HighGUI
OpenCV comes with its own library (highgui) that enables easy control of
several aspects of image work within a program. Its GUI methods are covered
here, but before displaying an image you might like
to know how to load some you 'prepared earlier' and save those that you
alter.
As the name 'highgui' suggests, the methods are kept at a remarkably high level of abstraction and give virtually no control over the internal workings of the interface. For example, you don't specify the file type - this is assumed to be represented by the file's extension.
Reading Files
Loading images couldn't be simpler. Be advised that the method here returns
a pointer to an IplImage, so you'll need to define one first. Assuming you
haven't, this would work:
IplImage *img = cvLoadImage(char[ ] filename, int type);
The filename can also include a path if the image is not in the same
folder. type refers to whether or not the image is colour, and can
be entered as an integer or a constant, depending on personal preference:
CV_LOAD_IMAGE_COLOR = 1 forces the resultant IplImage to be colour.
CV_LOAD_IMAGE_GRAYSCALE = 0 forces a greyscale IplImage.
CV_LOAD_IMAGE_UNCHANGED = -1 generates a file-dependent represenation.
If the file is colour, you get a 3-channel IplImage. If not, you get grayscale.
Writing Files
This is just as simple as before, but without the complication of a type.
cvSaveImage(char[ ] filename, IplImage* img) saves img as
filename, again using the extension specified to determine the filetype.
![[Valid HTML 4.01!]](/vision/valid-html401.gif){kind=small}