MacOS安装OpenCV(opencv-python)

下载源码

下载zip文件

gitee: https://gitee.com/mirrors/opencv/tree/4.5.3(建议，速度快)
github: https://github.com/opencv/opencv/tree/4.5.3

git clone项目

gitee

git clone https://gitee.com/mirrors/opencv.git

github

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

构建源码

创建文件夹(与opencv源码同级)

mkdir build_opencv

进入新建的文件夹

cd build_opencv

构建

1、cmake(缺少自行下载)

# CMAKE_INSTALL_PREFIX=./构建后输出文件的位置
# ../opencv源码文件的位置
cmake -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=./ ../opencv

2、make
3、make install

与python建立连接

安装opencv-python依赖包

pip3 install opencv-python

进入python的依赖包site-packages下的cv2目录

cd /Users/xxxx/.virtualenvs/xxxx/lib/python3.9/site-packages/cv2

使该目录下的cv2.cpython-39-darwin.so与build_opencv内的cv2.cpython-39-darwin.so建立软链接

build_opencv内的cv2.cpython-39-darwin.so路径：

/Users/xxxx/opencv/build_opencv/lib/python3.9/site-packages/cv2/python-3.9/cv2.cpython-39-darwin.so

建立软连接

ln -sf /Users/xxxx/opencv/build_opencv/lib/python3.9/site-packages/cv2/python-3.9/cv2.cpython-39-darwin.so ./cv2.cpython-39-darwin.so

导入cv2

使用虚拟环境建立的软链接先进入虚拟环境

# 进入python
python
# 导入cv2
import cv2
# 查看opencv版本
cv2.__version__