ubuntu2004安装opencv，支持python接口

「爱情、让人受尽委屈。」 2022-12-27 08:48 270阅读 0赞

opencv下载

[https://github.com/opencv][https_github.com_opencv]

使用cmake-gui配置

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3pkZjUxMTM5MTU2NQ_size_16_color_FFFFFF_t_70][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3pkZjUxMTM5MTU2NQ_size_16_color_FFFFFF_t_70 1][]

需要安装numpy

sudo apt-get install python-numpy

另外一个就是，如果编译时使用了编译选项：  
\-D BUILD\_opencv\_world=NO 那么最后在build下面的lib文件夹下也找不到cv2.so的

cmake -D CMAKE\_BUILD\_TYPE=RELEASE -D CMAKE\_INSTALL\_PREFIX=/usr/local -D INSTALL\_PYTHON\_EXAMPLES=ON -D INSTALL\_C\_EXAMPLES=OFF -D PYTHON\_EXECUTABLE=/usr/bin/python3.7m -D PYTHON\_INCLUDE\_DIR=/usr/include/python3.7m -D PYTHON\_LIBRARY=/usr/bin/libpython3.7m.so -D PYTHON3\_PACKAGES\_PATH=/usr/lib/python3.7/site-packages -D BUILD\_EXAMPLES=ON …

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3pkZjUxMTM5MTU2NQ_size_16_color_FFFFFF_t_70 2][]

CMakeCache.txt节选:

//Value Computed by CMake  
OpenCV\_BINARY\_DIR:STATIC=/home/andy/zdfwork/openCV/pc/opencv-3.4.12/build

//The directory containing a CMake configuration file for OpenCV\_HAL.  
OpenCV\_HAL\_DIR:PATH=OpenCV\_HAL\_DIR-NOTFOUND

//Value Computed by CMake  
OpenCV\_SOURCE\_DIR:STATIC=/home/andy/zdfwork/openCV/pc/opencv-3.4.12

//pkg-config executable  
PKG\_CONFIG\_EXECUTABLE:FILEPATH=/usr/bin/pkg-config

//Path to a library.  
PNG\_LIBRARY\_DEBUG:FILEPATH=PNG\_LIBRARY\_DEBUG-NOTFOUND

//Path to a library.  
PNG\_LIBRARY\_RELEASE:FILEPATH=/usr/lib/x86\_64-linux-gnu/libpng.so

//Path to a file.  
PNG\_PNG\_INCLUDE\_DIR:PATH=/usr/include

//Force rebuilding .proto files (protoc should be available)  
PROTOBUF\_UPDATE\_FILES:BOOL=OFF

//Path to Python interpreter  
PYTHON2\_EXECUTABLE:FILEPATH=

//Python include dir  
PYTHON2\_INCLUDE\_DIR:PATH=

//Python include dir 2  
PYTHON2\_INCLUDE\_DIR2:PATH=

//Path to Python library  
PYTHON2\_LIBRARY:FILEPATH=

//Path to Python debug  
PYTHON2\_LIBRARY\_DEBUG:FILEPATH=

//Path to numpy headers  
PYTHON2\_NUMPY\_INCLUDE\_DIRS:PATH=

//Where to install the python packages.  
PYTHON2\_PACKAGES\_PATH:PATH=

//Path to Python interpreter  
PYTHON3\_EXECUTABLE:FILEPATH=/usr/bin/python3

//Python include dir  
PYTHON3\_INCLUDE\_DIR:PATH=/usr/include/python3.8

//Python include dir 2  
PYTHON3\_INCLUDE\_DIR2:PATH=/usr/include/python3.8

//Path to Python library  
PYTHON3\_LIBRARY:FILEPATH=/usr/lib/x86\_64-linux-gnu/libpython3.8.so

//Path to Python debug  
PYTHON3\_LIBRARY\_DEBUG:FILEPATH=

//Path to numpy headers  
PYTHON3\_NUMPY\_INCLUDE\_DIRS:PATH=/usr/include/numpy/

//Where to install the python packages.  
PYTHON3\_PACKAGES\_PATH:PATH=/lib/python3.8/dist-packages

PYTHON\_DEFAULT\_EXECUTABLE:FILEPATH=/usr/bin/python3

PYTHON\_EXECUTABLE:FILEPATH=/usr/bin/python3

PYTHON\_INCLUDE\_DIRS:PATH=/usr/include/python3.8

PYTHON\_LIBRARY:FILEPATH=/usr/lib/x86\_64-linux-gnu/libpython3.8.so.1

做好以上配置后，执行make -j4

大约等待15分钟后编译完成，可以看到，在opencv编译目录的lib/python3/下面

lib/python3/cv2.cpython-38-x86\_64-linux-gnu.so

ls lib

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3pkZjUxMTM5MTU2NQ_size_16_color_FFFFFF_t_70 3][]

ls lib/python3/

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3pkZjUxMTM5MTU2NQ_size_16_color_FFFFFF_t_70 4][]

然后执行

sudo make install

编译完成后，把编译出来的cv2.cpython-36m-x86\_64-linux-gnu.so库(opencv-python)拷到anaconda的python库包下，并执行以下指令

sudo cp lib/python3/cv2.cpython-38-x86\_64-linux-gnu.so /usr/local/lib/python3.8/dist-packages/ -ar

sudo ln -s /usr/local/lib/python3.8/dist-packages/cv2.cpython-38-x86\_64-linux-gnu.so /usr/lib/python3/dist-packages/cv2.so

sudo apt-get install python3-pip

sudo pip3 install --upgrade pip

sudo pip install -U numpy

测试：（依次执行一下指令）

python3

import cv2

cv2.\_\_version\_\_

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3pkZjUxMTM5MTU2NQ_size_16_color_FFFFFF_t_70 5][]

表明配置成功！

[https_github.com_opencv]: https://github.com/opencv
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3pkZjUxMTM5MTU2NQ_size_16_color_FFFFFF_t_70]: /images/20221120/6af836cb4d6044aea679ce342604dc8e.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3pkZjUxMTM5MTU2NQ_size_16_color_FFFFFF_t_70 1]: /images/20221120/1d7f2283150d438a89f7af1595534df9.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3pkZjUxMTM5MTU2NQ_size_16_color_FFFFFF_t_70 2]: /images/20221120/c586ea6993004289831975c537deac24.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3pkZjUxMTM5MTU2NQ_size_16_color_FFFFFF_t_70 3]: /images/20221120/d3db3c5a684f432892833ce0a8767617.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3pkZjUxMTM5MTU2NQ_size_16_color_FFFFFF_t_70 4]: /images/20221120/b2268fa31e194cfba45ec1eba9d08893.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3pkZjUxMTM5MTU2NQ_size_16_color_FFFFFF_t_70 5]: /images/20221120/097cd347409641499bbf6ffbc00aafc5.png