Ubuntu16.04编译opencv3.1 CUDA9.0

1.官网下载opencv3.1.0源代码；https://github.com/opencv/opencv/archive/3.1.0.zip

2.修改cmake文件夹中的FindCUDA.cmake文件；

(1)查找“unset(CUDA_nppi_LIBRARY CACHE)”；

替换为：

  unset(CUDA_nppial_LIBRARY CACHE)
  unset(CUDA_nppicc_LIBRARY CACHE)
  unset(CUDA_nppicom_LIBRARY CACHE)
  unset(CUDA_nppidei_LIBRARY CACHE)
  unset(CUDA_nppif_LIBRARY CACHE)
  unset(CUDA_nppig_LIBRARY CACHE)
  unset(CUDA_nppim_LIBRARY CACHE)
  unset(CUDA_nppist_LIBRARY CACHE)
  unset(CUDA_nppisu_LIBRARY CACHE)
  unset(CUDA_nppitc_LIBRARY CACHE)

(2)查找"find_cuda_helper_libs(nppi)"；

替换为：

  find_cuda_helper_libs(nppial)
  find_cuda_helper_libs(nppicc)
  find_cuda_helper_libs(nppicom)
  find_cuda_helper_libs(nppidei)
  find_cuda_helper_libs(nppif)
  find_cuda_helper_libs(nppig)
  find_cuda_helper_libs(nppim)
  find_cuda_helper_libs(nppist)
  find_cuda_helper_libs(nppisu)
  find_cuda_helper_libs(nppitc)

3.安装cmake工具；

sudo apt-get install cmake

4.安装依赖库；

sudo apt-get install build-essential libgtk2.0-dev libavcodec-dev libavformat-dev libjpeg.dev libtiff4.dev libswscale-dev libjasper-dev qtbase5-dev

5.将头文件cuda_fp16.h添加至 opencv\modules\cudev\include\opencv2\cudev\common.hpp；

即在common.hpp中添加：

#include <cuda_fp16.h>

6.修改opencv3.1.0/modules/cudalegacy/src/graphcuts.cpp文件；

#if !defined (HAVE_CUDA) || defined (CUDA_DISABLER)
修改为：
#if !defined (HAVE_CUDA) || defined (CUDA_DISABLER) || (CUDART_VERSION>=8000)

7.解压下载的opencv3.1.0源码，进入到opencv3.1.0文件夹，创建release文件夹并进入；

mkdir release && cd release

8.cmake配置编译 ；

cmake -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=/usr/local -D WITH_TBB=ON -D BUILD_NEW_PYTHON_SUPPORT=ON -D WITH_V4L=ON -D WITH_QT=ON -D WITH_OPENGL=ON  -D CUDA_GENERATION=Kepler -D BUILD_PNG=ON ..

9.make编译和安装；

make -j$(nproc)
sudo make install

10.安装完成！