1.下载opencv4.5.0和opencv_contrib4.5.0源码

opencv_contrib4.5.0:

• svn : svn://116.56.140.138/soft/opencv_contrib-4.5.0.zip

• github：https://github.com/opencv/opencv_contrib/archive/4.5.0.zip

opencv4.5.0

• svn：svn://116.56.140.138/soft/opencv-4.5.0.zip

• github：https://github.com/opencv/opencv/archive/4.5.0.zip

2.编译cuda版的opencv

解压

• unzip opencv-4.5.0.zip

• unzip opencv_contrib-4.5.0.zip -d opencv-4.5.0

生成MakeFile文件

• cd opencv-4.5.0

• mkdir build

• cd build

• cmake

cmake  \
-D CMAKE_BUILD_TYPE=RELEASE \
-D CMAKE_INSTALL_PREFIX=/usr/local \
-D WITH_TBB=ON \
-D WITH_V4L=ON \
-D WITH_QT=OFF \
-D WITH_OPENGL=ON \
-D WITH_CUDA=ON \
-D CUDA_GENERATION="Auto" \
-D ENABLE_FAST_MATH=1 \
-D CUDA_FAST_MATH=1 \
-D CUDA_NVCC_FLAGS="-D_FORCE_INLINES" \
-D WITH_CUBLAS=1 \
-D OPENCV_EXTRA_MODULES_PATH=../opencv_contrib-4.5.0/modules \
-D BUILD_opencv_hdf=ON ..

执行MakeFile

• 查看cpu总个数、总核数、总线程数

cat /proc/cpuinfo |grep 'physical id'|sort -u|wc -l
cat /proc/cpuinfo |grep 'core id'|sort -u|wc -l
cat /proc/cpuinfo |grep 'processor'|sort -u|wc -l

• make，其中-j表示最多允许多少个编译命令同时执行，一般是cpu核数的两倍

make -j20
make install

• 如果之前make失败了，再次make时必须clean build目录

make clean