Robert Kern schrieb:
Probably, you need to use zeros(..., dtype=uint8). When you use dtype=int, that will result in dtype=int arrays. I suspect that matplotlib is then interpreting that to mean that you want it to treat the input as scalar data (which it will pass through a colormap) rather than an RGB image.
Thanks Robert, that was exactly the problem. Now I'am really wondering how one can know such details. Well, with your answer I searched again in the mathplotlib documentation and under the function imshow(X, ...) I indeed found the hint, that X has to be an uint8 or float array or PIL image, but before I hadn't known where to search.
So again, thank you Sebastian -- http://mail.python.org/mailman/listinfo/python-list
