Linux 编译安装 OpenCV (Ubuntu, Redhat)

文章目录

1. 1. 我的系统配置和安装版本
2. 2. 编译环境准备
3. 3. 下载OpenCV安装包
4. 4. 编译安装
5. 5. 环境变量修改
6. 6. 测试

今天在编译安装opencv时遇到的一些问题，记录下来。

我的系统配置和安装版本

ubuntu server 12.04 和 redhat 6.4 两台机器
安装opencv 2.4.8

编译环境准备

对于ubuntu的机器，首先安装以下软件：

# 安装基本编程环境
sudo apt-get install build-essential  
# 安装编解码支持库
sudo apt-get install  libgtk2.0-dev libavcodec-dev libavformat-dev  libtiff4-dev  libswscale-dev libjasper-dev  
# 安装cmake
sudo apt-get install cmake 
# 安装pkg-config
sudo apt-get install pkg-config

对于redhat的机器，使用yum安装以下包：

yum install cmake gcc gcc-c++ gtk+-devel gimp-devel gimp-devel-tools gimp-help-browser zlib-devel libtiff-devel libjpeg-devel libpng-devel gstreamer-devel libavc1394-devel libraw1394-devel libdc1394-devel jasper-devel jasper-utils swig python libtool nasm

下载OpenCV安装包

到官网下载OpenCV的安装包，我下载的是2.4.8的版本，但这个版本我安装的时候有bug，后面会介绍。（建议下载2.4.11及以上版本，安装步骤基本类似）

下载后解压，进入目录

unzip -v opencv-2.4.8.zip
#进入解压后的目录
cd opencv-2.4.8

编译安装

直接在当前目录下编译即可，并不一定要建立子目录

cmake -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=/usr/local .

make

make -j8    #（-j8表示使用8个线程编译，可以大大加快编译速度）
sudo make install

我主要在此处遇到了问题，因为我的电脑安装了GPU卡，因此OpenCV会自动编译CUDA相关的代码，可以直接关掉( 在cmake时加上-D WITH_CUDA=OFF )，但我需要编译。因此默认是打开的，因此在make时遇到以下错误：a storage class is not allowed in an explicit specialization

/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(51): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(52): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(53): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(54): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(55): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(56): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(57): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(58): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(61): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(62): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(63): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(64): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(65): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(66): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(67): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(68): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(119): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(120): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(121): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(122): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(123): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(124): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(125): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(126): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(127): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(128): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(129): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(130): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(131): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(132): error: a storage class is not allowed in an explicit specialization
/root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp(133): error: a storage class is not allowed in an explicit specialization
31 errors detected in the compilation of "/tmp/tmpxft_00006af4_00000000-9_NCVPyramid.compute_35.cpp1.ii".
CMake Error at cuda_compile_generated_NCVPyramid.cu.o.cmake:266 (message):
  Error generating file
  /root/opencv-2.4.8/modules/gpu/CMakeFiles/cuda_compile.dir/src/nvidia/core/./cuda_compile_generated_NCVPyramid.cu.o

make[2]: *** [modules/gpu/CMakeFiles/cuda_compile.dir/src/nvidia/core/./cuda_compile_generated_NCVPyramid.cu.o] Error 1
make[2]: *** Waiting for unfinished jobs....
make[1]: *** [modules/gpu/CMakeFiles/opencv_gpu.dir/all] Error 2
make: *** [all] Error 2

我安装的CUDA是6.5版本，计算能力是3.5，针对以上错误，我找到的说在cmake时加上以上参数的方法都是没用：

-D CUDA_GENERATION=Auto
或
-D CUDA_GENERATION="" -D CUDA_ARCH_BIN=2.0（或3.5，5.0）

最后找的解决办法是：
打开错误的文件：

# 注意修改为你的编译路径
vim /root/opencv-2.4.8/modules/gpu/src/nvidia/core/NCVPixelOperations.hpp

将其中template声明中的所有static全部删除。可以使用vim的全局替换，或者其他编辑工具全局替换即可(template<> static inline 替换为template<> inline )：

:%s/template<> static inline/template<> inline/g

这个问题的原因是OpenCV 2.4.8 与CUDA 6.5不兼容的一个bug（好像2.4系列都有问题），已在后续版本中修复，参考http://code.opencv.org/issues/3814

修改过后重新执行以下，切记有多核的一定使用多线程，编译CUDA慢的要哭：

cmake -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=/usr/local .
make -j8    #（-j8表示使用8个线程编译，可以大大加快编译速度）
sudo make install

环境变量修改

在/etc/ld.so.conf 文件中添加 /usr/local/lib ， 然后执行

sudo ldconfig -v

为程序指定openvc的头文件位置, 使用pkg-config命令来完成。首先在 /etc/profile 中添加

export  PKG_CONFIG_PATH=$PKG_CONFIG_PATH:/usr/local/lib/pkgconfig

然后source生效

source /etc/profile

再执行以下命令可以打印opencv的配置信息

pkg-config --libs opencv

root@ubuntu:~# pkg-config  --libs opencv
/usr/local/lib/libopencv_calib3d.so /usr/local/lib/libopencv_contrib.so /usr/local/lib/libopencv_core.so /usr/local/lib/libopencv_features2d.so /usr/local/lib/libopencv_flann.so /usr/local/lib/libopencv_gpu.so /usr/local/lib/libopencv_highgui.so /usr/local/lib/libopencv_imgproc.so /usr/local/lib/libopencv_legacy.so /usr/local/lib/libopencv_ml.so /usr/local/lib/libopencv_nonfree.so /usr/local/lib/libopencv_objdetect.so /usr/local/lib/libopencv_ocl.so /usr/local/lib/libopencv_photo.so /usr/local/lib/libopencv_stitching.so /usr/local/lib/libopencv_superres.so /usr/local/lib/libopencv_ts.a /usr/local/lib/libopencv_video.so /usr/local/lib/libopencv_videostab.so /usr/local/cuda/lib64/libcufft.so /usr/local/cuda/lib64/libnpps.so /usr/local/cuda/lib64/libnppi.so /usr/local/cuda/lib64/libnppc.so /usr/local/cuda/lib64/libcudart.so -lrt -lpthread -lm -ldl

测试

在解压的opencv目录下找到 samples/c/build_all.sh，运行该文件

cd /root/opencv-2.4.8/samples/c/
chmod a+x build_all.sh
./build_all.sh

选取生成的可执行文件进行测试，比如find_obg：

root@ubuntu:~/opencv-2.4.8/samples/c# ls
adaptiveskindetector      delaunay               mser_sample
adaptiveskindetector.cpp  delaunay.c             mser_sample.cpp
agaricus-lepiota.data     example_cmake          mushroom
airplane.jpg              facedetect             mushroom.cpp
baboon200.jpg             facedetect.cmd         one_way_sample
baboon200_rotated.jpg     facedetect.cpp         one_way_sample.cpp
baboon.jpg                fback_c                one_way_train_0000.jpg
bgfg_codebook             fback_c.c              one_way_train_0001.jpg
bgfg_codebook.cpp         find_obj               one_way_train_images.txt
blobtrack_sample          find_obj_calonder      polar_transforms
blobtrack_sample.cpp      find_obj_calonder.cpp  polar_transforms.c
box_in_scene.png          find_obj.cpp           puzzle.png
box.png                   find_obj_ferns         pyramid_segmentation
build_all.sh              find_obj_ferns.cpp     pyramid_segmentation.c
calonder_params.xml       fruits.jpg             scene_l.bmp
cat.jpg                   intrinsics.yml         scene_r.bmp
cat.xml                   JCB.png                smiledetect
CMakeLists.txt            latentsvmdetect        smiledetect.cpp
contours                  latentsvmdetect.cpp    stuff.jpg
contours.c                lena.jpg               tree.avi
convert_cascade           morphology             tree_engine
convert_cascade.c         morphology.c           tree_engine.cpp
cvsample.dsp              motempl                waveform.data
cvsample.vs2005.vcproj    motempl.c

./find_obj

安装结束。

---

转载请标明文章出处。本文内容为实践后的原创整理，如果侵犯了您的版权，请联系我进行删除，邮件：yanhealth@163.com

参考资料：
[1] http://docs.opencv.org/2.4/doc/tutorials/introduction/linux_install/linux_install.html#linux-installation
[2] http://blog.csdn.net/surgewong/article/details/39078251
[3] http://blog.csdn.net/xuejiren/article/details/24347555