3.8.0 libgeos-c1v5 libgeotiff5 libgfortran-8-dev libgfortran-9-dev libgl2ps1.4 libhdf4-0-alt libhdf5-103 libhdf5-openmpi-103
  libhwloc-dev libhwloc-plugins libhwloc15 libibverbs-dev libibverbs1 libilmbase24 libkmlbase1 libkmldom1 libkmlengine1 liblept5 libminizip1
  libmpx2 libnetcdf-c++4 libnetcdf15 libnl-3-dev libnl-route-3-dev libnuma-dev libodbc1 libogdi4.1 libopencv-calib3d4.2 libopencv-contrib4.2
  libopencv-core4.2 libopencv-dnn4.2 libopencv-features2d4.2 libopencv-flann4.2 libopencv-highgui4.2 libopencv-imgcodecs4.2
  libopencv-imgproc4.2 libopencv-ml4.2 libopencv-objdetect4.2 libopencv-photo4.2 libopencv-shape4.2 libopencv-stitching4.2
  libopencv-superres4.2 libopencv-video4.2 libopencv-videoio4.2 libopencv-videostab4.2 libopencv-viz4.2 libopenexr24 libopenmpi-dev libopenmpi3
  libpmix2 libpq5 libproj15 libpsm-infinipath1 libpsm2-2 libqhull7 librdmacm1 libsocket++1 libspatialite7 libsuperlu5 libsz2 libtbb2
  libtesseract4 liburiparser1 libvtk6.3 libxerces-c3.2 libxnvctrl0 odbcinst odbcinst1debian2 openmpi-bin openmpi-common proj-bin proj-data
  python3-numpy python3-opencv
升級了 0 個軟件包，新安裝了 105 個軟件包，要卸載 0 個軟件包，有 70 個軟件包未被升級。
需要下載 104 MB 的歸檔。
解壓縮後會消耗 397 MB 的額外空間。
您希望繼續執行嗎？ [Y/n] y
獲取:1 http://cn.archive.ubuntu.com/ubuntu focal/universe amd64 gcc-8-base amd64 8.4.0-3ubuntu2 [18.7 kB]
. . . . . .

它會安裝非常多在 Python 中調用 OpenCV 接口所需的依賴，如 libopencv-core4.2、libopencv-superres4.2、libopencv-objdetect4.2 和 libopencv-features2d4.2 等。注意，安裝的這些 OpenCV 依賴庫的包只包含了動態鏈接庫二進制文件，而沒有 C++ OpenCV 開發所需要的頭文件等內容。

打開一個 Python IDLE （或者 IPython）並在 Python 終端中輸入如下代碼：

import cv2 as cv
print(cv.__version__)

執行這段代碼，輸出了適當的 OpenCV 版本號而沒有任何報錯，則恭喜你！！！你已經成功地安裝了 OpenCV-Python。

這很簡單。但這有一個問題。Apt 倉庫可能總是不包含最新版本的 OpenCV。比如筆者在 Ubuntu 20.04 上通過上面的操作安裝 OpenCV，得到的版本爲 4.2.0，最新的版本則爲 4.5.5：

$ python3
Python 3.8.10 (default, Nov 26 2021, 20:14:08) 
[GCC 9.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import cv2
>>> print(cv2.__version__)
4.2.0

對於 Python API，最新版本將始終包含更好的支持和最新的錯誤修復。

因此，獲取最新源代碼的下一個方法，即從源代碼編譯。同時，如果在某個時間點，你想爲 OpenCV 貢獻代碼，你也將需要這個。

從源碼構建

乍一看去，從源碼編譯可能有點複雜，但一旦你成功了，你會發現其實也沒什麼複雜的。

首先我們將安裝一些依賴。有些是必須的，有些是可選的。如果你不需要，可以跳過可選的依賴。

必須的構建依賴

我們需要 CMake 來配置安裝，需要 GCC 來編譯。需要 Python-devel 和 Numpy 來構建 Python 綁定等。

sudo apt-get install cmake
sudo apt-get install gcc g++

要支持 python2：

sudo apt-get install python-dev python-numpy

要支持 python3：

sudo apt-get install python3-dev python3-numpy

接着，需要 GTK 來支持 GUI 功能，Camera 支持 (v4l)，媒體支持 (ffmpeg，gstreamer) 等等。

sudo apt-get install libavcodec-dev libavformat-dev libswscale-dev
sudo apt-get install libgstreamer-plugins-base1.0-dev libgstreamer1.0-dev

要支持 gtk2：

sudo apt-get install libgtk2.0-dev

要支持 gtk3：

sudo apt-get install libgtk-3-dev

可選的依賴

對於在 Ubuntu 機器上安裝 OpenCV，上面的依賴已經足夠了。但根據你的需求，你可能需要一些額外的依賴。這樣的額外依賴列表如下。你可以根據你的需求把它們留在那裏或安裝它們。

OpenCV 附帶了對圖像格式的支持，如 PNG、JPEG、JPEG2000、TIFF、WebP 等。但它可能有點老。如果你想獲得最新的庫，你可以爲這些格式的系統庫安裝開發文件。

sudo apt-get install libpng-dev
sudo apt-get install libjpeg-dev
sudo apt-get install libopenexr-dev
sudo apt-get install libtiff-dev
sudo apt-get install libwebp-dev

注意： 如果你正在使用 Ubuntu 16.04，也可以安裝 libjasper-dev 來給你的系統添加對 JPEG2000 格式的系統級支持。

下載 OpenCV

獲得 OpenCV 的源碼有兩種方式：

從 OpenCV 的 Github 倉庫下載最新的源碼。（如果你想要給 OpenCV 貢獻代碼也可以通過它。要下載源碼，首先需要安裝 Git。）

$ sudo apt-get install git
$ git clone https://github.com/opencv/opencv.git

這將在當前目錄創建一個名爲 "opencv" 的目錄。代碼克隆過程可能需要一些時間，這依賴於你的網絡連接情況。

除了直接克隆 OpenCV 的 Github 倉庫之外，還有另外一種獲取 OpenCV 源碼的方法。OpenCV 官方提供了已經打包好的發行版源碼的壓縮包，可以下載這些壓縮包獲取源碼，如：

# Install minimal prerequisites
$ sudo apt update && sudo apt install -y wget unzip

# Download and unpack sources
$ wget -O opencv.zip https://github.com/opencv/opencv/archive/4.x.zip
$ unzip opencv.zip
$ mv opencv-4.x opencv

解壓之後並重命名源碼目錄，OpenCV 源碼將位於 opencv 目錄下。

現在打開一個終端窗口，切換到下載的 "opencv" 目錄。創建一個名爲 "build" 的目錄並切換到它下面。

$ mkdir build
$ cd build

當然也可以在 OpenCV 源碼目錄的同級目錄中創建 "build" 目錄，這同樣需要切換到新建的目錄下。注意，後面執行 CMake 配置構建時，需要正確引用 OpenCV 源碼目錄。

配置和安裝

現在我們已經擁有了所有必須的依賴，讓我們開始安裝 OpenCV 吧。安裝必須通過 CMake 來配置。它指定了將要安裝哪些模塊、安裝路徑、將使用哪些附加庫、是否要編譯文檔和示例等。這些工作中的大部分將通過默認參數的良好配置自動完成。

以下命令通常用於配置 OpenCV 庫構建（在 "build" 目錄中執行）：

$ cmake ../

OpenCV 默認假設構建類型爲 "Release"，安裝路徑爲 "/usr/local"。關於 CMake 選項的其它信息請參考 OpenCV C++ 編譯指南。

你應該能夠在你的 CMake 輸出中看到如下這些行（它們意味着已經發現了適當的 Python）：

--   Other third-party libraries:
--     VA:                          YES
--     Lapack:                      NO
--     Eigen:                       NO
--     Custom HAL:                  NO
--     Protobuf:                    build (3.19.1)
-- 
--   OpenCL:                        YES (INTELVA)
--     Include path:                /media/data/my_multimedia/opencv-4.x/3rdparty/include/opencl/1.2
--     Link libraries:              Dynamic load
-- 
--   Python 2:
--     Interpreter:                 /usr/bin/python2.7 (ver 2.7.18)
--     Libraries:                   /usr/lib/x86_64-linux-gnu/libpython2.7.so (ver 2.7.18)
--     numpy:                       /usr/lib/python2.7/dist-packages/numpy/core/include (ver 1.16.5)
--     install path:                lib/python2.7/dist-packages/cv2/python-2.7
-- 
--   Python 3:
--     Interpreter:                 /usr/bin/python3 (ver 3.8.10)
--     Libraries:                   /usr/lib/x86_64-linux-gnu/libpython3.8.so (ver 3.8.10)
--     numpy:                       /usr/lib/python3/dist-packages/numpy/core/include (ver 1.17.4)
--     install path:                lib/python3.8/site-packages/cv2/python-3.8
-- 
--   Python (for build):            /usr/bin/python2.7
-- 
--   Java:                          
--     ant:                         NO
--     JNI:                         /usr/lib/jvm/default-java/include /usr/lib/jvm/default-java/include/linux /usr/lib/jvm/default-java/include
--     Java wrappers:               NO
--     Java tests:                  NO
-- 
--   Install to:                    /usr/local

cmake 命令的參數是包含 CMakelists.txt 文件的目錄的路徑，即源碼目錄。如果 "build" 目錄在 "opencv" 源碼目錄的同級目錄中創建，則爲（在 "build" 目錄中執行）：

# Configure
$ cmake ../opencv

現在可以使用 "make" 命令來構建文件：

$ make
. . . . . .
Scanning dependencies of target opencv_version
[100%] Building CXX object apps/version/CMakeFiles/opencv_version.dir/opencv_version.cpp.o
[100%] Linking CXX executable ../../bin/opencv_version
[100%] Built target opencv_version
Scanning dependencies of target opencv_model_diagnostics
[100%] Building CXX object apps/model-diagnostics/CMakeFiles/opencv_model_diagnostics.dir/model_diagnostics.cpp.o
[100%] Linking CXX executable ../../bin/opencv_model_diagnostics
[100%] Built target opencv_model_diagnostics

構建也可以通過 cmake 命令完成：

$ cmake --build .

成功構建之後，可以在 build/lib目錄下找到庫文件，並在 build/bin 目錄下找到可執行文件（測試，示例，和 apps）：

build$ ls bin
opencv_annotation               opencv_perf_optflow      opencv_test_bgsegm        opencv_test_intensity_transform  opencv_test_stitching
opencv_interactive-calibration  opencv_perf_photo        opencv_test_bioinspired   opencv_test_line_descriptor      opencv_test_structured_light
opencv_model_diagnostics        opencv_perf_reg          opencv_test_calib3d       opencv_test_mcc                  opencv_test_superres
opencv_perf_aruco               opencv_perf_rgbd         opencv_test_core          opencv_test_ml                   opencv_test_text
opencv_perf_bioinspired         opencv_perf_stereo       opencv_test_dnn           opencv_test_objdetect            opencv_test_tracking
opencv_perf_calib3d             opencv_perf_stitching    opencv_test_dnn_superres  opencv_test_optflow              opencv_test_video
opencv_perf_core                opencv_perf_superres     opencv_test_face          opencv_test_phase_unwrapping     opencv_test_videoio
opencv_perf_dnn                 opencv_perf_tracking     opencv_test_features2d    opencv_test_photo                opencv_test_videostab
opencv_perf_dnn_superres        opencv_perf_video        opencv_test_flann         opencv_test_quality              opencv_test_wechat_qrcode
opencv_perf_features2d          opencv_perf_videoio      opencv_test_fuzzy         opencv_test_rapid                opencv_test_xfeatures2d
opencv_perf_gapi                opencv_perf_xfeatures2d  opencv_test_gapi          opencv_test_reg                  opencv_test_ximgproc
opencv_perf_imgcodecs           opencv_perf_ximgproc     opencv_test_highgui       opencv_test_rgbd                 opencv_test_xphoto
opencv_perf_imgproc             opencv_perf_xphoto       opencv_test_imgcodecs     opencv_test_saliency             opencv_version
opencv_perf_line_descriptor     opencv_test_aruco        opencv_test_img_hash      opencv_test_shape                opencv_visualisation
opencv_perf_objdetect           opencv_test_barcode      opencv_test_imgproc       opencv_test_stereo               opencv_waldboost_detector
build$ ls lib
cv2.so                            libopencv_features2d.so.405             libopencv_objdetect.so               libopencv_structured_light.so.4.5.5
libopencv_aruco.so                libopencv_features2d.so.4.5.5           libopencv_objdetect.so.405           libopencv_superres.so
libopencv_aruco.so.405            libopencv_flann.so                      libopencv_objdetect.so.4.5.5         libopencv_superres.so.405
libopencv_aruco.so.4.5.5          libopencv_flann.so.405                  libopencv_optflow.so                 libopencv_superres.so.4.5.5
libopencv_barcode.so              libopencv_flann.so.4.5.5                libopencv_optflow.so.405             libopencv_surface_matching.so
libopencv_barcode.so.405          libopencv_freetype.so                   libopencv_optflow.so.4.5.5           libopencv_surface_matching.so.405
libopencv_barcode.so.4.5.5        libopencv_freetype.so.405               libopencv_phase_unwrapping.so        libopencv_surface_matching.so.4.5.5
libopencv_bgsegm.so               libopencv_freetype.so.4.5.5             libopencv_phase_unwrapping.so.405    libopencv_text.so
libopencv_bgsegm.so.405           libopencv_fuzzy.so                      libopencv_phase_unwrapping.so.4.5.5  libopencv_text.so.405
libopencv_bgsegm.so.4.5.5         libopencv_fuzzy.so.405                  libopencv_photo.so                   libopencv_text.so.4.5.5
libopencv_bioinspired.so          libopencv_fuzzy.so.4.5.5                libopencv_photo.so.405               libopencv_tracking.so
libopencv_bioinspired.so.405      libopencv_gapi.so                       libopencv_photo.so.4.5.5             libopencv_tracking.so.405
libopencv_bioinspired.so.4.5.5    libopencv_gapi.so.405                   libopencv_plot.so                    libopencv_tracking.so.4.5.5
libopencv_calib3d.so              libopencv_gapi.so.4.5.5                 libopencv_plot.so.405                libopencv_ts.a
libopencv_calib3d.so.405          libopencv_hfs.so                        libopencv_plot.so.4.5.5              libopencv_videoio.so
libopencv_calib3d.so.4.5.5        libopencv_hfs.so.405                    libopencv_quality.so                 libopencv_videoio.so.405
libopencv_ccalib.so               libopencv_hfs.so.4.5.5                  libopencv_quality.so.405             libopencv_videoio.so.4.5.5
libopencv_ccalib.so.405           libopencv_highgui.so                    libopencv_quality.so.4.5.5           libopencv_video.so
libopencv_ccalib.so.4.5.5         libopencv_highgui.so.405                libopencv_rapid.so                   libopencv_video.so.405
libopencv_core.so                 libopencv_highgui.so.4.5.5              libopencv_rapid.so.405               libopencv_video.so.4.5.5
libopencv_core.so.405             libopencv_imgcodecs.so                  libopencv_rapid.so.4.5.5             libopencv_videostab.so
libopencv_core.so.4.5.5           libopencv_imgcodecs.so.405              libopencv_reg.so                     libopencv_videostab.so.405
libopencv_datasets.so             libopencv_imgcodecs.so.4.5.5            libopencv_reg.so.405                 libopencv_videostab.so.4.5.5
libopencv_datasets.so.405         libopencv_img_hash.so                   libopencv_reg.so.4.5.5               libopencv_wechat_qrcode.so
libopencv_datasets.so.4.5.5       libopencv_img_hash.so.405               libopencv_rgbd.so                    libopencv_wechat_qrcode.so.405
libopencv_dnn_objdetect.so        libopencv_img_hash.so.4.5.5             libopencv_rgbd.so.405                libopencv_wechat_qrcode.so.4.5.5
libopencv_dnn_objdetect.so.405    libopencv_imgproc.so                    libopencv_rgbd.so.4.5.5              libopencv_xfeatures2d.so
libopencv_dnn_objdetect.so.4.5.5  libopencv_imgproc.so.405                libopencv_saliency.so                libopencv_xfeatures2d.so.405
libopencv_dnn.so                  libopencv_imgproc.so.4.5.5              libopencv_saliency.so.405            libopencv_xfeatures2d.so.4.5.5
libopencv_dnn.so.405              libopencv_intensity_transform.so        libopencv_saliency.so.4.5.5          libopencv_ximgproc.so
libopencv_dnn.so.4.5.5            libopencv_intensity_transform.so.405    libopencv_shape.so                   libopencv_ximgproc.so.405
libopencv_dnn_superres.so         libopencv_intensity_transform.so.4.5.5  libopencv_shape.so.405               libopencv_ximgproc.so.4.5.5
libopencv_dnn_superres.so.405     libopencv_line_descriptor.so            libopencv_shape.so.4.5.5             libopencv_xobjdetect.so
libopencv_dnn_superres.so.4.5.5   libopencv_line_descriptor.so.405        libopencv_stereo.so                  libopencv_xobjdetect.so.405
libopencv_dpm.so                  libopencv_line_descriptor.so.4.5.5      libopencv_stereo.so.405              libopencv_xobjdetect.so.4.5.5
libopencv_dpm.so.405              libopencv_mcc.so                        libopencv_stereo.so.4.5.5            libopencv_xphoto.so
libopencv_dpm.so.4.5.5            libopencv_mcc.so.405                    libopencv_stitching.so               libopencv_xphoto.so.405
libopencv_face.so                 libopencv_mcc.so.4.5.5                  libopencv_stitching.so.405           libopencv_xphoto.so.4.5.5
libopencv_face.so.405             libopencv_ml.so                         libopencv_stitching.so.4.5.5         python3
libopencv_face.so.4.5.5           libopencv_ml.so.405                     libopencv_structured_light.so
libopencv_features2d.so           libopencv_ml.so.4.5.5                   libopencv_structured_light.so.405

編譯出來的可執行文件和庫文件非常多。

CMake 包文件位於構建根目錄：

build$ ls OpenCVConfig*.cmake
OpenCVConfig.cmake  OpenCVConfig-version.cmake
build$ ls OpenCVModules.cmake 
OpenCVModules.cmake

它們用於支持 CMake 的 find_library。

默認情況，OpenCV 將安裝在 /usr/local 目錄下，所有文件將被拷貝如下位置：

/usr/local/bin - 可執行文件
/usr/local/lib - 庫文件 (.so)
/usr/local/cmake/opencv4 - cmake 包
/usr/local/include/opencv4 - 頭文件
/usr/local/share/opencv4 - 其它文件（比如 XML 格式的訓練的級聯）

由於 /usr/local 爲 root 用戶所有，因而安裝應該以超級用戶特權執行 (sudo)，使用 "make install" 命令來安裝它：

$ sudo make install

安裝結束。所有的文件都被安裝在了 "/usr/local/" 目錄下。打開一個終端並嘗試導入 "cv2"：

import cv2 as cv
print(cv.__version__)

構建配置和安裝過程的進一步說明

這一部分將提供更多構建過程的細節說明，並描述了其它可選的方法和工具。請參考 OpenCV 安裝概述瞭解一般的安裝細節說明，並參考 OpenCV 配置選項參考瞭解配置選項文檔。

編譯器和構建工具

要編譯 OpenCV，我們就需要 C++ 編譯器。通常是 G++/GCC 或 Clang/LLVM。上面我們看到了 G++/GCC 的安裝方法，這裏給出安裝 Clang/LLVM 的命令：

$ sudo apt install -y clang

無論何種編譯工具，構建 OpenCV 都需要用 CMake 來執行構建配置。

CMake 可以爲不同的構建系統生成構建腳本，除了上面看到的 make，在 Ubuntu 上還可以用 ninja 來構建。Ninja 的安裝命令如下：

$ sudo apt install -y ninja-build

配置和構建

CMake 默認爲 make 構建系統生成構建腳本，如上面那樣。但我們可以通過 CMake 命令行參數強制爲 ninja 構建系統生成構建腳本，如：

$ cmake -GNinja ../opencv

再提一點關於安裝目標目錄配置的問題。

安裝過程僅將文件複製到預定義的位置並進行少量修補。使用這種方法，無論是通過 make，還是通過 ninja，安裝都不會將 opencv 集成到系統包註冊表中，如無法通過 apt 來管理這些安裝。因此，也無法自動卸載 opencv。一般來說，最好不要進行系統範圍的安裝，因爲可能與系統包衝突。

安裝根目錄可以通過 CMAKE_INSTALL_PREFIX 配置參數改變，比如 -DCMAKE_INSTALL_PREFIX=$HOME/.local 指明安裝到當前用戶的本地目錄。安裝佈局可以通過 OPENCV_*_INSTALL_PATH 參數修改。詳情請參考 OpenCV 配置選項參考。

另外，OpenCV 有一些所謂的 "extra" 模塊，它們也提供了一些功能，源碼位於單獨的 git 倉庫內。新模塊常常不具有穩定的 API，且還沒經過良好地測試。由於 OpenCV 庫想要提供良好的二進制兼容性，並嘗試提供良好的性能和穩定性，這樣，新模塊不應該作爲正式的 OpenCV 分發版的一部分發型。

所有的新模塊單獨開發，並首先發布到 opencv_contrib 倉庫中。隨後模塊變得成熟並流行時，再被移入 OpenCV 主倉庫中，開發團隊再爲這個模塊提供產品級質量的支持。在構建時，想要一起編譯 opencv_contrib 的話，首先需要下載 opencv_contrib 的源碼。方法同樣有兩個，一是下載官方發佈的源碼包：

$ wget -O opencv_contrib.zip https://github.com/opencv/opencv_contrib/archive/4.x.zip
$ unzip opencv_contrib.zip

二是克隆 Github 倉庫：

$ git clone https://github.com/opencv/opencv_contrib.git

假設 opencv_contrib 的源碼被下載到了 opencv 源碼的同級目錄中。

在通過 CMake 執行構建配置時，通過 -DOPENCV_EXTRA_MODULES_PATH 參數選項指定 opencv_contrib 的模塊的路徑。

這樣，配置通過 ninja 構建，連同 opencv_contrib 一起構建，並將 OpenCV 安裝在用戶本地目錄的命令如下：

$ cmake -GNinja -DCMAKE_INSTALL_PREFIX=$HOME/.local -DOPENCV_EXTRA_MODULES_PATH=../opencv_contrib-4.x/modules ../opencv

對於構建，在使用 make 構建系統時，可以給 make 傳參數，以啓用多線程構建，如：

$ make -j4

運行 ninja 執行構建的命令如下：

$ ninja

運行 ninja 執行安裝的命令如下：

$ ninja install
[0/1] Install the project...
-- Install configuration: "Release"
-- Installing: /home/zhangsan/.local/share/licenses/opencv4/opencl-headers-LICENSE.txt
-- Installing: /home/zhangsan/.local/share/licenses/opencv4/ade-LICENSE
-- Installing: /home/zhangsan/.local/include/opencv4/opencv2/cvconfig.h
-- Installing: /home/zhangsan/.local/include/opencv4/opencv2/opencv_modules.hpp
-- Installing: /home/zhangsan/.local/lib/cmake/opencv4/OpenCVModules.cmake
-- Installing: /home/zhangsan/.local/lib/cmake/opencv4/OpenCVModules-release.cmake
-- Installing: /home/zhangsan/.local/lib/cmake/opencv4/OpenCVConfig-version.cmake
-- Installing: /home/zhangsan/.local/lib/cmake/opencv4/OpenCVConfig.cmake
-- Installing: /home/zhangsan/.local/bin/setup_vars_opencv4.sh
-- Installing: /home/zhangsan/.local/share/opencv4/valgrind.supp
-- Installing: /home/zhangsan/.local/share/opencv4/valgrind_3rdparty.supp
-- Installing: /home/zhangsan/.local/share/licenses/opencv4/libopenjp2-README.md
-- Installing: /home/zhangsan/.local/share/licenses/opencv4/libopenjp2-LICENSE
-- Installing: /home/zhangsan/.local/share/licenses/opencv4/protobuf-LICENSE
-- Installing: /home/zhangsan/.local/share/licenses/opencv4/protobuf-README.md
-- Installing: /home/zhangsan/.local/share/licenses/opencv4/quirc-LICENSE
-- Installing: /home/zhangsan/.local/share/licenses/opencv4/ittnotify-LICENSE.BSD
-- Installing: /home/zhangsan/.local/share/licenses/opencv4/ittnotify-LICENSE.GPL
-- Installing: /home/zhangsan/.local/include/opencv4/opencv2/opencv.hpp
. . . . . .
-- Installing: /home/zhangsan/.local/lib/libopencv_optflow.so.4.5.5
-- Installing: /home/zhangsan/.local/lib/libopencv_optflow.so.405
-- Set runtime path of "/home/zhangsan/.local/lib/libopencv_optflow.so.4.5.5" to "/home/zhangsan/.local/lib"
-- Installing: /home/zhangsan/.local/lib/libopencv_optflow.so
-- Installing: /home/zhangsan/.local/include/opencv4/opencv2/optflow.hpp
-- Installing: /home/zhangsan/.local/include/opencv4/opencv2/optflow/motempl.hpp
-- Installing: /home/zhangsan/.local/include/opencv4/opencv2/optflow/pcaflow.hpp
-- Installing: /home/zhangsan/.local/include/opencv4/opencv2/optflow/rlofflow.hpp
-- Installing: /home/zhangsan/.local/include/opencv4/opencv2/optflow/sparse_matching_gpc.hpp
-- Installing: /home/zhangsan/.local/lib/libopencv_stitching.so.4.5.5
-- Installing: /home/zhangsan/.local/lib/libopencv_stitching.so.405
. . . . . .
-- Installing: /home/zhangsan/.local/include/opencv4/opencv2/stereo/stereo.hpp
-- Installing: /home/zhangsan/.local/lib/python2.7/dist-packages/cv2/__init__.py
-- Installing: /home/zhangsan/.local/lib/python2.7/dist-packages/cv2/load_config_py2.py
-- Installing: /home/zhangsan/.local/lib/python2.7/dist-packages/cv2/load_config_py3.py
-- Installing: /home/zhangsan/.local/lib/python2.7/dist-packages/cv2/config.py
-- Installing: /home/zhangsan/.local/lib/python2.7/dist-packages/cv2/misc/__init__.py
-- Installing: /home/zhangsan/.local/lib/python2.7/dist-packages/cv2/misc/version.py
-- Installing: /home/zhangsan/.local/lib/python2.7/dist-packages/cv2/mat_wrapper/__init__.py
-- Installing: /home/zhangsan/.local/lib/python2.7/dist-packages/cv2/utils/__init__.py
-- Installing: /home/zhangsan/.local/lib/python2.7/dist-packages/cv2/gapi/__init__.py
-- Installing: /home/zhangsan/.local/lib/python2.7/dist-packages/cv2/python-2.7/cv2.so
-- Set runtime path of "/home/zhangsan/.local/lib/python2.7/dist-packages/cv2/python-2.7/cv2.so" to "/home/zhangsan/.local/lib"
-- Installing: /home/zhangsan/.local/lib/python2.7/dist-packages/cv2/config-2.7.py
-- Installing: /home/zhangsan/.local/lib/python3.8/site-packages/cv2/__init__.py
-- Installing: /home/zhangsan/.local/lib/python3.8/site-packages/cv2/load_config_py2.py
-- Installing: /home/zhangsan/.local/lib/python3.8/site-packages/cv2/load_config_py3.py
-- Installing: /home/zhangsan/.local/lib/python3.8/site-packages/cv2/config.py
-- Installing: /home/zhangsan/.local/lib/python3.8/site-packages/cv2/misc/__init__.py
-- Installing: /home/zhangsan/.local/lib/python3.8/site-packages/cv2/misc/version.py
-- Installing: /home/zhangsan/.local/lib/python3.8/site-packages/cv2/mat_wrapper/__init__.py
-- Installing: /home/zhangsan/.local/lib/python3.8/site-packages/cv2/utils/__init__.py
-- Installing: /home/zhangsan/.local/lib/python3.8/site-packages/cv2/gapi/__init__.py
-- Installing: /home/zhangsan/.local/lib/python3.8/site-packages/cv2/python-3.8/cv2.cpython-38-x86_64-linux-gnu.so
-- Set runtime path of "/home/zhangsan/.local/lib/python3.8/site-packages/cv2/python-3.8/cv2.cpython-38-x86_64-linux-gnu.so" to "/home/zhangsan/.local/lib"
-- Installing: /home/zhangsan/.local/lib/python3.8/site-packages/cv2/config-3.8.py
. . . . . .
-- Installing: /home/zhangsan/.local/share/opencv4/haarcascades/haarcascade_smile.xml
-- Installing: /home/zhangsan/.local/share/opencv4/haarcascades/haarcascade_upperbody.xml
-- Installing: /home/zhangsan/.local/share/opencv4/lbpcascades/lbpcascade_frontalcatface.xml
-- Installing: /home/zhangsan/.local/share/opencv4/lbpcascades/lbpcascade_frontalface.xml
-- Installing: /home/zhangsan/.local/share/opencv4/lbpcascades/lbpcascade_frontalface_improved.xml
-- Installing: /home/zhangsan/.local/share/opencv4/lbpcascades/lbpcascade_profileface.xml
-- Installing: /home/zhangsan/.local/share/opencv4/lbpcascades/lbpcascade_silverware.xml
-- Installing: /home/zhangsan/.local/bin/opencv_annotation
-- Set runtime path of "/home/zhangsan/.local/bin/opencv_annotation" to "/home/zhangsan/.local/lib"
-- Installing: /home/zhangsan/.local/bin/opencv_visualisation
-- Set runtime path of "/home/zhangsan/.local/bin/opencv_visualisation" to "/home/zhangsan/.local/lib"
-- Installing: /home/zhangsan/.local/bin/opencv_interactive-calibration
-- Set runtime path of "/home/zhangsan/.local/bin/opencv_interactive-calibration" to "/home/zhangsan/.local/lib"
-- Installing: /home/zhangsan/.local/bin/opencv_version
-- Set runtime path of "/home/zhangsan/.local/bin/opencv_version" to "/home/zhangsan/.local/lib"
-- Installing: /home/zhangsan/.local/bin/opencv_model_diagnostics
-- Set runtime path of "/home/zhangsan/.local/bin/opencv_model_diagnostics" to "/home/zhangsan/.local/lib"

由於指定的安裝路徑爲用戶本地路徑，因而不需要 sudo。在安裝過程的輸出中，還是能發現不少信息，如 Python 綁定的安裝路徑，頭文件的具體安裝路徑等。

在 Ubuntu 系統上搭建 OpenCV 的 C++ 開發環境

這裏說明配置 CMake 和 Eclipse 構建基於 OpenCV 的 C++ 項目的方法。

準備依賴

構建依賴可以分爲兩部分：

(1). 開發和構建工具，包括 Eclipse，CMake，GCC/G++，Clang/LLVM，Git 等。如果還沒有安裝 Eclipse（含 CDT），可以到 Eclipse 的網站上，下載 Eclipse IDE for C/C++ Developers。根據自己的系統平臺類型選擇下載的版本。其它工具可以參考上面的說明進行安裝，這裏不再贅述。
(2). 用於 C/C++ 開發的 OpenCV 庫開發包，主要包括動態鏈接庫，靜態鏈接庫和頭文件。

對於 OpenCV 庫開發包安裝，有兩種方法：

從源碼構建安裝。如上面從源碼構建安裝 OpenCV-Python 的過程，該過程除了可以構建安裝 OpenCV 的 Python 綁定之外，也同樣會構建安裝C/C++ 開發所需要的所有庫文件、頭文件等內容。對於這種方法，這裏不再贅述。
從 Apt 倉庫安裝預編譯的二進制安裝包。同樣，用這種方式安裝的 OpenCV 庫開發包版本可能比較老。但這種方法更簡單，對於只想利用 OpenCV 的 API 開發應用，而無心研究 OpenCV 的源碼或向 OpenCV 項目貢獻源碼的同學，這種方法比較方便。

對於從從 Apt 倉庫安裝預編譯的二進制安裝包，搜索 Apt 倉庫中的 OpenCV 開發包，可以看到如下內容：

build$ sudo apt-cache search libopencv
libopencv-apps-dev - Opencv_apps Robot OS package - development files
libopencv-apps1d - opencv_apps Robot OS package - runtime files
libopencv-calib3d-dev - development files for libopencv-calib3d4.2
libopencv-calib3d4.2 - computer vision Camera Calibration library
libopencv-contrib-dev - development files for libopencv-contrib4.2
libopencv-contrib4.2 - computer vision contrlib library
libopencv-core-dev - development files for libopencv-core4.2
libopencv-core4.2 - computer vision core library
libopencv-dev - development files for opencv
libopencv-dnn-dev - development files for libopencv-dnn4.2
libopencv-dnn4.2 - computer vision Deep neural network module
libopencv-features2d-dev - development files for libopencv-features2d4.2
libopencv-features2d4.2 - computer vision Feature Detection and Descriptor Extraction library
libopencv-flann-dev - development files for libopencv-flann4.2
libopencv-flann4.2 - computer vision Clustering and Search in Multi-Dimensional spaces library
libopencv-highgui-dev - development files for libopencv-highgui4.2
libopencv-highgui4.2 - computer vision High-level GUI and Media I/O library
libopencv-imgcodecs-dev - development files for libopencv-imgcodecs4.2
libopencv-imgcodecs4.2 - computer vision Image Codecs library
libopencv-imgproc-dev - development files for libopencv-imgproc4.2
libopencv-imgproc4.2 - computer vision Image Processing library
libopencv-ml-dev - development files for libopencv-ml4.2
libopencv-ml4.2 - computer vision Machine Learning library
libopencv-objdetect-dev - development files for libopencv-objdetect4.2
libopencv-objdetect4.2 - computer vision Object Detection library
libopencv-photo-dev - development files for libopencv-photo4.2
libopencv-photo4.2 - computer vision computational photography library
libopencv-shape-dev - development files for libopencv-shape4.2
libopencv-shape4.2 - computer vision shape descriptors and matchers library
libopencv-stitching-dev - development files for libopencv-stitching4.2
libopencv-stitching4.2 - computer vision image stitching library
libopencv-superres-dev - development files for libopencv-superres4.2
libopencv-superres4.2 - computer vision Super Resolution library
libopencv-ts-dev - development files for TS library of OpenCV (Open Computer Vision)
libopencv-video-dev - development files for libopencv-video4.2
libopencv-video4.2 - computer vision Video analysis library
libopencv-videoio-dev - development files for libopencv-videoio4.2
libopencv-videoio4.2 - computer vision Video I/O library
libopencv-videostab-dev - development files for libopencv-videostab4.2
libopencv-videostab4.2 - computer vision video stabilization library
libopencv-viz-dev - development files for libopencv-viz4.2
libopencv-viz4.2 - computer vision 3D data visualization library
libopencv4.2-java - Java bindings for the computer vision library
libopencv4.2-jni - Java jni library for the computer vision library

上面找到的這些包中，包名含有 "libopencv" 字樣，後綴爲 "-dev" 的包即爲我們需要安裝的 OpenCV 庫開發包。可以按需安裝這些包，也可以一次性安裝所有這些包。在終端中通過如下命令安裝所有 OpenCV 庫開發包（以 root 用戶）：

$ sudo apt install libopencv-apps-dev libopencv-calib3d-dev libopencv-contrib-dev libopencv-core-dev libopencv-dev libopencv-dnn-dev libopencv-features2d-dev libopencv-flann-dev libopencv-highgui-dev libopencv-imgcodecs-dev libopencv-imgproc-dev libopencv-ml-dev libopencv-objdetect-dev libopencv-photo-dev libopencv-shape-dev libopencv-stitching-dev libopencv-superres-dev libopencv-ts-dev libopencv-video-dev libopencv-videoio-dev libopencv-videostab-dev libopencv-viz-dev

這裏隨便找一個包，看下它安裝的文件及各個文件的安裝路徑：

$ dpkg -L libopencv-core-dev 
/.
/usr
/usr/include
/usr/include/opencv4
/usr/include/opencv4/opencv2
/usr/include/opencv4/opencv2/core
/usr/include/opencv4/opencv2/core/affine.hpp
/usr/include/opencv4/opencv2/core/async.hpp
/usr/include/opencv4/opencv2/core/base.hpp
/usr/include/opencv4/opencv2/core/bindings_utils.hpp
/usr/include/opencv4/opencv2/core/bufferpool.hpp
/usr/include/opencv4/opencv2/core/check.hpp
/usr/include/opencv4/opencv2/core/core.hpp
/usr/include/opencv4/opencv2/core/core_c.h
/usr/include/opencv4/opencv2/core/cuda
/usr/include/opencv4/opencv2/core/cuda/block.hpp
/usr/include/opencv4/opencv2/core/cuda/border_interpolate.hpp
/usr/include/opencv4/opencv2/core/cuda/color.hpp
/usr/include/opencv4/opencv2/core/cuda/common.hpp
/usr/include/opencv4/opencv2/core/cuda/datamov_utils.hpp
/usr/include/opencv4/opencv2/core/cuda/detail
/usr/include/opencv4/opencv2/core/cuda/detail/color_detail.hpp
/usr/include/opencv4/opencv2/core/cuda/detail/reduce.hpp
/usr/include/opencv4/opencv2/core/cuda/detail/reduce_key_val.hpp
/usr/include/opencv4/opencv2/core/cuda/detail/transform_detail.hpp
/usr/include/opencv4/opencv2/core/cuda/detail/type_traits_detail.hpp
/usr/include/opencv4/opencv2/core/cuda/detail/vec_distance_detail.hpp
/usr/include/opencv4/opencv2/core/cuda/dynamic_smem.hpp
/usr/include/opencv4/opencv2/core/cuda/emulation.hpp
/usr/include/opencv4/opencv2/core/cuda/filters.hpp
/usr/include/opencv4/opencv2/core/cuda/funcattrib.hpp
/usr/include/opencv4/opencv2/core/cuda/functional.hpp
/usr/include/opencv4/opencv2/core/cuda/limits.hpp
/usr/include/opencv4/opencv2/core/cuda/reduce.hpp
/usr/include/opencv4/opencv2/core/cuda/saturate_cast.hpp
/usr/include/opencv4/opencv2/core/cuda/scan.hpp
/usr/include/opencv4/opencv2/core/cuda/simd_functions.hpp
/usr/include/opencv4/opencv2/core/cuda/transform.hpp
/usr/include/opencv4/opencv2/core/cuda/type_traits.hpp
/usr/include/opencv4/opencv2/core/cuda/utility.hpp
/usr/include/opencv4/opencv2/core/cuda/vec_distance.hpp
/usr/include/opencv4/opencv2/core/cuda/vec_math.hpp
/usr/include/opencv4/opencv2/core/cuda/vec_traits.hpp
/usr/include/opencv4/opencv2/core/cuda/warp.hpp
/usr/include/opencv4/opencv2/core/cuda/warp_reduce.hpp
/usr/include/opencv4/opencv2/core/cuda/warp_shuffle.hpp
/usr/include/opencv4/opencv2/core/cuda.hpp
/usr/include/opencv4/opencv2/core/cuda.inl.hpp
/usr/include/opencv4/opencv2/core/cuda_stream_accessor.hpp
/usr/include/opencv4/opencv2/core/cuda_types.hpp
/usr/include/opencv4/opencv2/core/cv_cpu_dispatch.h
/usr/include/opencv4/opencv2/core/cv_cpu_helper.h
/usr/include/opencv4/opencv2/core/cvdef.h
/usr/include/opencv4/opencv2/core/cvstd.hpp
/usr/include/opencv4/opencv2/core/cvstd.inl.hpp
/usr/include/opencv4/opencv2/core/cvstd_wrapper.hpp
/usr/include/opencv4/opencv2/core/detail
/usr/include/opencv4/opencv2/core/detail/async_promise.hpp
/usr/include/opencv4/opencv2/core/detail/exception_ptr.hpp
/usr/include/opencv4/opencv2/core/directx.hpp
/usr/include/opencv4/opencv2/core/eigen.hpp
/usr/include/opencv4/opencv2/core/fast_math.hpp
/usr/include/opencv4/opencv2/core/hal
/usr/include/opencv4/opencv2/core/hal/hal.hpp
/usr/include/opencv4/opencv2/core/hal/interface.h
/usr/include/opencv4/opencv2/core/hal/intrin.hpp
/usr/include/opencv4/opencv2/core/hal/intrin_avx.hpp
/usr/include/opencv4/opencv2/core/hal/intrin_avx512.hpp
/usr/include/opencv4/opencv2/core/hal/intrin_cpp.hpp
/usr/include/opencv4/opencv2/core/hal/intrin_forward.hpp
/usr/include/opencv4/opencv2/core/hal/intrin_msa.hpp
/usr/include/opencv4/opencv2/core/hal/intrin_neon.hpp
/usr/include/opencv4/opencv2/core/hal/intrin_sse.hpp
/usr/include/opencv4/opencv2/core/hal/intrin_sse_em.hpp
/usr/include/opencv4/opencv2/core/hal/intrin_vsx.hpp
/usr/include/opencv4/opencv2/core/hal/intrin_wasm.hpp
/usr/include/opencv4/opencv2/core/hal/msa_macros.h
/usr/include/opencv4/opencv2/core/hal/simd_utils.impl.hpp
/usr/include/opencv4/opencv2/core/mat.hpp
/usr/include/opencv4/opencv2/core/mat.inl.hpp
/usr/include/opencv4/opencv2/core/matx.hpp
/usr/include/opencv4/opencv2/core/neon_utils.hpp
/usr/include/opencv4/opencv2/core/ocl.hpp
/usr/include/opencv4/opencv2/core/ocl_genbase.hpp
/usr/include/opencv4/opencv2/core/opencl
/usr/include/opencv4/opencv2/core/opencl/ocl_defs.hpp
/usr/include/opencv4/opencv2/core/opencl/opencl_info.hpp
/usr/include/opencv4/opencv2/core/opencl/opencl_svm.hpp
/usr/include/opencv4/opencv2/core/opencl/runtime
/usr/include/opencv4/opencv2/core/opencl/runtime/autogenerated
/usr/include/opencv4/opencv2/core/opencl/runtime/autogenerated/opencl_clamdblas.hpp
/usr/include/opencv4/opencv2/core/opencl/runtime/autogenerated/opencl_clamdfft.hpp
/usr/include/opencv4/opencv2/core/opencl/runtime/autogenerated/opencl_core.hpp
/usr/include/opencv4/opencv2/core/opencl/runtime/autogenerated/opencl_core_wrappers.hpp
/usr/include/opencv4/opencv2/core/opencl/runtime/autogenerated/opencl_gl.hpp
/usr/include/opencv4/opencv2/core/opencl/runtime/autogenerated/opencl_gl_wrappers.hpp
/usr/include/opencv4/opencv2/core/opencl/runtime/opencl_clamdblas.hpp
/usr/include/opencv4/opencv2/core/opencl/runtime/opencl_clamdfft.hpp
/usr/include/opencv4/opencv2/core/opencl/runtime/opencl_core.hpp
/usr/include/opencv4/opencv2/core/opencl/runtime/opencl_core_wrappers.hpp
/usr/include/opencv4/opencv2/core/opencl/runtime/opencl_gl.hpp
/usr/include/opencv4/opencv2/core/opencl/runtime/opencl_gl_wrappers.hpp
/usr/include/opencv4/opencv2/core/opencl/runtime/opencl_svm_20.hpp
/usr/include/opencv4/opencv2/core/opencl/runtime/opencl_svm_definitions.hpp
/usr/include/opencv4/opencv2/core/opencl/runtime/opencl_svm_hsa_extension.hpp
/usr/include/opencv4/opencv2/core/opengl.hpp
/usr/include/opencv4/opencv2/core/operations.hpp
/usr/include/opencv4/opencv2/core/optim.hpp
/usr/include/opencv4/opencv2/core/ovx.hpp
/usr/include/opencv4/opencv2/core/persistence.hpp
/usr/include/opencv4/opencv2/core/saturate.hpp
/usr/include/opencv4/opencv2/core/simd_intrinsics.hpp
/usr/include/opencv4/opencv2/core/softfloat.hpp
/usr/include/opencv4/opencv2/core/sse_utils.hpp
/usr/include/opencv4/opencv2/core/traits.hpp
/usr/include/opencv4/opencv2/core/types.hpp
/usr/include/opencv4/opencv2/core/types_c.h
/usr/include/opencv4/opencv2/core/utility.hpp
/usr/include/opencv4/opencv2/core/utils
/usr/include/opencv4/opencv2/core/utils/allocator_stats.hpp
/usr/include/opencv4/opencv2/core/utils/allocator_stats.impl.hpp
/usr/include/opencv4/opencv2/core/utils/filesystem.hpp
/usr/include/opencv4/opencv2/core/utils/instrumentation.hpp
/usr/include/opencv4/opencv2/core/utils/logger.defines.hpp
/usr/include/opencv4/opencv2/core/utils/logger.hpp
/usr/include/opencv4/opencv2/core/utils/logtag.hpp
/usr/include/opencv4/opencv2/core/utils/tls.hpp
/usr/include/opencv4/opencv2/core/utils/trace.hpp
/usr/include/opencv4/opencv2/core/va_intel.hpp
/usr/include/opencv4/opencv2/core/version.hpp
/usr/include/opencv4/opencv2/core/vsx_utils.hpp
/usr/include/opencv4/opencv2/core.hpp
/usr/include/opencv4/opencv2/cvconfig.h
/usr/include/opencv4/opencv2/opencv.hpp
/usr/include/opencv4/opencv2/opencv_modules.hpp
/usr/lib
/usr/lib/x86_64-linux-gnu
/usr/lib/x86_64-linux-gnu/libopencv_core.a
/usr/share
/usr/share/doc
/usr/share/doc/libopencv-core-dev
/usr/share/doc/libopencv-core-dev/README.Debian
/usr/share/doc/libopencv-core-dev/copyright
/usr/lib/x86_64-linux-gnu/libopencv_core.so
/usr/share/doc/libopencv-core-dev/changelog.Debian.gz

注意，頭文件是安裝在 /usr/include/opencv4/ 目錄下的，但這個目錄通常不是編譯工具默認的頭文件搜索目錄。包含頭文件時，要麼寫成 <opencv4/opencv2/opencv.hpp> 這樣，要麼寫成 <opencv2/opencv.hpp> 並定製頭文件的搜索路徑。

在 Eclipse 中構建使用 OpenCV（CDT 插件）的應用

首先要創建一個工程。具體過程如下：

啓動 Eclipse。在 Eclipse 的安裝目錄中運行可執行文件。
轉到 File -> New -> C/C++ Project

爲這個工程選擇一個名字，比如 DisplayImage。對於這個例子來說一個 Empty Project 應該就可以了。

其它配置保持默認值。點擊 Finish。
新建的工程（在這種情況下是 DisplayImage）應該出現在 Project Navigator（通常出現在窗口的左側）。

現在來給工程添加一個使用 OpenCV 的源文件：
- 鼠標右鍵點擊 DisplayImage （在 Navigator 中）。New -> Folder。

將目錄命名爲 src，然後點擊 Finish。
右鍵點擊新創建的目錄。選擇 New source file。
將其命名爲 DisplayImage.cpp。點擊 Finish。

現在我們有了一個包含一個空 .cpp 文件的工程。讓我們在這個源文件中添加我們的示例代碼（換句話說，把下面的代碼片段拷貝過去）：

#include <iostream>
using namespace std;

#include <opencv2/opencv.hpp>

int main(int argc, char **argv) {
  cv::Mat image;
  image = cv::imread(argv[1], 1);
  if (argc != 2 || !image.data) {
    printf("No image data \n");
    return -1;
  }
  cv::namedWindow("Display Image", cv::WINDOW_AUTOSIZE);
  imshow("Display Image", image);
  cv::waitKey(0);
  return 0;
}

我們只差最後一步了：告訴 Eclipse OpenCV 的頭文件和庫文件在哪兒。要做到這一點，執行如下操作：

轉到 Project–>Properties
在 C/C++ Build 中，點中 Settings。在右邊，選擇 Tool Settings 標籤。這裏我們進入頭文件和庫信息設置： a. 在 GCC C++ Compiler 中，切換到 Includes。在 Include paths(-l) 中，應該包含安裝 opencv 的頭文件的目錄的路徑。對於默認的安裝，一般來說是 /usr/include/opencv4，對於上面將 OpenCV 庫安裝到用戶本地目錄的情況，則是 /home/zhangsan/.local/include/opencv4。

b. 現在切換到 GCC C++ Linker，在這裏需要填兩個地方：首先是必須在 Library search path (-L) 寫入 opencv 庫所在的目錄的路徑，默認情況下這個路徑爲：

/usr/local/lib

這個路徑一般都是鏈接器默認的庫文件搜索路徑，因而一般不用填寫。在 Libraries(-l) 中添加需要的 OpenCV 庫。通常只需要添加下面列表中的前 4 個就夠了（對於簡單的應用程序來說）。這裏我們把所有計劃要使用的庫都加上。 opencv_core opencv_imgproc opencv_imgcodecs opencv_highgui opencv_ml opencv_videoio opencv_video opencv_features2d opencv_calib3d opencv_objdetect opencv_flann

現在點擊 OK。

這個工程應該已經準備好構建了。轉到 Project->Build Project 在終端中應該能看到類似這樣的內容：

檢查你的目錄的話，應該能夠看到可執行文件。

運行應用

現在，我們已經有可執行文件可以運行了。如果是把 OpenCV 安裝在默認的 /usr/local 目錄的話，則一般情況下 OpenCV 庫文件的安裝目錄已經在動態鏈接器的庫文件搜索路徑了。使用終端，我們可以像這樣做運行它：

$ cd ~/workspace/DisplayImage
$ ./Debug/DisplayImage ~/HappyLittleFish.jpg

如果我們將 OpenCV 安裝在了用戶本地目錄的話，像上面那樣執行可執行文件會報錯，找不到動態鏈接庫，如：

$ ./Debug/DisplayImage ~/HappyLittleFish.jpg
Debug/DisplayImage: error while loading shared libraries: libopencv_core.so.405: cannot open shared object file: No such file or directory

此時我們可以將 OpenCV 庫文件的安裝路徑添加進 /etc/ld.so.conf 並執行 sudo ldconfig 以更新動態鏈接器的庫文件搜索路徑；也可以設置 LD_LIBRARY_PATH 指向 OpenCV 庫文件的安裝路徑，然後再執行可執行文件，如：

$ export LD_LIBRARY_PATH=~/.local/lib
$ ./Debug/DisplayImage ~/HappyLittleFish.jpg

假設要使用作爲路徑參數的圖片位於 ~/HappyLittleFish.jpg。我們也可以在 Eclipse 中執行可執行文件：

轉到 Run->Run Configurations
在 C/C++ Application 下，將看到可執行文件的名字（如果沒有，則鼠標左鍵雙擊 C/C++ Application）。選擇該名字（在這個例子中是 DisplayImage）。
現在在窗口的右側，選擇 Arguments 標籤。把我們想打開的圖片的路徑（可以寫絕對路徑，也可以寫相對於 workspace/DisplayImage 目錄的相對路徑）寫進去。讓我們使用 ~/HappyLittleFish.jpg：

如果我們沒有把 OpenCV 安裝在默認位置，同樣需要設置環境變量 LD_LIBRARY_PATH 指向 OpenCV 庫文件的安裝路徑。同樣是在上面的窗口中，選中 Environment 標籤。添加環境變量 LD_LIBRARY_PATH，如：

點擊 Apply 按鈕，然後運行。這將彈出一個 OpenCV 窗口並顯示一張魚圖（或其它你想顯示的）：

現在已經完全做好了使用 Eclipse 開發 OpenCV 應用的準備了。

V2：在 Eclipse （包含 CDT 插件）中使用 CMake+OpenCV

比如已經在一個稱爲 helloopencv 的目錄下創建了一個名爲 helloworld.cpp 的新文件，其內容如下：

#include <opencv2/opencv.hpp>
using namespace cv;
int main ( int argc, char **argv )
{
  Mat img(480, 640, CV_8U);
  putText(img, "Hello World!", Point( 200, 400 ), FONT_HERSHEY_SIMPLEX | FONT_ITALIC, 1.0, Scalar( 255, 255, 0 ));
  imshow("My Window", img);
  waitKey();
  return 0;
}

在 helloopencv 目錄下放一個 CMakeLists.txt 文件，其內容如下：

cmake_minimum_required(VERSION 3.6)

PROJECT( helloworld_proj )
FIND_PACKAGE( OpenCV REQUIRED )
ADD_EXECUTABLE( helloworld helloworld.cpp )
TARGET_LINK_LIBRARIES(helloworld 
                      ${OpenCV_LIBS})

在 helloopencv 目錄下創建 build 目錄，並切換進去，執行 cmake 命令生成構建腳本：

$ cd helloopencv
$ mkdir build
$ cd build
$ cmake -DOpenCV_DIR=/home/zhangsan/.local/lib/cmake/opencv4 ..
-- Configuring done
-- Generating done
-- Build files have been written to: /home/zhangsan/workspace/helloopencv/build

這裏定義 OpenCV_DIR 指向 opencv 庫的 cmake 文件安裝路徑。

運行 make -j4（-j4 是可選的，它只是告訴編譯器用 4 個線程來編譯），或者 cmake --build .。確保它構建完成。
啓動 Eclipse。把 workspace 放在某些目錄下，但不能在 helloopencv 或 helloopencv\build 下。
轉到 File -> Import...。然後打開 C/C++ filter。選中 Existing Code as a Makefile Project。
命名這個工程，比如 helloworld。瀏覽到代碼目錄位置 helloopencv 目錄。在 "Toolchain for Indexer Settings" 中選擇 Linux GCC 並按下 Finish。
在 Project Explorer 部分選中工程，右鍵單擊。選擇 Properties。在 C/C++ Build 下，在 Builder Settings 標籤中，設置 Build directory:：從類似於 ${workspace_loc:/helloopencv} 改爲 ${workspace_loc:/helloopencv}/build，即執行 cmake 並生成 Makefile 文件的目錄。也可以選擇修改構建命令 Build command:，反選 Use default build command，然後在 Build command: 中填入自己需要的構建命令。在 Behavior 標籤中，選擇 Use custom build arguments，還可以輸入想要的構建參數，比如加上 VERBOSE=1 -j4 參數，告訴編譯器生成詳細的符號文件用於調試，並且以 4 個併發線程來編譯。
轉到 Project->Build Project 構建工程。

在命令行及 Eclipse 中運行編譯出來的二進制可執行的方法如上文所述。

通過 GCC 和 CMake 使用 OpenCV

在自己的代碼中使用 OpenCV 最簡單的方式就是使用 CMake 了。這有這樣一些優勢（摘自 Wiki）：

當在 Linux 和 Windows 之間移植時，不需要修改任何東西。
可以通過 CMake 簡單地與其它工具結合使用（比如 Qt，ITK 和 VTK）

如果你還不熟悉 CMake，可以參考它的網站上的教程。

步驟

創建一個使用 OpenCV 的程序

隨便使用任何一個你所熟悉的喜歡的編輯器，或者集成開發環境，創建一個類似於名爲 DisplayImage.cpp 的源文件，並輸入如下這樣簡單程序的源碼：

#include <stdio.h>
#include <opencv2/opencv.hpp>
using namespace cv;
int main(int argc, char **argv) {
  if (argc != 2) {
    printf("usage: DisplayImage.out <Image_Path>\n");
    return -1;
  }
  Mat image;
  image = imread(argv[1], 1);
  if (!image.data) {
    printf("No image data \n");
    return -1;
  }
  namedWindow("Display Image", WINDOW_AUTOSIZE);
  imshow("Display Image", image);
  waitKey(0);
  return 0;
}

把這個源文件放在某個目錄下，如 DisplayImage/src。

創建一個 CMake 文件

然後必須創建你的 CMakeLists.txt，比如在 DisplayImage 目錄下。它看起來像這樣：

cmake_minimum_required(VERSION 3.8)

project(DisplayImage)

find_package(OpenCV REQUIRED)

include_directories(${OpenCV_INCLUDE_DIRS})

add_executable(DisplayImage 
               src/DisplayImage.cpp)

target_link_libraries(DisplayImage ${OpenCV_LIBS})

生成可執行文件

這個部分很簡單，就像在任何其它使用 CMake 的項目裏那樣處理：

$ cd DisplayImage
$ cmake -Bbuild -DOpenCV_DIR=/home/zhangsan/.local/lib/cmake/opencv4 .
$ cmake --build build

CMake 的 -B 選項用於指定生成構建腳本的目錄。

結果

現在應該有了可執行文件（在這個例子中稱爲 DisplayImage）。只需要給它傳入一個圖片文件的路徑作爲參數來運行它，比如：

$ build/DisplayImage ..opencv/samples/data/lena.jpg

這應該能得到一個漂亮的窗口，就像下面展示的這個：

參考文檔

OpenCV Releases OpenCV 4.5.5 文檔 OpenCV-Python Tutorials OpenCV-Python Tutorials - Introduction to OpenCV OpenCV-Python Tutorials - Introduction to OpenCV-Python Tutorials OpenCV-Python Tutorials - Install OpenCV-Python in Ubuntu

OpenCV Tutorials OpenCV Tutorials - Introduction to OpenCV OpenCV Tutorials - Installation in Linux OpenCV Tutorials - Using OpenCV with Eclipse (plugin CDT) OpenCV Tutorials - Using OpenCV with gcc and CMake

opencv_contrib

Done.

OpenCV_001-在 Ubuntu 上搭建 OpenCV 開發環境