How might one extract all images from a pdf document, at native resolution and format? (Meaning extract tiff as tiff, jpeg as jpeg, etc. and without resampling). Layout is unimportant, I don't care were the source image is located on the page.
I'm using python 2.7 but can use 3.x if required.
Often in a PDF, the image is simply stored as-is. For example, a PDF with a jpg inserted will have a range of bytes somewhere in the middle that when extracted is a valid jpg file. You can use this to very simply extract byte ranges from the PDF. I wrote about this some time ago, with sample code: Extracting JPGs from PDFs.
In Python with PyPDF2 and Pillow libraries it is simple:
import PyPDF2
from PIL import Image
if __name__ == '__main__':
input1 = PyPDF2.PdfFileReader(open("input.pdf", "rb"))
page0 = input1.getPage(0)
xObject = page0['/Resources']['/XObject'].getObject()
for obj in xObject:
if xObject[obj]['/Subtype'] == '/Image':
size = (xObject[obj]['/Width'], xObject[obj]['/Height'])
data = xObject[obj].getData()
if xObject[obj]['/ColorSpace'] == '/DeviceRGB':
mode = "RGB"
else:
mode = "P"
if xObject[obj]['/Filter'] == '/FlateDecode':
img = Image.frombytes(mode, size, data)
img.save(obj[1:] + ".png")
elif xObject[obj]['/Filter'] == '/DCTDecode':
img = open(obj[1:] + ".jpg", "wb")
img.write(data)
img.close()
elif xObject[obj]['/Filter'] == '/JPXDecode':
img = open(obj[1:] + ".jp2", "wb")
img.write(data)
img.close()
In Python with PyPDF2 for CCITTFaxDecode filter:
import PyPDF2
import struct
"""
Links:
PDF format: http://www.adobe.com/content/dam/Adobe/en/devnet/acrobat/pdfs/pdf_reference_1-7.pdf
CCITT Group 4: https://www.itu.int/rec/dologin_pub.asp?lang=e&id=T-REC-T.6-198811-I!!PDF-E&type=items
Extract images from pdf: http://stackoverflow.com/questions/2693820/extract-images-from-pdf-without-resampling-in-python
Extract images coded with CCITTFaxDecode in .net: http://stackoverflow.com/questions/2641770/extracting-image-from-pdf-with-ccittfaxdecode-filter
TIFF format and tags: http://www.awaresystems.be/imaging/tiff/faq.html
"""
def tiff_header_for_CCITT(width, height, img_size, CCITT_group=4):
tiff_header_struct = '<' + '2s' + 'h' + 'l' + 'h' + 'hhll' * 8 + 'h'
return struct.pack(tiff_header_struct,
b'II', # Byte order indication: Little indian
42, # Version number (always 42)
8, # Offset to first IFD
8, # Number of tags in IFD
256, 4, 1, width, # ImageWidth, LONG, 1, width
257, 4, 1, height, # ImageLength, LONG, 1, lenght
258, 3, 1, 1, # BitsPerSample, SHORT, 1, 1
259, 3, 1, CCITT_group, # Compression, SHORT, 1, 4 = CCITT Group 4 fax encoding
262, 3, 1, 0, # Threshholding, SHORT, 1, 0 = WhiteIsZero
273, 4, 1, struct.calcsize(tiff_header_struct), # StripOffsets, LONG, 1, len of header
278, 4, 1, height, # RowsPerStrip, LONG, 1, lenght
279, 4, 1, img_size, # StripByteCounts, LONG, 1, size of image
0 # last IFD
)
pdf_filename = 'scan.pdf'
pdf_file = open(pdf_filename, 'rb')
cond_scan_reader = PyPDF2.PdfFileReader(pdf_file)
for i in range(0, cond_scan_reader.getNumPages()):
page = cond_scan_reader.getPage(i)
xObject = page['/Resources']['/XObject'].getObject()
for obj in xObject:
if xObject[obj]['/Subtype'] == '/Image':
"""
The CCITTFaxDecode filter decodes image data that has been encoded using
either Group 3 or Group 4 CCITT facsimile (fax) encoding. CCITT encoding is
designed to achieve efficient compression of monochrome (1 bit per pixel) image
data at relatively low resolutions, and so is useful only for bitmap image data, not
for color images, grayscale images, or general data.
K < 0 --- Pure two-dimensional encoding (Group 4)
K = 0 --- Pure one-dimensional encoding (Group 3, 1-D)
K > 0 --- Mixed one- and two-dimensional encoding (Group 3, 2-D)
"""
if xObject[obj]['/Filter'] == '/CCITTFaxDecode':
if xObject[obj]['/DecodeParms']['/K'] == -1:
CCITT_group = 4
else:
CCITT_group = 3
width = xObject[obj]['/Width']
height = xObject[obj]['/Height']
data = xObject[obj]._data # sorry, getData() does not work for CCITTFaxDecode
img_size = len(data)
tiff_header = tiff_header_for_CCITT(width, height, img_size, CCITT_group)
img_name = obj[1:] + '.tiff'
with open(img_name, 'wb') as img_file:
img_file.write(tiff_header + data)
#
# import io
# from PIL import Image
# im = Image.open(io.BytesIO(tiff_header + data))
pdf_file.close()
You can use the module PyMuPDF. This outputs all images as .png files, but worked out of the box and is fast.
import fitz
doc = fitz.open("file.pdf")
for i in range(len(doc)):
for img in doc.getPageImageList(i):
xref = img[0]
pix = fitz.Pixmap(doc, xref)
if pix.n < 5: # this is GRAY or RGB
pix.writePNG("p%s-%s.png" % (i, xref))
else: # CMYK: convert to RGB first
pix1 = fitz.Pixmap(fitz.csRGB, pix)
pix1.writePNG("p%s-%s.png" % (i, xref))
pix1 = None
pix = None
Libpoppler comes with a tool called "pdfimages" that does exactly this.
(On ubuntu systems it's in the poppler-utils package)
http://poppler.freedesktop.org/
http://en.wikipedia.org/wiki/Pdfimages
Windows binaries: http://blog.alivate.com.au/poppler-windows/
I started from the code of @sylvain
There was some flaws, like the exception NotImplementedError: unsupported filter /DCTDecode of getData, or the fact the code failed to find images in some pages because they were at a deeper level than the page.
There is my code :
import PyPDF2
from PIL import Image
import sys
from os import path
import warnings
warnings.filterwarnings("ignore")
number = 0
def recurse(page, xObject):
global number
xObject = xObject['/Resources']['/XObject'].getObject()
for obj in xObject:
if xObject[obj]['/Subtype'] == '/Image':
size = (xObject[obj]['/Width'], xObject[obj]['/Height'])
data = xObject[obj]._data
if xObject[obj]['/ColorSpace'] == '/DeviceRGB':
mode = "RGB"
else:
mode = "P"
imagename = "%s - p. %s - %s"%(abspath[:-4], p, obj[1:])
if xObject[obj]['/Filter'] == '/FlateDecode':
img = Image.frombytes(mode, size, data)
img.save(imagename + ".png")
number += 1
elif xObject[obj]['/Filter'] == '/DCTDecode':
img = open(imagename + ".jpg", "wb")
img.write(data)
img.close()
number += 1
elif xObject[obj]['/Filter'] == '/JPXDecode':
img = open(imagename + ".jp2", "wb")
img.write(data)
img.close()
number += 1
else:
recurse(page, xObject[obj])
try:
_, filename, *pages = sys.argv
*pages, = map(int, pages)
abspath = path.abspath(filename)
except BaseException:
print('Usage :\nPDF_extract_images file.pdf page1 page2 page3 …')
sys.exit()
file = PyPDF2.PdfFileReader(open(filename, "rb"))
for p in pages:
page0 = file.getPage(p-1)
recurse(p, page0)
print('%s extracted images'% number)
I prefer minecart as it is extremely easy to use. The below snippet show how to extract images from a pdf:
#pip install minecart
import minecart
pdffile = open('Invoices.pdf', 'rb')
doc = minecart.Document(pdffile)
page = doc.get_page(0) # getting a single page
#iterating through all pages
for page in doc.iter_pages():
im = page.images[0].as_pil() # requires pillow
display(im)
After some searching I found the following script which works really well with my PDF's. It does only tackle JPG, but it worked perfectly with my unprotected files. Also is does not require any outside libraries.
Not to take any credit, the script originates from Ned Batchelder, and not me. Python3 code: extract jpg's from pdf's. Quick and dirty
import sys
with open(sys.argv[1],"rb") as file:
file.seek(0)
pdf = file.read()
startmark = b"\xff\xd8"
startfix = 0
endmark = b"\xff\xd9"
endfix = 2
i = 0
njpg = 0
while True:
istream = pdf.find(b"stream", i)
if istream < 0:
break
istart = pdf.find(startmark, istream, istream + 20)
if istart < 0:
i = istream + 20
continue
iend = pdf.find(b"endstream", istart)
if iend < 0:
raise Exception("Didn't find end of stream!")
iend = pdf.find(endmark, iend - 20)
if iend < 0:
raise Exception("Didn't find end of JPG!")
istart += startfix
iend += endfix
print("JPG %d from %d to %d" % (njpg, istart, iend))
jpg = pdf[istart:iend]
with open("jpg%d.jpg" % njpg, "wb") as jpgfile:
jpgfile.write(jpg)
njpg += 1
i = iend
I installed ImageMagick on my server and then run commandline-calls through Popen:
#!/usr/bin/python
import sys
import os
import subprocess
import settings
IMAGE_PATH = os.path.join(settings.MEDIA_ROOT , 'pdf_input' )
def extract_images(pdf):
output = 'temp.png'
cmd = 'convert ' + os.path.join(IMAGE_PATH, pdf) + ' ' + os.path.join(IMAGE_PATH, output)
subprocess.Popen(cmd.split(), stderr=subprocess.STDOUT, stdout=subprocess.PIPE)
This will create an image for every page and store them as temp-0.png, temp-1.png .... This is only 'extraction' if you got a pdf with only images and no text.
Much easier solution:
Use the poppler-utils package. To install it use homebrew (homebrew is MacOS specific, but you can find the poppler-utils package for Widows or Linux here: https://poppler.freedesktop.org/). First line of code below installs poppler-utils using homebrew. After installation the second line (run from the command line) then extracts images from a PDF file and names them "image*". To run this program from within Python use the os or subprocess module. Third line is code using os module, beneath that is an example with subprocess (python 3.5 or later for run() function). More info here: https://www.cyberciti.biz/faq/easily-extract-images-from-pdf-file/
brew install poppler
pdfimages file.pdf image
import os
os.system('pdfimages file.pdf image')
or
import subprocess
subprocess.run('pdfimages file.pdf image', shell=True)
You could use pdfimages command in Ubuntu as well.
Install poppler lib using the below commands.
sudo apt install poppler-utils
sudo apt-get install python-poppler
pdfimages file.pdf image
List of files created are, (for eg.,. there are two images in pdf)
image-000.png
image-001.png
It works ! Now you can use a subprocess.run to run this from python.
As of February 2019, the solution given by @sylvain (at least on my setup) does not work without a small modification: xObject[obj]['/Filter'] is not a value, but a list, thus in order to make the script work, I had to modify the format checking as follows:
import PyPDF2, traceback
from PIL import Image
input1 = PyPDF2.PdfFileReader(open(src, "rb"))
nPages = input1.getNumPages()
print nPages
for i in range(nPages) :
print i
page0 = input1.getPage(i)
try :
xObject = page0['/Resources']['/XObject'].getObject()
except : xObject = []
for obj in xObject:
if xObject[obj]['/Subtype'] == '/Image':
size = (xObject[obj]['/Width'], xObject[obj]['/Height'])
data = xObject[obj].getData()
try :
if xObject[obj]['/ColorSpace'] == '/DeviceRGB':
mode = "RGB"
elif xObject[obj]['/ColorSpace'] == '/DeviceCMYK':
mode = "CMYK"
# will cause errors when saving
else:
mode = "P"
fn = 'p%03d-%s' % (i + 1, obj[1:])
print '\t', fn
if '/FlateDecode' in xObject[obj]['/Filter'] :
img = Image.frombytes(mode, size, data)
img.save(fn + ".png")
elif '/DCTDecode' in xObject[obj]['/Filter']:
img = open(fn + ".jpg", "wb")
img.write(data)
img.close()
elif '/JPXDecode' in xObject[obj]['/Filter'] :
img = open(fn + ".jp2", "wb")
img.write(data)
img.close()
elif '/LZWDecode' in xObject[obj]['/Filter'] :
img = open(fn + ".tif", "wb")
img.write(data)
img.close()
else :
print 'Unknown format:', xObject[obj]['/Filter']
except :
traceback.print_exc()
I added all of those together in PyPDFTK here.
My own contribution is handling of /Indexed files as such:
for obj in xObject:
if xObject[obj]['/Subtype'] == '/Image':
size = (xObject[obj]['/Width'], xObject[obj]['/Height'])
color_space = xObject[obj]['/ColorSpace']
if isinstance(color_space, pdf.generic.ArrayObject) and color_space[0] == '/Indexed':
color_space, base, hival, lookup = [v.getObject() for v in color_space] # pg 262
mode = img_modes[color_space]
if xObject[obj]['/Filter'] == '/FlateDecode':
data = xObject[obj].getData()
img = Image.frombytes(mode, size, data)
if color_space == '/Indexed':
img.putpalette(lookup.getData())
img = img.convert('RGB')
img.save("{}{:04}.png".format(filename_prefix, i))
Note that when /Indexed files are found, you can't just compare /ColorSpace to a string, because it comes as an ArrayObject. So, we have to check the array and retrieve the indexed palette (lookup in the code) and set it in the PIL Image object, otherwise it stays uninitialized (zero) and the whole image shows as black.
My first instinct was to save them as GIFs (which is an indexed format), but my tests turned out that PNGs were smaller and looked the same way.
I found those types of images when printing to PDF with Foxit Reader PDF Printer.
Here is my version from 2019 that recursively gets all images from PDF and reads them with PIL. Compatible with Python 2/3. I also found that sometimes image in PDF may be compressed by zlib, so my code supports decompression.
#!/usr/bin/env python3
try:
from StringIO import StringIO
except ImportError:
from io import BytesIO as StringIO
from PIL import Image
from PyPDF2 import PdfFileReader, generic
import zlib
def get_color_mode(obj):
try:
cspace = obj['/ColorSpace']
except KeyError:
return None
if cspace == '/DeviceRGB':
return "RGB"
elif cspace == '/DeviceCMYK':
return "CMYK"
elif cspace == '/DeviceGray':
return "P"
if isinstance(cspace, generic.ArrayObject) and cspace[0] == '/ICCBased':
color_map = obj['/ColorSpace'][1].getObject()['/N']
if color_map == 1:
return "P"
elif color_map == 3:
return "RGB"
elif color_map == 4:
return "CMYK"
def get_object_images(x_obj):
images = []
for obj_name in x_obj:
sub_obj = x_obj[obj_name]
if '/Resources' in sub_obj and '/XObject' in sub_obj['/Resources']:
images += get_object_images(sub_obj['/Resources']['/XObject'].getObject())
elif sub_obj['/Subtype'] == '/Image':
zlib_compressed = '/FlateDecode' in sub_obj.get('/Filter', '')
if zlib_compressed:
sub_obj._data = zlib.decompress(sub_obj._data)
images.append((
get_color_mode(sub_obj),
(sub_obj['/Width'], sub_obj['/Height']),
sub_obj._data
))
return images
def get_pdf_images(pdf_fp):
images = []
try:
pdf_in = PdfFileReader(open(pdf_fp, "rb"))
except:
return images
for p_n in range(pdf_in.numPages):
page = pdf_in.getPage(p_n)
try:
page_x_obj = page['/Resources']['/XObject'].getObject()
except KeyError:
continue
images += get_object_images(page_x_obj)
return images
if __name__ == "__main__":
pdf_fp = "test.pdf"
for image in get_pdf_images(pdf_fp):
(mode, size, data) = image
try:
img = Image.open(StringIO(data))
except Exception as e:
print ("Failed to read image with PIL: {}".format(e))
continue
# Do whatever you want with the image
来源：https://stackoverflow.com/questions/2693820/extract-images-from-pdf-without-resampling-in-python