读取pdf图片并识别内容

import fitz
import time
import re
import os

from PIL import Image
from aip import AipOcr

import numpy as np


def pdf2pic(path, pic_path):
    '''
    # 从pdf中提取图片
    :param path: pdf的路径
    :param pic_path: 图片保存的路径
    :return:
    '''
    t0 = time.perf_counter()
    # 使用正则表达式来查找图片
    checkXO = r"/Type(?= */XObject)" 
    checkIM = r"/Subtype(?= */Image)"  


    # 打开pdf
    doc = fitz.open(path)
    # 图片计数
    imgcount = 0
    lenXREF = doc.xref_length()
 
    # 打印PDF的信息
    print("文件名:{}, 页数: {}, 对象: {}".format(path, len(doc), lenXREF - 1))

        # 遍历每一个对象
    for i in range(1, lenXREF):
        # 定义对象字符串
        text = doc.xref_object(i)
        isXObject = re.search(checkXO, text)
        # 使用正则表达式查看是否是图片
        isImage = re.search(checkIM, text)
        # 如果不是对象也不是图片，则continue
        if not isXObject or not isImage:
            continue
        imgcount += 1
        # 根据索引生成图像
        pix = fitz.Pixmap(doc, i)
        # 根据pdf的路径生成图片的名称
        new_name = f"{imgcount}.png"


        # 如果pix.n<5,可以直接存为PNG
        if pix.n < 5:
            pix.writePNG(os.path.join(pic_path, new_name))
        # 否则先转换CMYK
        else:
            pix0 = fitz.Pixmap(fitz.csRGB, pix)
            pix0.writePNG(os.path.join(pic_path, new_name))
            pix0 = None
        # 释放资源
        pix = None
        t1 = time.perf_counter()
        print("运行时间:{}s".format(t1 - t0))
        print("提取了{}张图片".format(imgcount))

class Tap_img:

	APP_ID = ''
	API_KEY = ''
	SECRET_KEY = ''

	def ResizeImage(self,img_path):
	    '''改变图片尺寸'''
	    filein = img_path
	    fileout = img_path
	    width = 500
	    height = 900
	    img = Image.open(filein)
	    out = img.resize((width, height),Image.ANTIALIAS)
	    out.save(fileout)
	    img.close()


	def get_file_content(self,filepath):
	    '''读取图片内容'''
	    
	    client = AipOcr(self.APP_ID, self.API_KEY, self.SECRET_KEY)
	    with open(filepath, 'rb') as fp:
	        image = fp.read()
	    fp.close()
	    # 定义参数变量
	    options = {
	        # 定义图像方向
	        'detect-direction': 'true',
	        'language-type': 'CHN_ENG'
	    }
	    result = client.general(image, options)
	    content_list=[]
	    for word in result['words_result']:
	        content_list.append(word['words'])

	    res=''.join(content_list).split('%')[:-1]

	    res_dict = {}
	    for x in res:
	    	key=re.findall('[\u4e99-\u9fa5]+',x)[0]
	    	value=re.findall('\d+',x)[0]
	    	res_dict[key]=value+'%'


	    return res_dict


if __name__=='__main__':
    # 提取pdf的图片保存在本地
    path = '1.pdf'
    pic_path = './img'
    pdf2pic(path, pic_path)

    
    ###将图片内容提取出来
    img=Tap_img()
    content=img.get_file_content('img/2.png')
    print(content)