这篇文章将说明在 Ubuntu Linux 系统上搭建 OpenCV 开发环境的过程,以支持 Python 和 C++ 编程语言开发基于 OpenCV 的应用,或学习 OpenCV。本文说明的过程在 Ubuntu 20.04 版本的系统上经过测试验证。
在 Ubuntu 系统上安装 OpenCV-Python
在 Ubuntu 系统中安装 OpenCV-Python 有两种方式:
- 从 Ubuntu 仓库安装可用的预编译二进制文件
- 由源码编译。
这里这两种方式都会介绍。
另一个非常重要的问题是必须的附加库。OpenCV-Python 只需要 Numpy。但在这个教程中,我们还使用 Matplotlib 进行一些简单而漂亮的绘图。Matplotlib 是可选的,但强烈建议安装它。类似地,我们也将看到 IPython,一个交互式 Python 终端,同样强烈建议安装它。
从预编译二进制文件安装 OpenCV-Python
在仅仅想要编程和开发 OpenCV 应用程序时,这个方法最好。
在终端中通过如下命令安装 python3-opencv 包(以 root 用户):
$ sudo apt-get install python3-opencv
正在读取软件包列表... 完成
正在分析软件包的依赖关系树
正在读取状态信息... 完成
下列软件包是自动安装的并且现在不需要了:
ibus-data python3-ibus-1.0
使用'sudo apt autoremove'来卸载它(它们)。
将会同时安装下列软件:
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
建议安装:
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】软件包将被安装:
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
升级了 0 个软件包,新安装了 105 个软件包,要卸载 0 个软件包,有 70 个软件包未被升级。
需要下载 104 MB 的归档。
解压缩后会消耗 397 MB 的额外空间。
您希望继续执行吗? [Y/n] y
获取:1 http://cn.archive.ubuntu.com/ubuntu focal/universe amd64 gcc-8-base amd64 8.4.0-3ubuntu2 [18.7 kB]
. . . . . .
它会安装非常多在 Python 中调用 OpenCV 接口所需的依赖,如 libopencv-core4.2
、libopencv-superres4.2
、libopencv-objdetect4.2
和 libopencv-features2d4.2
等。注意,安装的这些 OpenCV 依赖库的包只包含了动态链接库二进制文件,而没有 C++ OpenCV 开发所需要的头文件等内容。
打开一个 Python IDLE (或者 IPython)并在 Python 终端中输入如下代码:
import cv2 as cv
print(cv.__version__)
执行这段代码,输出了适当的 OpenCV 版本号而没有任何报错,则恭喜你!!!你已经成功地安装了 OpenCV-Python。
这很简单。但这有一个问题。Apt 仓库可能总是不包含最新版本的 OpenCV。比如笔者在 Ubuntu 20.04 上通过上面的操作安装 OpenCV,得到的版本为 4.2.0,最新的版本则为 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
对于 Python API,最新版本将始终包含更好的支持和最新的错误修复。
因此,获取最新源代码的下一个方法,即从源代码编译。同时,如果在某个时间点,你想为 OpenCV 贡献代码,你也将需要这个。
从源码构建
乍一看去,从源码编译可能有点复杂,但一旦你成功了,你会发现其实也没什么复杂的。
首先我们将安装一些依赖。有些是必须的,有些是可选的。如果你不需要,可以跳过可选的依赖。
必须的构建依赖
我们需要 CMake 来配置安装,需要 GCC 来编译。需要 Python-devel 和 Numpy 来构建 Python 绑定等。
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 来支持 GUI 功能,Camera 支持 (v4l),媒体支持 (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
可选的依赖
对于在 Ubuntu 机器上安装 OpenCV,上面的依赖已经足够了。但根据你的需求,你可能需要一些额外的依赖。这样的额外依赖列表如下。你可以根据你的需求把它们留在那里或安装它们。
OpenCV 附带了对图像格式的支持,如 PNG、JPEG、JPEG2000、TIFF、WebP 等。但它可能有点老。如果你想获得最新的库,你可以为这些格式的系统库安装开发文件。
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,也可以安装
libjasper-dev
来给你的系统添加对 JPEG2000 格式的系统级支持。
下载 OpenCV
获得 OpenCV 的源码有两种方式:
- 从 OpenCV 的 Github 仓库 下载最新的源码。(如果你想要给 OpenCV 贡献代码也可以通过它。要下载源码,首先需要安装 Git。)
$ sudo apt-get install git
$ git clone https://github.com/opencv/opencv.git
这将在当前目录创建一个名为 “opencv” 的目录。代码克隆过程可能需要一些时间,这依赖于你的网络连接情况。
- 除了直接克隆 OpenCV 的 Github 仓库 之外,还有另外一种获取 OpenCV 源码的方法。OpenCV 官方提供了已经打包好的发行版源码的压缩包,可以下载这些压缩包获取源码,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
解压之后并重命名源码目录,OpenCV 源码将位于 opencv 目录下。
现在打开一个终端窗口,切换到下载的 “opencv” 目录。创建一个名为 “build” 的目录并切换到它下面。
$ mkdir build
$ cd build
当然也可以在 OpenCV 源码目录的同级目录中创建 “build” 目录,这同样需要切换到新建的目录下。注意,后面执行 CMake 配置构建时,需要正确引用 OpenCV 源码目录。
配置和安装
现在我们已经拥有了所有必须的依赖,让我们开始安装 OpenCV 吧。安装必须通过 CMake 来配置。它指定了将要安装哪些模块、安装路径、将使用哪些附加库、是否要编译文档和示例等。这些工作中的大部分将通过默认参数的良好配置自动完成。
以下命令通常用于配置 OpenCV 库构建(在 “build” 目录中执行):
$ cmake ../
OpenCV 默认假设构建类型为 “Release”,安装路径为 “/usr/local”。关于 CMake 选项的其它信息请参考 OpenCV C++ 编译指南。
你应该能够在你的 CMake 输出中看到如下这些行(它们意味着已经发现了适当的 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
命令的参数是包含 CMakelists.txt 文件的目录的路径,即源码目录。如果 “build” 目录在 “opencv” 源码目录的同级目录中创建,则为(在 “build” 目录中执行):
# Configure
$ cmake ../opencv
现在可以使用 “make” 命令来构建文件:
$ 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
构建也可以通过 cmake
命令完成:
$ cmake --build .
成功构建之后,可以在 build/lib
目录下找到库文件,并在 build/bin
目录下找到可执行文件(测试,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
编译出来的可执行文件和库文件非常多。
CMake 包文件位于构建根目录:
build$ ls OpenCVConfig*.cmake
OpenCVConfig.cmake OpenCVConfig-version.cmake
build$ ls OpenCVModules.cmake
OpenCVModules.cmake
它们用于支持 CMake 的 find_library。
默认情况,OpenCV 将安装在 /usr/local
目录下,所有文件将被拷贝如下位置:
/usr/local/bin
- 可执行文件/usr/local/lib
- 库文件 (.so)/usr/local/cmake/opencv4
- cmake 包/usr/local/include/opencv4
- 头文件/usr/local/share/opencv4
- 其它文件(比如 XML 格式的训练的级联)
由于 /usr/local
为 root 用户所有,因而安装应该以超级用户特权执行 (sudo
),使用 “make install” 命令来安装它:
$ sudo make install
安装结束。所有的文件都被安装在了 “/usr/local/” 目录下。打开一个终端并尝试导入 “cv2”:
import cv2 as cv
print(cv.__version__)
构建配置和安装过程的进一步说明
这一部分将提供更多构建过程的细节说明,并描述了其它可选的方法和工具。请参考 OpenCV 安装概述 了解一般的安装细节说明,并参考 OpenCV 配置选项参考 了解配置选项文档。
编译器和构建工具
要编译 OpenCV,我们就需要 C++ 编译器。通常是 G++/GCC 或 Clang/LLVM。上面我们看到了 G++/GCC 的安装方法,这里给出安装 Clang/LLVM 的命令:
$ sudo apt install -y clang
无论何种编译工具,构建 OpenCV 都需要用 CMake 来执行构建配置。
CMake 可以为不同的构建系统生成构建脚本,除了上面看到的 make,在 Ubuntu 上还可以用 ninja 来构建。Ninja 的安装命令如下:
$ sudo apt install -y ninja-build
配置和构建
CMake 默认为 make 构建系统生成构建脚本,如上面那样。但我们可以通过 CMake 命令行参数强制为 ninja 构建系统生成构建脚本,如:
$ cmake -GNinja ../opencv
再提一点关于安装目标目录配置的问题。
安装过程仅将文件复制到预定义的位置并进行少量修补。使用这种方法,无论是通过 make,还是通过 ninja,安装都不会将 opencv 集成到系统包注册表中,如无法通过 apt 来管理这些安装。因此,也无法自动卸载 opencv。一般来说,最好不要进行系统范围的安装,因为可能与系统包冲突。
安装根目录可以通过 CMAKE_INSTALL_PREFIX
配置参数改变,比如 -DCMAKE_INSTALL_PREFIX=$HOME/.local
指明安装到当前用户的本地目录。安装布局可以通过 OPENCV_*_INSTALL_PATH
参数修改。详情请参考 OpenCV 配置选项参考。
另外,OpenCV 有一些所谓的 “extra” 模块,它们也提供了一些功能,源码位于单独的 git 仓库内。新模块常常不具有稳定的 API,且还没经过良好地测试。由于 OpenCV 库想要提供良好的二进制兼容性,并尝试提供良好的性能和稳定性,这样,新模块不应该作为正式的 OpenCV 分发版的一部分发型。
所有的新模块单独开发,并首先发布到 opencv_contrib
仓库中。随后模块变得成熟并流行时,再被移入 OpenCV 主仓库中,开发团队再为这个模块提供产品级质量的支持。在构建时,想要一起编译 opencv_contrib
的话,首先需要下载 opencv_contrib
的源码。方法同样有两个,一是下载官方发布的源码包:
$ wget -O opencv_contrib.zip https://github.com/opencv/opencv_contrib/archive/4.x.zip
$ unzip opencv_contrib.zip
二是克隆 Github 仓库:
$ git clone https://github.com/opencv/opencv_contrib.git
假设 opencv_contrib
的源码被下载到了 opencv
源码的同级目录中。
在通过 CMake 执行构建配置时,通过 -DOPENCV_EXTRA_MODULES_PATH
参数选项指定 opencv_contrib
的模块的路径。
这样,配置通过 ninja 构建,连同 opencv_contrib
一起构建,并将 OpenCV 安装在用户本地目录的命令如下:
$ cmake -GNinja -DCMAKE_INSTALL_PREFIX=$HOME/.local -DOPENCV_EXTRA_MODULES_PATH=../opencv_contrib-4.x/modules ../opencv
对于构建,在使用 make 构建系统时,可以给 make 传参数,以启用多线程构建,如:
$ make -j4
运行 ninja 执行构建的命令如下:
$ ninja
运行 ninja 执行安装的命令如下:
$ 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"
由于指定的安装路径为用户本地路径,因而不需要 sudo
。在安装过程的输出中,还是能发现不少信息,如 Python 绑定的安装路径,头文件的具体安装路径等。
在 Ubuntu 系统上搭建 OpenCV 的 C++ 开发环境
这里说明配置 CMake 和 Eclipse 构建基于 OpenCV 的 C++ 项目的方法。
准备依赖
构建依赖可以分为两部分:
-
(1). 开发和构建工具,包括 Eclipse,CMake,GCC/G++,Clang/LLVM,Git 等。如果还没有安装 Eclipse(含 CDT),可以到 Eclipse 的网站上,下载 Eclipse IDE for C/C++ Developers。根据自己的系统平台类型选择下载的版本。其它工具可以参考上面的说明进行安装,这里不再赘述。
-
(2). 用于 C/C++ 开发的 OpenCV 库开发包,主要包括动态链接库,静态链接库和头文件。
对于 OpenCV 库开发包安装,有两种方法:
-
从源码构建安装。如上面从源码构建安装 OpenCV-Python 的过程,该过程除了可以构建安装 OpenCV 的 Python 绑定之外,也同样会构建安装C/C++ 开发所需要的所有库文件、头文件等内容。对于这种方法,这里不再赘述。
-
从 Apt 仓库安装预编译的二进制安装包。同样,用这种方式安装的 OpenCV 库开发包版本可能比较老。但这种方法更简单,对于只想利用 OpenCV 的 API 开发应用,而无心研究 OpenCV 的 源码或向 OpenCV 项目贡献源码的同学,这种方法比较方便。
对于从从 Apt 仓库安装预编译的二进制安装包,搜索 Apt 仓库中的 OpenCV 开发包,可以看到如下内容:
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
上面找到的这些包中,包名含有 “libopencv” 字样,后缀为 “-dev” 的包即为我们需要安装的 OpenCV 库开发包。可以按需安装这些包,也可以一次性安装所有这些包。在终端中通过如下命令安装所有 OpenCV 库开发包(以 root 用户):
$ 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
这里随便找一个包,看下它安装的文件及各个文件的安装路径:
$ 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,头文件是安装在 /usr/include/opencv4/
目录下的,但这个目录通常不是编译工具默认的头文件搜索目录。包含头文件时,要么写成 <opencv4/opencv2/opencv.hpp>
这样,要么写成 <opencv2/opencv.hpp>
并定制头文件的搜索路径。
在 Eclipse 中构建使用 OpenCV(CDT 插件)的应用
首先要创建一个工程。具体过程如下:
-
启动 Eclipse。在 Eclipse 的安装目录中运行可执行文件。
-
转到 File -> New -> C/C++ Project
- 为这个工程选择一个名字,比如 DisplayImage。对于这个例子来说一个 Empty Project 应该就可以了。
-
其它配置保持默认值。点击 Finish。
-
新建的工程(在这种情况下是 DisplayImage)应该出现在 Project Navigator(通常出现在窗口的左侧)。
- 现在来给工程添加一个使用 OpenCV 的源文件:
- 鼠标右键点击 DisplayImage (在 Navigator 中)。New -> Folder。
-
将目录命名为 src,然后点击 Finish。
-
右键点击新创建的 目录。选择 New source file。
-
将其命名为 DisplayImage.cpp。点击 Finish。
- 现在我们有了一个包含一个空 .cpp 文件的工程。让我们在这个源文件中添加我们的示例代码(换句话说,把下面的代码片段拷贝过去):
#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;
}
- 我们只差最后一步了:告诉 Eclipse OpenCV 的头文件和库文件在哪儿。要做到这一点,执行如下操作:
-
转到 Project–>Properties
-
在 C/C++ Build 中,点中 Settings。在右边,选择 Tool Settings 标签。这里我们进入头文件和库信息设置:
a. 在 GCC C++ Compiler 中,切换到 Includes。在 Include paths(-l) 中,应该包含安装 opencv 的头文件的目录的路径。对于默认的安装,一般来说是 /usr/include/opencv4,对于上面将 OpenCV 库安装到用户本地目录的情况,则是 /home/zhangsan/.local/include/opencv4。
b. 现在切换到 GCC C++ Linker,在这里需要填两个地方:
首先是必须在 Library search path (-L) 写入 opencv 库所在的目录的路径,默认情况下这个路径为:
/usr/local/lib
这个路径一般都是链接器默认的库文件搜索路径,因而一般不用填写。
在 Libraries(-l) 中添加需要的 OpenCV 库。通常只需要添加下面列表中的前 4 个就够了(对于简单的应用程序来说)。这里我们把所有计划要使用的库都加上。
opencv_core opencv_imgproc opencv_imgcodecs opencv_highgui opencv_ml opencv_videoio opencv_video opencv_features2d opencv_calib3d opencv_objdetect opencv_flann
现在点击 OK。
- 这个工程应该已经准备好构建了。转到 Project->Build Project
在终端中应该能看到类似这样的内容:
检查你的目录的话,应该能够看到可执行文件。
运行应用
现在,我们已经有可执行文件可以运行了。如果是把 OpenCV 安装在默认的 /usr/local
目录的话,则一般情况下 OpenCV 库文件的安装目录已经在动态链接器的库文件搜索路径了。使用终端,我们可以像这样做运行它:
$ cd ~/workspace/DisplayImage
$ ./Debug/DisplayImage ~/HappyLittleFish.jpg
如果我们将 OpenCV 安装在了用户本地目录的话,像上面那样执行可执行文件会报错,找不到动态链接库,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
此时我们可以将 OpenCV 库文件的安装路径添加进 /etc/ld.so.conf
并执行 sudo ldconfig
以更新动态链接器的库文件搜索路径;也可以设置 LD_LIBRARY_PATH
指向 OpenCV 库文件的安装路径,然后再执行可执行文件,如:
$ export LD_LIBRARY_PATH=~/.local/lib
$ ./Debug/DisplayImage ~/HappyLittleFish.jpg
假设要使用作为路径参数的图片位于 ~/HappyLittleFish.jpg
。我们也可以在 Eclipse 中执行可执行文件:
-
转到 Run->Run Configurations
-
在 C/C++ Application 下,将看到可执行文件的名字(如果没有,则鼠标左键双击 C/C++ Application)。选择该名字(在这个例子中是 DisplayImage)。
-
现在在窗口的右侧,选择 Arguments 标签。把我们想打开的图片的路径(可以写绝对路径,也可以写相对于 workspace/DisplayImage 目录的相对路径)写进去。让我们使用 ~/HappyLittleFish.jpg:
- 如果我们没有把 OpenCV 安装在默认位置,同样需要设置环境变量
LD_LIBRARY_PATH
指向 OpenCV 库文件的安装路径。同样是在上面的窗口中,选中 Environment 标签。添加环境变量LD_LIBRARY_PATH
,如:
- 点击 Apply 按钮,然后运行。这将弹出一个 OpenCV 窗口并显示一张鱼图(或其它你想显示的):
- 现在已经完全做好了使用 Eclipse 开发 OpenCV 应用的准备了。
V2:在 Eclipse (包含 CDT 插件)中使用 CMake+OpenCV
比如已经在一个称为 helloopencv 的目录下创建了一个名为 helloworld.cpp 的新文件,其内容如下:
#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 目录下放一个
CMakeLists.txt
文件,其内容如下:
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 目录下创建 build 目录,并切换进去,执行
cmake
命令生成构建脚本:
$ 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
这里定义 OpenCV_DIR 指向 opencv 库的 cmake 文件安装路径。
-
运行
make -j4
(-j4 是可选的,它只是告诉编译器用 4 个线程来编译),或者cmake --build .
。确保它构建完成。 -
启动 Eclipse。把 workspace 放在某些目录下,但 不能 在
helloopencv
或helloopencv\build
下。 -
转到 File -> Import…。然后打开 C/C++ filter。选中 Existing Code as a Makefile Project。
-
命名这个工程,比如 helloworld。浏览到代码目录位置 helloopencv 目录。在 “Toolchain for Indexer Settings” 中选择 Linux GCC 并按下 Finish。
-
在 Project Explorer 部分选中工程,右键单击。选择 Properties。在 C/C++ Build 下,在 Builder Settings 标签中,设置 Build directory::从类似于
${workspace_loc:/helloopencv}
改为${workspace_loc:/helloopencv}/build
,即执行 cmake 并生成 Makefile 文件的目录。
也可以选择修改构建命令 Build command:,反选 Use default build command,然后在 Build command: 中填入自己需要的构建命令。在 Behavior 标签中,选择 Use custom build arguments,还可以输入想要的构建参数,比如加上VERBOSE=1 -j4
参数,告诉编译器生成详细的符号文件用于调试,并且以 4 个并发线程来编译。 -
转到 Project->Build Project 构建工程。
在命令行及 Eclipse 中运行编译出来的二进制可执行的方法如上文所述。
通过 GCC 和 CMake 使用 OpenCV
在自己的代码中使用 OpenCV 最简单的方式就是使用 CMake 了。这有这样一些优势(摘自 Wiki):
- 当在 Linux 和 Windows 之间移植时,不需要修改任何东西。
- 可以通过 CMake 简单地与其它工具结合使用(比如 Qt,ITK 和 VTK)
如果你还不熟悉 CMake,可以参考它的网站上的 教程。
步骤
创建一个使用 OpenCV 的程序
随便使用任何一个你所熟悉的喜欢的编辑器,或者集成开发环境,创建一个类似于名为 DisplayImage.cpp 的源文件,并输入如下这样简单程序的源码:
#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;
}
把这个源文件放在某个目录下,如 DisplayImage/src
。
创建一个 CMake 文件
然后必须创建你的 CMakeLists.txt,比如在 DisplayImage
目录下。它看起来像这样:
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})
生成可执行文件
这个部分很简单,就像在任何其它使用 CMake 的项目里那样处理:
$ cd DisplayImage
$ cmake -Bbuild -DOpenCV_DIR=/home/zhangsan/.local/lib/cmake/opencv4 .
$ cmake --build build
CMake 的 -B
选项用于指定生成构建脚本的目录。
结果
现在应该有了可执行文件(在这个例子中称为 DisplayImage)。只需要给它传入一个图片文件的路径作为参数来运行它,比如:
$ build/DisplayImage ..opencv/samples/data/lena.jpg
这应该能得到一个漂亮的窗口,就像下面展示的这个:
参考文档
OpenCV Releases
OpenCV 4.5.5 文档
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
Done.