ubuntu 16.04 cuda-9.0 编译opencv2.4.13报错

cmake:

cmake -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=/usr/local/opencv2.4.13 -D WITH_CUDA=ON -D ENABLE_FAST_MATH=ON -D CUDA_FAST_MATH=ON -D WITH_CUBLAS=1 -D WITH_NVCUVID=on -D CUDA_GENERATION=Auto -D WITH_GTK=ON ..

cmake时报错如下：

CMake Warning at cmake/OpenCVPackaging.cmake:23 (message):
  CPACK_PACKAGE_VERSION does not match version provided by version.hpp
  header!
Call Stack (most recent call first):
  CMakeLists.txt:1103 (include)


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:
opencv_dep_CUDA_nppi_LIBRARY

原因解析：

cuda9不再支持opencv 2.0架构

解决方案如下：

1）.找到cmake文件夹中FindCUDA.cmake文件

找到行

find_cuda_helper_libs(nppi)
改为
  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)
2）.找到行
set(CUDA_npp_LIBRARY "${CUDA_nppc_LIBRARY};${CUDA_nppi_LIBRARY};${CUDA_npps_LIBRARY}")
改为
set(CUDA_npp_LIBRARY "${CUDA_nppc_LIBRARY};${CUDA_nppial_LIBRARY};${CUDA_nppicc_LIBRARY};${CUDA_nppicom_LIBRARY};${CUDA_nppidei_LIBRARY};${CUDA_nppif_LIBRARY};${CUDA_nppig_LIBRARY};${CUDA_nppim_LIBRARY};${CUDA_nppist_LIBRARY};${CUDA_nppisu_LIBRARY};${CUDA_nppitc_LIBRARY};${CUDA_npps_LIBRARY}")
3）.找到行
unset(CUDA_nppi_LIBRARY CACHE)
改为
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)

4）.找到文件OpenCVDetectCUDA.cmake
修改以下几行

 ...
  set(__cuda_arch_ptx "")
  if(CUDA_GENERATION STREQUAL "Fermi")
    set(__cuda_arch_bin "2.0")
  elseif(CUDA_GENERATION STREQUAL "Kepler")
    set(__cuda_arch_bin "3.0 3.5 3.7")
  ...
改为
  ...
  set(__cuda_arch_ptx "")
  if(CUDA_GENERATION STREQUAL "Kepler")
    set(__cuda_arch_bin "3.0 3.5 3.7")
  elseif(CUDA_GENERATION STREQUAL "Maxwell")
    set(__cuda_arch_bin "5.0 5.2")
  ...

5）.cuda9中有一个单独的halffloat(cuda_fp16.h)头文件,也应该被包括在opencv的目录里

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
重新生成即可

 

试了下这篇博客的方案成功了。即

cmake 时指定CUDA_nppi_LIBRARY=true:

~# cmake .. -DCMAKE_BUILD_TYPE=Release -DCUDA_nppi_LIBRARY=true

编译成功

如果还报错，参考https://blog.csdn.net/dawn_chen121/article/details/82828629