[Image-SIG] Creating an image from <canvas>'s getImageData() data

[Fredrik Lundh]

On Tue, Dec 9, 2008 at 5:35 PM, Fredrik Lundh <fredrik at pythonware.com> wrote:

> The "raw" data decoder expects raw binary bytes; in this case, it
> expects four bytes per pixel (<red> <green> <blue> <alpha>), repeated
> 80x80 times to fill the entire image.  What does the data you get back
> from getImageData() look like if you print (a portion of) it?  e.g.
> say
>
>    print repr(data[:40])
>
> (That "replace" you're using makes me think that you have data in text
> form, not binary form...)

And for the record, here's a fairly efficient way to convert
comma-separated pixel values into a PIL Image object:

>>> data = "1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16"

>>> import Image
>>> import array
>>> a = array.array("b", map(int, data.split(",")))
>>> a
array('b', [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16])

# turn it into a grayscale image (one byte per pixel)
>>> i = Image.frombuffer("L", (4, 4), a, "raw", "L", 0, 1)
>>> i.mode
'L'
>>> list(i.getdata())
[1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]

# turn it into an RGBA image (four bytes per pixel)
>>> i = Image.frombuffer("RGBA", (1, 4), a, "raw", "RGBA", 0, 1)
>>> i.mode
'RGBA'
>>> list(i.getdata())
[(1, 2, 3, 4), (5, 6, 7, 8), (9, 10, 11, 12), (13, 14, 15, 16)]

</F>