Hi, this is a question I have posted earlier, but unfortunately I didn't get any answer. if anybody has any hint on how to do this, I would be most graceful!! Thanks in advance! I would like to visualize an image after a rotation: ==> this means to view each squared pixels as "rotated" (seen as an rotated square). I have in fact several images which I need to plot on the same figure (with subplot), each of these having different "rotation angles". The rough solution would be to rotate the data itself (x and y) and use imshow after some rebinning on a squared grid. But this would not be showing the original data, which is what I wish to do. Is it possible to do this in mpl? thanks! Eric
I think you can get the effect you describe using pcolor or pcolormesh; the latter is faster but due to a bug it doesn't handle alpha values other than 1. You will have to generate arrays with the pixel corners (not centers) as you want them to be after your rotation. You will want to use the shading='flat' kwarg. Eric Eric Emsellem wrote: > Hi, > > this is a question I have posted earlier, but unfortunately I didn't get > any answer. > if anybody has any hint on how to do this, I would be most graceful!! > Thanks in advance! > > I would like to visualize an image after a rotation: > ==> this means to view each squared pixels as "rotated" (seen as an > rotated square). I have in fact several images which I need to plot on > the same figure (with subplot), each of these having different "rotation > angles". > > The rough solution would be to rotate the data itself (x and y) and use > imshow after some rebinning on a squared grid. But this would not be > showing the original data, which is what I wish to do. > Is it possible to do this in mpl? > > thanks! > > Eric > > ------------------------------------------------------------------------- > Take Surveys. Earn Cash. Influence the Future of IT > Join SourceForge.net's Techsay panel and you'll get the chance to share your > opinions on IT & business topics through brief surveys -- and earn cash > http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV > _______________________________________________ > Matplotlib-users mailing list > Mat...@li... > https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>>>> "Eric" == Eric Emsellem <ems...@ob...> writes:
Eric> Hi, this is a question I have posted earlier, but
Eric> unfortunately I didn't get any answer. if anybody has any
Eric> hint on how to do this, I would be most graceful!! Thanks
Eric> in advance!
I looked at this a bit -- the underlying image extension code handles
image rotations but it is not exposed at the python level. I spent
some time working on an image class that would handle rotations (in
this test code below I just hardcoded the rotation for testing). The
missing part is to get the extent and image placement algorithms to do
the layout properly in the presence of rotation (eg handling extent
and corners properly below). But this should give the enterprising
developer a head start if they want to run with with. Basically, I
just copied the guts out of the axes.image.AxesImage.make_image code
to experiment with adding a rotation
from matplotlib.image import AxesImage
from pylab import subplot, show, nx
class RotatedImage(AxesImage):
def make_image(self):
from matplotlib.colors import normalize, colorConverter
from matplotlib.numerix import arange, asarray, UInt8, Float32, repeat, NewAxis, typecode
import matplotlib._image as _image
if self._A is not None:
if self._imcache is None:
if typecode(self._A) == UInt8:
im = _image.frombyte(self._A, 0)
im.is_grayscale = False
else:
x = self.to_rgba(self._A, self._alpha)
im = _image.fromarray(x, 0)
if len(self._A.shape) == 2:
im.is_grayscale = self.cmap.is_gray()
else:
im.is_grayscale = False
self._imcache = im
else:
im = self._imcache
else:
raise RuntimeError('You must first set the image array or the image attribute')
bg = colorConverter.to_rgba(self.axes.get_frame().get_facecolor(), 0)
if self.origin=='upper':
im.flipud_in()
im.set_bg( *bg)
im.set_interpolation(self._interpd[self._interpolation])
# image input dimensions
numrows, numcols = im.get_size()
im.reset_matrix()
xmin, xmax, ymin, ymax = self.get_extent()
dxintv = xmax-xmin
dyintv = ymax-ymin
# the viewport scale factor
sx = dxintv/self.axes.viewLim.width()
sy = dyintv/self.axes.viewLim.height()
if im.get_interpolation()!=_image.NEAREST:
im.apply_translation(-1, -1)
# the viewport translation
tx = (xmin-self.axes.viewLim.xmin())/dxintv * numcols
#if flipy:
# ty = -(ymax-self.axes.viewLim.ymax())/dyintv * numrows
#else:
# ty = (ymin-self.axes.viewLim.ymin())/dyintv * numrows
ty = (ymin-self.axes.viewLim.ymin())/dyintv * numrows
l, b, widthDisplay, heightDisplay = self.axes.bbox.get_bounds()
im.apply_translation(tx, ty)
im.apply_scaling(sx, sy)
# resize viewport to display
rx = widthDisplay / numcols
ry = heightDisplay / numrows
im.apply_scaling(rx, ry)
im.apply_rotation(45.)
#print tx, ty, sx, sy, rx, ry, widthDisplay, heightDisplay
im.resize(int(widthDisplay+0.5), int(heightDisplay+0.5),
norm=self._filternorm, radius=self._filterrad)
if self.origin=='upper':
im.flipud_in()
return im
ax = subplot(111)
im = RotatedImage(ax, interpolation='nearest')
im.set_data(nx.mlab.rand(10,10))
xmin, xmax, ymin, ymax = im.get_extent()
corners = (xmin, ymin), (xmax, ymax)
ax.update_datalim(corners)
ax.set_xlim((xmin, xmax))
ax.set_ylim((ymin, ymax))
ax.images.append(im)
show()
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> <html> <head> <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type"> </head> <body bgcolor="#ffffff" text="#000000"> beautiful!<br> I was in the meantime working out something using the suggestion Eric sent me (e.g. via pcolormesh) and I could work most of things I wanted out, but at the price of a rather stupidly looking management of the axis and rotation (maybe there is a way for improvement here too...). <br> Your solution may indeed be a cleaner way forward! I'll try both ways and see how I manage to get things done (the main problem being my rather poor ability in writing advanced codes in python...<br> <br> thanks &<br> cheers!<br> Eric<br> <br> John Hunter wrote: <blockquote cite="mid...@pe..." type="cite"> <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite"> <pre wrap="">"Eric" == Eric Emsellem <a class="moz-txt-link-rfc2396E" href="mailto:ems...@ob..."><ems...@ob...></a> writes: </pre> </blockquote> </blockquote> </blockquote> </blockquote> </blockquote> <pre wrap=""><!----> Eric> Hi, this is a question I have posted earlier, but Eric> unfortunately I didn't get any answer. if anybody has any Eric> hint on how to do this, I would be most graceful!! Thanks Eric> in advance! I looked at this a bit -- the underlying image extension code handles image rotations but it is not exposed at the python level. I spent some time working on an image class that would handle rotations (in this test code below I just hardcoded the rotation for testing). The missing part is to get the extent and image placement algorithms to do the layout properly in the presence of rotation (eg handling extent and corners properly below). But this should give the enterprising developer a head start if they want to run with with. Basically, I just copied the guts out of the axes.image.AxesImage.make_image code to experiment with adding a rotation from matplotlib.image import AxesImage from pylab import subplot, show, nx ... </pre> </blockquote> <br> <pre class="moz-signature" cols="72">-- ==================================================================== Eric Emsellem <a class="moz-txt-link-abbreviated" href="mailto:ems...@ob...">ems...@ob...</a> Centre de Recherche Astrophysique de Lyon 9 av. Charles-Andre tel: +33 (0)4 78 86 83 84 69561 Saint-Genis Laval Cedex fax: +33 (0)4 78 86 83 86 France <a class="moz-txt-link-freetext" href="http://www-obs.univ-lyon1.fr/eric.emsellem">http://www-obs.univ-lyon1.fr/eric.emsellem</a> ==================================================================== </pre> </body> </html>