从带有图像的扫描 pdf 中提取文本？

Question

我试图从计算机创建的 pdf 中提取文本并且它有效，但我无法从扫描的 pdf 中提取文本，您可以在此处找到，其中包含图像和几页，例如：

这是我使用的代码：

# libraries
## split
from PyPDF2 import PdfFileWriter, PdfFileReader
## read 
import sys
from pdfminer.pdfinterp import PDFResourceManager, PDFPageInterpreter
from pdfminer.pdfpage import PDFPage
from pdfminer.converter import XMLConverter, HTMLConverter, TextConverter
from pdfminer.layout import LAParams
import io
# remove files
import os

# split in case there is several pages
def pdfspliter(filename):
    inputpdf = PdfFileReader(open(filename, "rb"))

    for i in range(inputpdf.numPages):
        output = PdfFileWriter()
        output.addPage(inputpdf.getPage(i))
        with open("document-page%s.pdf" % i, "wb") as outputStream:
            output.write(outputStream)
        pdfparser("document-page%s.pdf" % i)
        os.remove("document-page%s.pdf" % i)

# read a given page
def pdfparser(data):

    fp = open(data, 'rb')
    rsrcmgr = PDFResourceManager()
    retstr = io.StringIO()
    codec = 'utf-8'
    laparams = LAParams()
    device = TextConverter(rsrcmgr, retstr, codec=codec, laparams=laparams)
    # Create a PDF interpreter object.
    interpreter = PDFPageInterpreter(rsrcmgr, device)
    # Process each page contained in the document.

    for page in PDFPage.get_pages(fp):
        interpreter.process_page(page)
        data =  retstr.getvalue()

    print(data)

if __name__ == '__main__':
    filename = sys.argv[1]
    pdfspliter(filename)

你能帮助从这种文件中提取文本吗？

使用 Tesseract OCR 更新

我用 Python 尝试使用 Tesseract OCR，它提取了 pdf 文本的一些页面，但确实需要时间，而且似乎停在了某个点：

# import the necessary packages
from PIL import Image
import pytesseract
import argparse
import cv2
import os
## split
from PyPDF2 import PdfFileWriter, PdfFileReader
# remove
import sys
# 
from pdf2image import convert_from_path
# import all files with a name
import glob

# functions
def pdfspliterimager(filename):
    inputpdf = PdfFileReader(open(filename, "rb"))
    for i in range(inputpdf.numPages):
        output = PdfFileWriter()
        output.addPage(inputpdf.getPage(i))
        with open("document-page%s.pdf" % i, "wb") as outputStream:
            output.write(outputStream)
        pages = convert_from_path("document-page%s.pdf" % i, 500)
        for page in pages:
            page.save('out%s.jpg'%i, 'JPEG')

        os.remove("document-page%s.pdf" % i)

# construct the argument parse and parse the arguments
ap = argparse.ArgumentParser()
ap.add_argument("-i", "--image", required=True,
    help="path to input image to be OCR'd")
ap.add_argument("-p", "--preprocess", type=str, default="thresh",
    help="type of preprocessing to be done")
args = vars(ap.parse_args())

# we test if it is a pdf
image_path = args["image"]
# if it is a pdf we convert it to an image
if image_path.endswith('.pdf'):
    pdfspliterimager(image_path)

# for all files with out in their name
file_names = glob.glob("out*")
for file_name in file_names:
    # load the image and convert it to grayscale
    image = cv2.imread(file_name)
    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)

    # check to see if we should apply thresholding to preprocess the
    # image
    if args["preprocess"] == "thresh":
        gray = cv2.threshold(gray, 0, 255,
            cv2.THRESH_BINARY | cv2.THRESH_OTSU)[1]

    # make a check to see if median blurring should be done to remove
    # noise
    elif args["preprocess"] == "blur":
        gray = cv2.medianBlur(gray, 3)

    # write the grayscale image to disk as a temporary file so we can
    # apply OCR to it
    filename = "{}.png".format(os.getpid())
    cv2.imwrite(filename, gray)

    # load the image as a PIL/Pillow image, apply OCR, and then delete
    # the temporary file
    text = pytesseract.image_to_string(Image.open(filename))
    os.remove(filename)
    print(text)

    # show the output images
    cv2.imshow("Image", image)
    cv2.imshow("Output", gray)
    cv2.waitKey(0)

Answer 1

使用 Python 对 PDF 文件进行 OCR

import os
import io
from PIL import Image
import pytesseract
from wand.image import Image as wi
import gc

def Get_text_from_image(pdf_path):
    pdf=wi(filename=pdf_path,resolution=300)
    pdfImg=pdf.convert('jpeg')
    imgBlobs=[]
    extracted_text=[]
    for img in pdfImg.sequence:
        page=wi(image=img)
        imgBlobs.append(page.make_blob('jpeg'))
    for imgBlob in imgBlobs:
        im=Image.open(io.BytesIO(imgBlob))
        text=pytesseract.image_to_string(im,lang='eng')
        extracted_text.append(text)
    return ([i.replace("\n","") for i in extracted_text])   

我做了一个小修改

下面的代码在代码末尾按顺序将 PDF 的所有页面转换为图像 我破坏了图像序列，因为它需要大量内存来处理

def Get_text_from_image(pdf_path):
    import pytesseract,io,gc
    from PIL import Image
    from wand.image import Image as wi
    import gc
    """ Extracting text content from Image  """

    pdf=wi(filename=pdf_path,resolution=300)                                                                                                                
    pdfImg=pdf.convert('jpeg')                                                                                                                                                                              
    imgBlobs=[]
    extracted_text=[]
    try:        
        for img in pdfImg.sequence:
            page=wi(image=img)
            imgBlobs.append(page.make_blob('jpeg'))
            for i in range(0,5):
                [gc.collect() for i in range(0,10)]

        for imgBlob in imgBlobs:
            im=Image.open(io.BytesIO(imgBlob))
            text=pytesseract.image_to_string(im,lang='eng')
            text = text.replace(r"\n", " ")
            extracted_text.append(text)
            for i in range(0,5):
                [gc.collect() for i in range(0,10)]
        return (''.join([i.replace("\n"," ").replace("\n\n"," ") for i in extracted_text]))
        [gc.collect() for i in range(0,10)]
    finally:
        [gc.collect() for i in range(0,10)]
        img.destroy()