Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

Error Installing OpenCV-2.4.3 on Ubuntu

When I try to cmake to OpenCv-2.4.3 on Ubuntu,I get this Error:

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_nvcuvid_LIBRARY (ADVANCED)

And then ,I disabled "with_cuda" option at CMakeLists.txt .After that,I get this warning:

The source directory is the same as binary directory. "make clean" may damage the source tree

I don't have "makefile" at my build folder ,so I cannot install OpenCV-2.4.3.

like image 522
Pekaziz Avatar asked Dec 11 '22 21:12

Pekaziz


2 Answers

I have just fixed this. OS: Linux Mint 13 / 64bit, CUDA 5.0. It should also work on Ubuntu.

The problem is that you need to create a link to libnvcuvid.so located in /usr/lib/nvidia-current/ from /usr/lib/.

Please execute this in a terminal:

su -c 'ln -s /usr/lib/nvidia-current/libnvcuvid.so /usr/lib/libnvcuvid.so && ln -s /usr/lib/nvidia-current/libnvcuvid.so.1 /usr/lib/libnvcuvid.so.1'

Build files were generated without errors. I would also add that you shouldn't use apt-get for OpenCV, as the repositories contain older versions of the library.

like image 163
Dragos Stanciu Avatar answered Dec 14 '22 11:12

Dragos Stanciu


Where have you performed your make clean ?

If you follow the documentation correctly, you can see that it is advised to create a folder to store everything that is going to be compiled.

Ususally, you do that by performing a

mkdir build

You should not have any problem in following this guide step by step.

AS you can see, the cmake command is :

cmake -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=/usr/local -D BUILD_PYTHON_SUPPORT=ON -D BUILD_EXAMPLES=ON ..

The .. at the end are very important, because it means that you should perform the command in the directory where you want everything to be built.

If you directly ran your cmake into the source directory, you might want to unzip the package and start all over again ;).

You were right to change the variable for CUDA, especially if you don't have an NVidia graphical card.

The last message you indicate is a warning, not a error. And you can solve it by following what I said before.

If you only want to use OpenCV, and not develop for it, you migh as well as Barnabas said directly use you package manager.

In this case, follow his link

like image 33
jlengrand Avatar answered Dec 14 '22 10:12

jlengrand