I need to use scanner in 16 bit depth and color mode, so i modified python-imaging-sane (that doesn't support RGB tiff with 16 bit depth) to receive from a scanner (epson v500) an image in a Python string.
this is the modified function that i use to send data from the scanner to a python string:
#define READSIZE 32768
static PyObject *
SaneDev_read(SaneDevObject *self, PyObject *args)
{
SANE_Status st;
unsigned char c_buf[READSIZE];
SANE_Int len, maxlen;
maxlen = READSIZE;
if (!PyArg_ParseTuple(args, ""))
{
return NULL;
}
st = (int) sane_read(self->h, (SANE_Byte *) c_buf, maxlen, &len);
return Py_BuildValue("is#", st, c_buf, len);
}
i use this python script to receive and use data:
import sane
import pgmagick
import cStringIO
sane.init()
s = sane.open(sane.get_devices()[0][0])
s.mode = 'Color'
s.source = 'Transparency Unity'
s.tl_x = 15
s.tl_y = 30
s.br_x = 52
s.br_y = 55
s.x_resolution = 1600
s.y_resolution = 1600
s.depth = 16
fbuffer = cStringIO.StringIO()
s.start()
par = s.get_parameters()
print "par = ", par
st = 0
while st is 0:
st, buf = s.read()
fbuffer.write(buf)
s.cancel()
data = fbuffer.getvalue()
fbuffer.close()
px = par[2][0]
py = par[2][1]
bytesperlines = par[4]
depth = par[3]
size = "%sx%s" % (px, py)
blob = pgmagick.Blob(data)
im = pgmagick.Image()
im.density("1600x1600")
im.depth(depth)
im.size(size)
im.magick('RGB')
im.resolutionUnits(gm.ResolutionType.PixelsPerInchResolution)
im.read(blob)
im.write("img.tiff")
the script works very well with 8 bit depth, but with depth set to 16 bit i obtain an image with wrong colors;
these are two examples:
where is the problem?
EDIT: i use pgmagick, a python wrapper to graphicsmagick; graphicsmagick is compiled with quantum depth set to 16 bit.
You probably have the byte order wrong. Try swapping the two bytes that make up the sixteen bit data and see what you get.