dandelion1124's profile - activity

I tried to reproduce this issue on OpenCV 3.1+Windows 10.
But, I could not reproduce this issue.

So, could you please tell us information of your environment(Platform, Compiler, OpenCV version)?
You can get these information from cv::getBuildInformation().

And, this issue might be related to this ticket. https://github.com/Itseez/opencv/issu...

Did you specify WITH_NVCUVID in CMake?
If you use cudacodec module, you need to specify this option in CMake.

But, I can not find NVCUVENC in NVIDIA CUDA Toolkit v7.0.
So, I disabled encoding feature of cudacodec
I wrote a method to use decoding feature of cudacodec in my webpage(Content is Japanese).
https://goo.gl/i6HZwF

Sorry, I was overlooking opencv_world300d.lib.
I found the following point.

[Library path]

Compiler: Visual Studio 2012
Library path: C:\opencv\build\x64\vc12\lib

I think that you need to set C:\opencv\build\x64\vc11\lib for Visual Studio 2012(VC11).

[Solution platform]

I created an empty project in Visual Studio 2012 (win32 console application).

I think that you need to change Solution platform from Win32 to x64.

In OpenCV 2.4.x, cv::imread was provided in highgui module.
But, this function moved to imgcodecs module from OpenCV 3.0.

So, you need to include "opencv2\imgcodecs\imgcodecs.hpp" and
link opencv_imgcodecs300d.lib to use this function.

If you execute the following code, you can know the information about compiled OpenCV library.

#include <opencv2/core.hpp>
#include <iostream>

int main(int argc, char **argv)
{
    std::cout<< cv::getBuildInformation() << std::endl;
    return 0;
}

And, please tell me the version of Qt that you installed?

I found other solution. I replaced my "build_framework.py".
https://gist.github.com/atinfinity/cc...

Thank you for your comment!
I' ll try it.

The turbojpeg-static.lib provides 'TurboJPEG C API', too.
If you use this API, you need to link this library.

You can know 'TurboJPEG C API' by the following information.

• http://www.libjpeg-turbo.org/About/Tu...
• https://cdn.rawgit.com/libjpeg-turbo/...

Current OpenCV uses only 'libjpeg API'.
So, I used jpeg-static.lib.

I think that your platform is Windows(64bit).
Is my understanding correct?

If your answer is 'yes', I think that you need to install libjpeg-turbo-1.4.1-vc64.exe and add the entry of JPEG_LIBRARY(Value is C:/libjpeg-turbo64/lib/jpeg-static.lib) as FILEPATH. Please try it.

Previous version CUDA does not support VS2013.
From this background, my batch file specify VS2012.

But, the latest CUDA support VS2013. So, we can specify VS2013 now.
Does that answer your question?

Do you set the location of the cmake.exe to the environment variable (PATH)?
Because, this setting is required to run the cmake on command prompt.

"cmake.exe" is not recognized as an internal or external command....and...impossible to find the files: openCV.sln...

Maybe, I think that cmake.exe was not found.
As a result, OpenCV.sln was not generated.

Sorry, I misunderstood.
I made a batch file to build OpenCV(WITH_CUDA).
Please run this batch file in Visual Studio command prompt(VS2012).

https://github.com/atinfinity/lab/wik...

I explain the procedure using cmake-gui.

1. launch cmake-gui
2. input the path of your souce code
3. input the path for build(solution file is generated in this place)
4. push Configure button
5. select your compiler and architecture(win32 or x64)
6. push Finish button
7. edit the path of OpenCV_DIR
8. push Configure button
9. push Generate button

As a result, solution file is generated in the path for build.

I made CMakeLists.txt with CUDA.
Please use it as a reference.
https://gist.github.com/atinfinity/7a...

From OpenCV 3.0.0-rc1, --contrib option is added to platforms/ios/build_framework.py.
As a result, we can specify opencv_contrib's path by this option. https://github.com/Itseez/opencv/blob...

So, I tried to build opencv2.framework with opencv_contrib in the following steps.

1. get the code of opencv-3.0.0-rc1.
2. get the code of opencv_contrib.
3. run python opencv-3.0.0-rc1/platforms/ios/build_framework.py ios to generate opencv2.framework.
4. run python opencv-3.0.0-rc1/platforms/ios/build_framework.py --contrib opencv_contrib ios_contrib to generate opencv2_contrib.framework.

And, I made iOS project using Xcode. And, I imported opencv2.framework and opencv2_contrib.framework to my project.

After that I added the following code to my project.

#import <opencv2/core.hpp>
#import <opencv2/features2d.hpp>
#import <opencv2_contrib/xfeatures2d.hpp>

But, the following build error occurs.

/Users/xxxx/dev/ios_contrib/opencv2_contrib.framework/Headers/xfeatures2d.hpp:43:10: 'opencv2/xfeatures2d/nonfree.hpp' file not found

Is this procedure is wrong to use opencv2.framework with opencv_contrib?

I explicitly specified the following option in CMake command. As a result, CMake could find Qt.

-D Qt5Concurrent_DIR="C:\Qt\Qt5.4.0\5.4\msvc2012_opengl\lib\cmake\Qt5Concurrent" ^
-D Qt5Core_DIR="C:\Qt\Qt5.4.0\5.4\msvc2012_opengl\lib\cmake\Qt5Core" ^
-D Qt5Gui_DIR="C:\Qt\Qt5.4.0\5.4\msvc2012_opengl\lib\cmake\Qt5Gui" ^
-D Qt5OpenGL_DIR="C:\Qt\Qt5.4.0\5.4\msvc2012_opengl\lib\cmake\Qt5OpenGL" ^
-D Qt5Test_DIR="C:\Qt\Qt5.4.0\5.4\msvc2012_opengl\lib\cmake\Qt5Test" ^
-D Qt5Widgets_DIR="C:\Qt\Qt5.4.0\5.4\msvc2012_opengl\lib\cmake\Qt5Widgets" ^

I tried the following environment.

• OpenCV 2.4.9
• VS2012 Win32
• Qt5.4.0
• Windows 8.1 Pro

On the other hand, CMake can find Qt without this specify in Ubuntu14.10(https://gist.github.com/atinfinity/259249a2ef787e358e72) . So, could you please teach the result of following command?

dpkg -l | grep libqt

[before]

Mat img = imread("MyPic.JPG");

[after]

Mat img = imread("MyPic.JPG");
if(img.empty())
{
    cerr << "Cannot read image file" << endl;
    return -1;
}

Could you try to change your code? If the error message("Cannot read image file") appears, imread() failed.

[before] Mat img = imread("MyPic.JPG");

[after] Mat img = imread("MyPic.JPG"); if(img.empty()) { cerr << "Cannot read image file" << endl; return -1; }

Could you try to change your code? If the error message("Cannot read image file") appears, imread() failed.

Please let me explain further. If you call ocl::setUseOpenCL() explicitly, you can select the behavior of UMat. This function's argument is boolean value. If you use this function, you need to include <opencv2/core/ocl.hpp>.

"T-API" means "Transparent API".

I think that the following document is useful to understand "Transparent API". http://code.opencv.org/attachments/1591/OpenCV3_0_CVPR_2014.pptx

On my environment, I can select only GPU as OpenCL platform. So, I can not run my program without CUDA. And, there is the following explanation in Beignet's web page. Maybe, this explanation might be helpful for you.

• Note about OpenCV support

I think that you can use Intel OpenCL SDK as other choices. By the way, you can get build infomation of OpenCV to call getBuildInformation(). This information is very useful to report your environment.

I tried using only OpenCL(with CUDA). If you possible, could you please tell me the result of this program?

https://gist.github.com/atinfinity/276c02f2c24cde30ae5b