创建一个编译目录，并进入。
cuda的安装建议从官网上直接下载run文件，一次性完成cuda nvcc等的安装。
CUDA Toolkit 11.7 Downloads | NVIDIA Developer
wget https://developer.download.nvidia.com/compute/cuda/11.7.0/local_installers/cuda_11.7.0_515.43.04_linux.runsudo sh cuda_11.7.0_515.43.04_linux.run
默认全选安装，建议不要修改。
安装完cuda toolkit之后安装一些必要的软件包，此处如果安装失败，可以多试几次或者换一下源。
# Install prerequisites
sudo apt-get update && apt-get upgrade -y &&
apt-get install -y
build-essential
unzip
yasm
pkg-config
libswscale-dev
libtbb2
libtbb-dev
libjpeg-dev
libpng-dev
libtiff-dev
libavformat-dev
libpq-dev
libxine2-dev
libglew-dev
libtiff5-dev
zlib1g-dev
libjpeg-dev
libavcodec-dev
libavformat-dev
libavutil-dev
libpostproc-dev
libswscale-dev
libeigen3-dev
libtbb-dev
libgtk2.0-dev
下载opencv以及opencv-contrib
# Download and unpack sources wget https://github.com/opencv/opencv/archive/refs/tags/4.5.2.zip -O opencv.zip wget https://github.com/opencv/opencv_contrib/archive/refs/tags/4.5.2.zip -O opencv_contrib.zip unzip opencv.zip mv opencv-4.5.2 opencv unzip opencv_contrib.zip mv opencv_contrib-4.5.2 opencv_contrib
创建一个编译目录，并进入。
# Create build directory and switch into it mkdir opencv/build && cd opencv/build
执行cmake，亲测有效
cmake -DCMAKE_BUILD_TYPE=RELEASE -DCMAKE_INSTALL_PREFIX=/usr/local -DOPENCV_EXTRA_MODULES_PATH=../opencv_contrib/modules .. -DWITH_CUDA=1 -DENABLE_FAST_MATH=1 -DCUDA_FAST_MATH=1 -DWITH_CUBLAS=1 -DOPENCV_GENERATE_PKGCONFIG=1 -DCUDA_GENERATION=Pascal ..
执行编译和安装。
make -j8
安装可能耗时较长
make install
