worst guy ever's profile - activity

Intellisense will only predict compiler errors, this is a link stage error. You need to add the appropriate library directory to your project configuration. For a 32-bit project use opencv-2.4.x\build\x86\vc10\lib.

I sympathise with your confusion at the state of the official documentation, but your example is an issue with your Python syntax, not with OpenCV or its documentation. Importing like so:

import cv

will not add symbols in the cv module to the local namespace. You can either reference the cv module every time you want to use a symbol from it (e.g.: cv.cvtColor and cv.CV_RGB2GRAY) or import the symbols you need into the local namespace (e.g.: from cv import cvtColor, CV_RGB2GRAY) or just import everything: from cv import *.

I don't know of any better source of documentation, sorry.

Find them in opencv-2.4.0\build\x86\vc10\bin, substituting x86 and vc10 for whichever architecture and build system you're using.

No, just that they are unimportant. Some memory is being allocated on the heap during initialisation, and _CrtDumpMemoryLeaks() just gives you a summary of the heap. A few stray entries is perfectly OK and normal and you shouldn't spend another minute worrying about it.

These leaks are spurious: the program does nothing, and they only amount to a few hundred bytes, once-off, before the call to main(). The only use of this list is to compare against the leaks that show up just before the program terminates so that you can work out which of them are occurring in your own code.

The clue is in the error message:

'array op array' (where arrays have the same size and the same number of channels)

Your addition fails because panels[0] is a single channel image and image2 has as many channels as are in rain.jpg (presumably three). What effect are you trying to achieve?

What kind of camera parameters are you looking for specifically?