我的环境是cuda10.1
cmake是3.16.5
第一步骤：下载安装包
第二步骤：编译环境安装
第三步骤：解压opencv安装包，并且在opencv文件夹内新建build文件夹
第四步骤：打开cmake-gui，
第五步骤：进行编译
总结：
报错解决方法汇总：
ICV: Failed to download ICV package:ippicv_linux_20151201.tgz
/usr/include/c++/7/cstdlib:75:15: fatal error: stdlib.h: No such file or directory
error: ‘NppiGraphcutState’ has not been declared typedef NppStatus (sudo make compile )
Note: Class l has more than 1 base class (not supported by Python C extensions)Bases: cv::a, cv:😒, cv:😒, cv::A, cv::l, cv::g, cv:😮, cv::r, cv::i, cv::t, cv::h, cv::mOnly the first base class will be usedGenerator error: constant l_MAGIC_VAL (cname=cv::l::MAGIC_VAL) already exists
#error -- unsupported GNU version! gcc versions later than 8 are not supported!
CMake Error: The following variables are used in this project, but they are set to NOTFOUND.Please set them or make sure they are set and tested correctly in the CMake files:CUDA_nppi_LIBRARY (ADVANCED)
linked by target “opencv_cudev“ in directory /home/davidhopper/opencv/opencv-3.1.0/
nvcc fatal : Unsupported gpu architecture 'compute_20'CMake Error at cuda_compile_generated_gpu_mat.cu.o.cmake:206 (message): Error generating /home/danalelixin/Desktop/opencv-3.2.0/build/modules/core/CMakeFiles/cuda_compile.dir/src/cuda/./cuda_compile_generated_gpu_mat.cu.o
opencv-3.1.0/modules/videoio/src/cap_ffmpeg_impl.hpp:1313:21: error: ‘CODEC_FLAG_GLOBAL_HEADER’ was not declared in this scope c->flags |= CODEC_FLAG_GLOBAL_HEADER;
opencv-3.1.0/modules/videoio/src/cap_ffmpeg_impl.hpp:1525:35: error: ‘AVFMT_RAWPICTURE’ was not declared in this scope
第一步骤：下载安装包
安装包名字为opencv-3.1.0.zip
opencv的安装链接看这个（内有cmake下载地址）https://blog.csdn.net/m0_69253695/article/details/136314667?spm=1001.2014.3001.5502
第二步骤：编译环境安装
sudo apt-get install build-essential
sudo apt-get install cmake git libgtk2.0-dev pkg-config libavcodec-dev libavformat-dev libswscale-dev
sudo apt-get install python-dev python-numpy libtbb2 libtbb-dev libjpeg-dev libpng-dev
第三步骤：解压opencv安装包，并且在opencv文件夹内新建build文件夹
第四步骤：打开cmake-gui，
点configure，当没有出现红色报错后，即可点击generate
在build文件夹下点击鼠标右键，在终端中输入下面的命令也可以（但不建议这样，因为cmake的时候会有报错，需要在可视化界面内进行修改，比较方便，要不然还要修改命令就很麻烦）
sudo cmake -D CMAKE_BUILD_TYPE=Release -D CMAKE_INSTALL_PREFIX=/usr/local ..
第五步骤：进行编译
然后在build文件夹下右键，选择终端，然后输入下面这个命令
sudo make -j8等到100%之后再输入下面这个命令，就完成安装了
sudo make install总结：
安装opencv3.1.0这种老版本的opencv与新版本的opencv步骤是一样的，安装步骤不是重点，重点是安装过程中遇到的报错该怎么解决，会遇到很多不兼容的报错。详细的安装步骤可以看下面的连接https://blog.csdn.net/public669/article/details/99044895
报错解决方法汇总：
ICV: Failed to download ICV package:ippicv_linux_20151201.tgz
解决办法是下载ippicv_linux_20151201.tgz包，并将包拷贝到下面这个文件夹内
opencv-3.1.0/3rdparty/ippicv/downloads/linux-808b791a6eac9ed78d32a7666804320e
下面是下载的链接，下图中第一个包
https://codeload.github.com/opencv/opencv_3rdparty/zip/refs/heads/ippicv/master_20151201 
/usr/include/c++/7/cstdlib:75:15: fatal error: stdlib.h: No such file or directory
解决办法是在/build目录下的cmake中添加指令-D ENABLE_PRECOMPILED_HEADERS=OFF
或者在cmake-gui界面中将ENABLE_PRECOMPILED_HEADERS取消勾选
error: ‘NppiGraphcutState’ has not been declared typedef NppStatus (sudo make compile )
解决方法是打开下面的.cpp文件
/home/lvv/Downloads/opencv-3.1.0/modules/cudalegacy/src/graphcuts.cpp将#if !defined (HAVE_CUDA) || defined (CUDA_DISABLER)
改为：
#if !defined (HAVE_CUDA) || defined (CUDA_DISABLER) || (CUDART_VERSION >= 8000)
修改后如下图
Note: Class l has more than 1 base class (not supported by Python C extensions)
Bases: cv::a, cv:😒, cv:😒, cv::A, cv::l, cv::g, cv:😮, cv::r, cv::i, cv::t, cv::h, cv::m
Only the first base class will be used
Generator error: constant l_MAGIC_VAL (cname=cv::l::MAGIC_VAL) already exists
解决方法是在/build目录下的cmake中添加指令 -DBUILD_opencv_python2=OFF -DBUILD_opencv_pyhton3=OFF
或者在cmake-gui界面中将BUILD_opencv_python2和BUILD_opencv_pyhton3取消勾选
#error -- unsupported GNU version! gcc versions later than 8 are not supported!
这个报错意思是gcc版本不能高于8
所以安装gcc-8及g++-8
sudo apt-get install g++-8
sudo apt-get install gcc-8
安装成功后设置gcc与g++各版本的优先级，数字越小优先级越高，默认的优先级为50，故设置的比50小即可，这里选择20：
sudo update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-8 20
sudo update-alternatives --install /usr/bin/g++ g++ /usr/bin/g++-8 20
通过sudo update-alternatives --config gcc
sudo update-alternatives --config g++
可查看电脑上gcc与g++的可供选择版本，也可以通过上述命令手动选择编译所用版本
CMake Error: The following variables are used in this project, but they are set to NOTFOUND.
Please set them or make sure they are set and tested correctly in the CMake files:
CUDA_nppi_LIBRARY (ADVANCED)
linked by target “opencv_cudev“ in directory /home/davidhopper/opencv/opencv-3.1.0/
解决办法看这个链接https://blog.csdn.net/m0_69253695/article/details/143068384?spm=1001.2014.3001.5502
nvcc fatal : Unsupported gpu architecture 'compute_20'
CMake Error at cuda_compile_generated_gpu_mat.cu.o.cmake:206 (message):
Error generating
/home/danalelixin/Desktop/opencv-3.2.0/build/modules/core/CMakeFiles/cuda_compile.dir/src/cuda/./cuda_compile_generated_gpu_mat.cu.o
老毛病，还是cuda9不支持2.0架构了。/build目录下的cmake中添加指令
D CUDA_GENERATION=Kepler
opencv-3.1.0/modules/videoio/src/cap_ffmpeg_impl.hpp:1313:21: error: ‘CODEC_FLAG_GLOBAL_HEADER’ was not declared in this scope
c->flags |= CODEC_FLAG_GLOBAL_HEADER;
opencv-3.1.0/modules/videoio/src/cap_ffmpeg_impl.hpp:1525:35: error: ‘AVFMT_RAWPICTURE’ was not declared in this scope
就是ffmpeg新版本对老版本中的一些关键字的支持变了。
打开文件opencv/modules/videoio/src/cap_ffmpeg_impl.hpp ，在里面添加
#define AV_CODEC_FLAG_GLOBAL_HEADER (1 << 22)
#define CODEC_FLAG_GLOBAL_HEADER AV_CODEC_FLAG_GLOBAL_HEADER
#define AVFMT_RAWPICTURE 0x0020
重新cmake，有的时候添加语句后，还是没用，可能是添加在文件中的位置不对，换一个位置就好了，建议放在.hpp的最开头
