You can use any distribution but I suggest you use the Ubuntu distribution as it is a polished release and much easier for newbies to get a grip on. Linux distributions are closely related in speed to each other, so go with a x64 machine with lots of main & video memory would be my 2c offering.

Ubuntu (aka Debian) has a good package manager that allows you to get packages you need like Opencv, Gstreamer tools libs etc. easily. Video speed is more a function of video card characteristics so the faster it is and the more on-board memory it has is all that is needed IMHO. Multi-core CPU with DDR3 or DDR4 main memory is the best.

The following answers my predictament exactly.

IplImage* img = cvLoadImage( argv[1] );
cv::namedWindow( "Test3-Example1", CV_WINDOW_AUTOSIZE );
cvShowImage( "Test3-Example1", img );
cv::waitKey(0);
cvReleaseImage( &img );
cv::destroyWindow( "Test3-Example1" );
return 0;

LoadImage, ShowImage and ReleaseImage needed to have pre-appended "cv" with the third character capitalised.

I found this out by just trying different combinations of function calls from the CV library headers.

cv::namedWindow( "Test3-Example1", CV_WINDOW_AUTOSIZE ); cvShowImage( "Test3-Example1", img ); cv::waitKey(0); cvReleaseImage( &img ); cv::destroyWindow( "Test3-Example1" ); return 0;

Before I mark this as answered, what do I do with the original cv::calls such as showImage, loadImage, releaseImage which exists in the documentation but I am unable to get its header file to play ball? My original problem with the examples, not correlating with the header files still stands.

As I am new here, how do I mark a answer as correct?

[100%] Building CXX object CMakeFiles/test3.dir/test3.cpp.o Linking CXX executable test3 [100%] Built target test3

It all compiled and more importantly Linked includes correctly.

Is this a bug?

./test3 ../cat.jpg read correctly without problems.

Thanks for your sustained help. G.

I am sure I have correct name-space.

Herein is the test program in full (excluding the #include statements as there is not enough room to fit here).

gmk@Brushtail:~/test3$ cat test3.cpp

// From top-of-page Page 17 of the OpenCV book

using namespace cv;

include "my_includes.hpp"

int main(int argc, char** argv) {

IplImage* img = loadImage( argv[1] );

namedWindow( "Test3-Example1", CV_WINDOW_AUTOSIZE );
showImage( "Test3-Example1", img );
waitKey(0);
releaseImage( &img );
destroyWindow( "Test3-Example1" );
return 0;

}

In "my_include.hpp" header file I have all header files from opencv and opencv2, in an effort to over these problems.

Maybe I need to load/run OpenCV on a different Linux distribution (currently Ubuntu 12.04.2 LTS). Any further thoughts?

Hi Guanta; Thanks for your suggestions. I have tried them but unfortunately I ended up with the same "not declared in this scope" errors. I have gone as far as having modified each include statement to reflect the actual include address i.e.: #include "/usr/local/include/opencv2/<include filename>" but again without any change to the make errors.

I forgot to let you know that I did a "cmake ./" which did printout correctly the location of the opencv libraries, and complete without any errors, I followed by a "make".

Any further suggestions. Grahame

I have tried both the CLI and the Eclipse development environments, but each give me the same problems with the errors of:

In function ‘int main(int, char**)’: error: ‘loadImage’ was not declared in this scope; error: ‘showImage’ was not declared in this scope; error: ‘releaseImage’ was not declared in this scope;

I have the following include directives: '#include "cv.h"' '#include "highgui.h"' and I have the exported environmental $LD_LIBRARY_PATH="/usr/local/lib" which is correct as it contains opencv & opencv2 directories.

A> What include files are I missing, but surely the cv.h (aka cv.hpp) should have taken care of the declarations themselves?

B> Do I have to do a "nm <library_name>" to get the includes from each library or is there a easier way to get what header file prototypes are within which library ?

C> Could you advise what I need to do to fix the problems with the 2.5.4 release examples. Actually to assist newbies to CV, it would be a good idea to have all the tutorial examples tested under each new release so it doesn't turn newbies off, as they consider it all too hard just trying to break into OpenCV.

Thanks.