这是我第一次处理多维数组,我在访问元素时遇到了问题。我正在尝试获取图片的红色像素,但只是数组中的前 8 个元素。这是代码
import Image
import numpy as np
im = Image.open("C:\Users\Jones\Pictures\1.jpg")
pix = im.load()
r, g, b = np.array(im).T
print r[0:8]
由于您正在处理图像,r
因此是一个二维数组。要获取图像中的前 8 个像素,请尝试
r.flatten()[:8]
如果第一行少于 8 个像素,这将自动环绕。
你可以这样做:
r[0][:8]
但是请注意,如果第一行少于 8 个像素,这将不起作用。要解决此问题,请执行以下操作:
from itertools import chain
r = list(chain.from_iterable(r))
r[:8]
或(如果您不想import
使用整个模块):
r = [val for element in r for val in element]
r[:8]
你也想要所有的行吗?试试这个r[:,:8]
只想要第一行?试试这个r[0,:8]
我认为它可以更简单。此示例使用随机矩阵(这将是您的r
矩阵):
In [7]: from pylab import * # convention
In [8]: r = randint(0,10,(10,10)) # this is your image
In [9]: r
array([[7, 9, 5, 5, 6, 8, 1, 4, 3, 4],
[5, 4, 4, 4, 2, 6, 2, 6, 4, 2],
[1, 4, 9, 9, 2, 6, 1, 9, 0, 6],
[5, 9, 0, 7, 9, 9, 5, 2, 0, 7],
[8, 3, 3, 9, 0, 0, 5, 9, 2, 2],
[5, 3, 7, 8, 8, 1, 6, 3, 2, 0],
[0, 2, 5, 7, 0, 1, 0, 2, 1, 2],
[4, 0, 4, 5, 9, 9, 3, 8, 3, 7],
[4, 6, 9, 9, 5, 9, 3, 0, 5, 1],
[6, 9, 9, 0, 3, 4, 9, 7, 9, 6]])
然后,提取前 8 列并做一些事情
In [17]: r_8 = r[:,:8] # extract columns
In [18]: r_8
Out[18]:
array([[7, 9, 5, 5, 6, 8, 1, 4],
[5, 4, 4, 4, 2, 6, 2, 6],
[1, 4, 9, 9, 2, 6, 1, 9],
[5, 9, 0, 7, 9, 9, 5, 2],
[8, 3, 3, 9, 0, 0, 5, 9],
[5, 3, 7, 8, 8, 1, 6, 3],
[0, 2, 5, 7, 0, 1, 0, 2],
[4, 0, 4, 5, 9, 9, 3, 8],
[4, 6, 9, 9, 5, 9, 3, 0],
[6, 9, 9, 0, 3, 4, 9, 7]])
In [19]: r_8 = r_8 * 2 # do something
In [20]: r_8
Out[20]:
array([[14, 18, 10, 10, 12, 16, 2, 8],
[10, 8, 8, 8, 4, 12, 4, 12],
[ 2, 8, 18, 18, 4, 12, 2, 18],
[10, 18, 0, 14, 18, 18, 10, 4],
[16, 6, 6, 18, 0, 0, 10, 18],
[10, 6, 14, 16, 16, 2, 12, 6],
[ 0, 4, 10, 14, 0, 2, 0, 4],
[ 8, 0, 8, 10, 18, 18, 6, 16],
[ 8, 12, 18, 18, 10, 18, 6, 0],
[12, 18, 18, 0, 6, 8, 18, 14]])
现在,这就是诀窍。r
替换using中的前 8 列hstack
:
In [21]: r = hstack((r_8, r[:,8:])) # it replaces the FISRT 8 columns, note the indexing notation
In [22]: r
Out[22]:
array([[14, 18, 10, 10, 12, 16, 2, 8, 3, 4], # it does not touch the last 2 columns
[10, 8, 8, 8, 4, 12, 4, 12, 4, 2],
[ 2, 8, 18, 18, 4, 12, 2, 18, 0, 6],
[10, 18, 0, 14, 18, 18, 10, 4, 0, 7],
[16, 6, 6, 18, 0, 0, 10, 18, 2, 2],
[10, 6, 14, 16, 16, 2, 12, 6, 2, 0],
[ 0, 4, 10, 14, 0, 2, 0, 4, 1, 2],
[ 8, 0, 8, 10, 18, 18, 6, 16, 3, 7],
[ 8, 12, 18, 18, 10, 18, 6, 0, 5, 1],
[12, 18, 18, 0, 6, 8, 18, 14, 9, 6]])
编辑:至于 DSM 所指出的,OP 实际上是使用numpy array。
我收回我的答案,因为 nneonneo 是正确的