python - 尝试使用噪声、numpy 和 Image 创建噪声图像

Question

我正在尝试在 python 中创建一些柏林噪声图像，但遇到了一些问题。当我运行我的小脚本时出现异常。我显然不了解 Image 模块的使用，因为我尝试的所有操作都会导致 ValueError 异常，并显示消息“缓冲区不够大”

这是我到目前为止所得到的:

import numpy
from noise import pnoise2, snoise2
import Image


octaves = 1
freq = 16.0 * octaves
y_max = 5
x_max = 5
imarray = [[0 for x in range(y_max)] for x in range(x_max)]
totalcount = 0

for y in range(y_max):
    for x in range(x_max):
        val = "%s\n" % int(snoise2(x / freq, y / freq, octaves) * 127.0 + 128.0)
        imarray[y][x] = val
        totalcount += 1

arr = numpy.asarray(imarray).astype('uint8')

im = Image.fromarray(arr, 'RGBA')
im.save('./blah.png')

我遇到的异常是:

Connected to pydev debugger (build 143.1184)
Traceback (most recent call last):
  File "/Applications/PyCharm CE.app/Contents/helpers/pydev/pydevd.py", line 2407, in <module>
    globals = debugger.run(setup['file'], None, None, is_module)
  File "/Applications/PyCharm CE.app/Contents/helpers/pydev/pydevd.py", line 1798, in run
    launch(file, globals, locals)  # execute the script
  File "/Users/u4234/source/systools/load/noise_test.py", line 26, in <module>
    im = Image.fromarray(arr, 'RGBA')
  File "/Users/u4234/Library/Python/2.7/lib/python/site-packages/PIL-1.1.7-py2.7-macosx-10.10-x86_64.egg/Image.py", line 1902, in fromarray
    return frombuffer(mode, size, obj, "raw", rawmode, 0, 1)
  File "/Users/u4234/Library/Python/2.7/lib/python/site-packages/PIL-1.1.7-py2.7-macosx-10.10-x86_64.egg/Image.py", line 1853, in frombuffer
    core.map_buffer(data, size, decoder_name, None, 0, args)
ValueError: buffer is not large enough

Process finished with exit code 1

Answer 1

问题是您要求 PIL 创建一个新的 'RGBA' 图像，该图像需要 4 个 channel ，但您只传递一个二维数组，例如一个 channel 。

如下所示，给定这些数组:

arr1 = numpy.zeros((5, 5), dtype=numpy.uint8)  # A single 5x5 channel
arr2 = numpy.zeros((5, 5, 3), dtype=numpy.uint8)  # 3 5x5 channels

创建灰度'L'图像:

im = Image.fromarray(arr1, 'L')  # OK
im = Image.fromarray(arr2, 'L')  # ValueError: Too many dimensions.

创建没有 Alpha channel 的彩色图像'RGB':

im = Image.fromarray(arr1, 'RGB')  # ValueError: not enough image data
im = Image.fromarray(arr2, 'RGB')  # OK

请注意RGB模式的使用，因为您可能不关心Alpha channel ，否则只需将其他维度添加到numpy.zeros构造函数即可。

<小时/>

您可以将代码重写为，

arr = numpy.zeros((5, 5, 1), dtype=numpy.uint8)

y_max, x_max, channels = arr.shape

for y in range(y_max):
    for x in range(x_max):
        val = int(snoise2(x / freq, y / freq, octaves) * 127.0 + 128.0)
        arr[y,x,0] = val

im = Image.fromarray(arr, 'L')
im.save('./blah.png')

生成灰度图像。