Matplotlib: Quiver 2D数组

Question

我在使用Python的quiver函数时遇到了一些困难。

我想用quiver来创建一个二维的向量场，这个向量场是由一个包含两个向量分量的二维数组生成的。

给定数组b：

>>> b
VigraArray(shape=(512, 512, 2), axistags=x y c, dtype=float32, data=
[[[ 0.59471679  0.51902866  0.38904327 ..., -0.56878477 -0.50834674
 -0.48382956]
[ 0.58222073  0.50713873  0.37990916 ..., -0.56091702 -0.50057167
 -0.47613338]
[ 0.53815156  0.46551338  0.34787226 ..., -0.54245669 -0.48314109
 -0.45911735]
...,

然后使用quiver：

>>> X,Y = meshgrid(range(b.shape[0]),range(b.shape[1]))  
>>> quiver(X,Y,b[...,0],b[...,1])

返回结果是：

Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/usr/lib/pymodules/python2.7/matplotlib/pyplot.py", line 2892, in quiver
  ret = ax.quiver(*args, **kw)
File "/usr/lib/pymodules/python2.7/matplotlib/axes.py", line 6641, in quiver
  q = mquiver.Quiver(self, *args, **kw)
File "/usr/lib/pymodules/python2.7/matplotlib/quiver.py", line 419, in __init__
self.set_UVC(U, V, C)
File "/usr/lib/pymodules/python2.7/matplotlib/quiver.py", line 463, in set_UVC
  U = ma.masked_invalid(U, copy=False).ravel()
File "/usr/lib/python2.7/dist-packages/numpy/ma/core.py", line 3969, in ravel
  r._mask = ndarray.ravel(self._mask).reshape(r.shape)
File "/usr/lib/python2.7/dist-packages/vigra/arraytypes.py", line 1308, in reshape
  res = numpy.ndarray.reshape(self, shape, order)
TypeError: an integer is required

我之前在使用VigraArray和matplotlib时从来没有遇到过问题，所以我觉得这不是问题所在。谢谢大家的帮助！

Answer 1

实际上，似乎 pylab.quiver 不能正确处理 vigra.VigraArray。我找到了一种替代方法，可以正常工作：

b[...,0] = numpy.array(b[...,0])
b[...,1] = numpy.array(b[...,1])