Image.open 给出错误无法识别图像文件

新手上路,请多包涵

我正在尝试使用 PIL 的图像功能打开一个 geotiff 文件。它引发了错误:

 OSError: cannot identify image file 'Whatever\\image\\I\\use.tiff'

例如,我看到 这里 提出的问题,解决方案要么使用

Import Image

代替

From PIL import Image

我认为这是一个过时的解决方案;我无法导入图像。另一个解决方案是将 pillow 更新到 2.9,但 5 年后我们使用的是 5.0.0。我也尝试了 4.0.0,但我收到了同样的错误。是否有最新的解决方案?

这是我的代码, 是文件的链接:

 image_path = 'each\\image\\I\\use.tiff'

from PIL import Image
Image.open(image_path)

原文由 ArnJac 发布,翻译遵循 CC BY-SA 4.0 许可协议

阅读 2.2k
2 个回答

您有一个多通道 32 位 TIFF 图像,而 Pillow 尚不支持该格式。请参阅 问题 #1888

Pillow(和 PIL)目前能够打开每通道 8 位的多通道图像(例如 RGB),但如果它们是单通道(例如,灰度)。

[…]

要求

  • 我们应该能够支持常见的 GIS 格式以及高位深度 RGB(A) 图像。
  • 至少 4 个通道,但可能更多(参见 #1839)
  • 不同的像素格式,包括 I16、I32 和 Float。

我通过直接使用 TIFF 插件图像阅读器确定了这一点,并启用了调试模式:

 from PIL import TiffImagePlugin
TiffImagePlugin.DEBUG = True
with open(image_path, 'rb') as f:
    TiffImagePlugin.TiffImageFile(f)

其中包括输出:

 tag: BitsPerSample (258) - type: short (3) Tag Location: 46 - Data Location: 218 - value: (32, 32, 32)

(下面的完整调试输出)

您可以使用 Python GDAL 绑定 来读取此格式。您还可以使用 gdal_translate 命令行实用程序 将文件转换为 Pillow 可以 处理的格式;对于多波段,您必须降低到 8 位,或者移动到灰度。

例如,要将输入文件转换为 PNG,您可以使用:

 gdal_translate -of PNG rgb_CGI.tiff rgb_CGI.png

之后 Pillow 可以打开 PNG 文件。


Pillow 的 TIFF 插件的完整调试输出:

 >>> from PIL import TiffImagePlugin
>>> TiffImagePlugin.DEBUG = True
>>> with open(image_path, 'rb') as f:
...     TiffImagePlugin.TiffImageFile(f)
...
*** TiffImageFile._open ***
- __first: 8
- ifh:  b'II*\x00\x08\x00\x00\x00'
Seeking to frame 0, on frame -1, __next 8, location: 8
Loading tags, location: 8
tag: ImageWidth (256) - type: short (3) - value: 2924
tag: ImageLength (257) - type: short (3) - value: 2088
tag: BitsPerSample (258) - type: short (3) Tag Location: 46 - Data Location: 218 - value: (32, 32, 32)
tag: Compression (259) - type: short (3) - value: 1
tag: PhotometricInterpretation (262) - type: short (3) - value: 1
tag: StripOffsets (273) - type: long (4) Tag Location: 82 - Data Location: 8576 - value: <table: 8352 bytes>
tag: SamplesPerPixel (277) - type: short (3) - value: 3
tag: RowsPerStrip (278) - type: short (3) - value: 1
tag: StripByteCounts (279) - type: long (4) Tag Location: 118 - Data Location: 224 - value: <table: 8352 bytes>
tag: PlanarConfiguration (284) - type: short (3) - value: 1
tag: ExtraSamples (338) - type: short (3) - value: (0, 0)
tag: SampleFormat (339) - type: short (3) Tag Location: 154 - Data Location: 16928 - value: (2, 2, 2)
tag: ModelPixelScaleTag (33550) - type: double (12) Tag Location: 166 - Data Location: 16934 - value: (0.25, 0.25, 0.0)
tag: ModelTiepointTag (33922) - type: double (12) Tag Location: 178 - Data Location: 16958 - value: <table: 48 bytes>
tag: GeoKeyDirectoryTag (34735) - type: short (3) Tag Location: 190 - Data Location: 17006 - value: <table: 72 bytes>
tag: GeoDoubleParamsTag (34736) - type: double (12) Tag Location: 202 - Data Location: 17078 - value: <table: 56 bytes>
tag: GeoAsciiParamsTag (34737) - type: string (2) Tag Location: 214 - Data Location: 17134 - value: Amersfoort / RD New|Amersfoort|
tag: ImageWidth (256) - type: short (3) - value: 2924
tag: ImageLength (257) - type: short (3) - value: 2088
tag: BitsPerSample (258) - type: short (3) Tag Location: 46 - Data Location: 218 - value: (32, 32, 32)
tag: Compression (259) - type: short (3) - value: 1
tag: PhotometricInterpretation (262) - type: short (3) - value: 1
tag: StripOffsets (273) - type: long (4) Tag Location: 82 - Data Location: 8576 - value: <table: 8352 bytes>
tag: SamplesPerPixel (277) - type: short (3) - value: 3
tag: RowsPerStrip (278) - type: short (3) - value: 1
tag: StripByteCounts (279) - type: long (4) Tag Location: 118 - Data Location: 224 - value: <table: 8352 bytes>
tag: PlanarConfiguration (284) - type: short (3) - value: 1
tag: ExtraSamples (338) - type: short (3) - value: (0, 0)
tag: SampleFormat (339) - type: short (3) Tag Location: 154 - Data Location: 16928 - value: (2, 2, 2)
tag: ModelPixelScaleTag (33550) - type: double (12) Tag Location: 166 - Data Location: 16934 - value: (0.25, 0.25, 0.0)
tag: ModelTiepointTag (33922) - type: double (12) Tag Location: 178 - Data Location: 16958 - value: <table: 48 bytes>
tag: GeoKeyDirectoryTag (34735) - type: short (3) Tag Location: 190 - Data Location: 17006 - value: <table: 72 bytes>
tag: GeoDoubleParamsTag (34736) - type: double (12) Tag Location: 202 - Data Location: 17078 - value: <table: 56 bytes>
tag: GeoAsciiParamsTag (34737) - type: string (2) Tag Location: 214 - Data Location: 17134 - value: Amersfoort / RD New|Amersfoort|
*** Summary ***
- compression: raw
- photometric_interpretation: 1
- planar_configuration: 1
- fill_order: 1
- size: (2924, 2088)
format key: (b'II', 1, (2, 2, 2), 1, (32, 32, 32), (0, 0))
- unsupported format
Traceback (most recent call last):
  File "/Users/mjpieters/Development/venvs/stackoverflow-3.6/lib/python3.6/site-packages/PIL/TiffImagePlugin.py", line 1196, in _setup
    self.mode, rawmode = OPEN_INFO[key]
KeyError: (b'II', 1, (2, 2, 2), 1, (32, 32, 32), (0, 0))

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/mjpieters/Development/venvs/stackoverflow-3.6/lib/python3.6/site-packages/PIL/ImageFile.py", line 102, in __init__
    self._open()
  File "/Users/mjpieters/Development/venvs/stackoverflow-3.6/lib/python3.6/site-packages/PIL/TiffImagePlugin.py", line 950, in _open
    self._seek(0)
  File "/Users/mjpieters/Development/venvs/stackoverflow-3.6/lib/python3.6/site-packages/PIL/TiffImagePlugin.py", line 1017, in _seek
    self._setup()
  File "/Users/mjpieters/Development/venvs/stackoverflow-3.6/lib/python3.6/site-packages/PIL/TiffImagePlugin.py", line 1200, in _setup
    raise SyntaxError("unknown pixel mode")
SyntaxError: unknown pixel mode

原文由 Martijn Pieters 发布,翻译遵循 CC BY-SA 3.0 许可协议

pyvips 可以像这样直接使用 TIFF。它将您的文件加载为具有两个 alpha 通道的单声道 int32 图像。

 $ python3
Python 3.8.2 (default, Apr 27 2020, 15:53:34)
[GCC 9.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import pyvips
>>> x = pyvips.Image.new_from_file("rgb_CGI.tiff")
>>> x.width, x.height, x.bands, x.format, x.interpretation
(2924, 2088, 3, 'int', 'b-w')
>>> # getpoint reads out a pixel
>>> x.getpoint(10, 10)
[156.0, 141.0, 133.0]
>>> # avg finds the image average
>>> x.avg()
125.31475912560515
>>>

它忽略了额外的 GDAL 标签,但您可以进行所有常见的 2D 图像处理,而且它 比 PIL 快很多

原文由 jcupitt 发布,翻译遵循 CC BY-SA 4.0 许可协议

撰写回答
你尚未登录,登录后可以
  • 和开发者交流问题的细节
  • 关注并接收问题和回答的更新提醒
  • 参与内容的编辑和改进,让解决方法与时俱进
推荐问题