目录
- 配置需求
- 提取出票日期
- 提取金额
- 提取出票人
- 提取付款行
- 提取汇票到账日期
- 提取票据单据
- 附上源码全部
每天面对成堆的发票,无论是发票还是承兑单据,抑或是其他各类公司数据要从照片、PDF等不同格式的内容中提取,我们都有必要进行快速办公的能力提升。
因此,我们的目标要求就十分明显了,首先要从图片中获取数据,其次将数据统一导入到EXCEL中。
配置需求
1.ImageMagick
2.tesseract-OCR
3.python3.7
4.from PIL import Image as PI
5.import io
6.imjsport os
7.import pyocr.builders
8.from cnocr import CnOcr
9.import xlwt
分析上图发现票据金额为“贰拾万元整”,数据金额为大写中文,因此在导入Excel之前我们需要将金额票据的数据转换成数php字的格式,基于此,我们需要首先完成大写汉字和数字的转换。
def chineseNumber2Int(strNum: str): result = 0 temp = 1 # 存放一个单位的数字如:十万 count = 0 # 判断是否有chArr cnArr = ['壹', '贰', '叁', '肆', '伍', '陆', '柒', '捌', '玖'] chArr = ['拾', '佰', '仟', '万', '亿'] for i in range(len(strNum)): b = True c = strNum[i] for j in range(len(cnArr)): if c == cnArr[j]: if count != 0: result += temp count = 0 temp = j + 1 b = False break if b: for j in range(len(chArr)): if c == chArr[j]: if j == 0: temp *= 10 elif j == 1: temp *= 100 elif j == 2: temp *= 1000 elif j == 3: temp *= 10000 elif j == 4: temp *= 100000000 count += 1 if i == len(strNum) - 1: result += temp return result
通过上述代码即可实现大写字母与数字的转换,例如输入“贰拾万元整”即可导出“200000”,再将其转换成数字后即可极大地简化表格的操作,也可以在完成表格操作的同时有利于数据归档。
接下来,我们需要分析发票的内部内容,分析下图可知,我们需要获取以下几个数据内容:“出票日期”、“汇票到账日期SZAeztB”、“票据号码”、“收款人”、“票据金额”、“出票人”,可以通过画图软件获取精准定位。
如图,小黑点即鼠标所在地,画图软件左下角即他的坐标。
提取出票日期
def text1(new_img): #提取出票日期 left = 80 top = 143 right = 162 bottom = 162 image_text1 = new_img.crop((left, top, right, bottom)) #展示图片 #image_text1.show() txt1 = tool.image_to_string(image_text1) print(txt1) retuSZAeztBrn str(txt1)
提取金额
def text2(new_img): #提取金额 left = 224 top = 355 right = 585 bottom = 380 image_text2 = new_img.crop((left, top, right, bottom)) #展示图片 #image_text2.show() image_text2.save("img/tmp.png") temp = ocr.ocr("img/tmp.png")js temp="".join(temp[0]) txt2=chineseNumber2Int(temp) print(txt2) return txt2
提取出票人
def text3(new_img): #提取出票人 left = 177 top = 207 right = 506 bottom = 231 image_text3 = new_img.crop((left, top, right, bottom)) #展示图片 #image_text3.show() image_text3.save("img/tmp.png") temp = ocr.ocr("img/tmp.png") txt3="".join(temp[0]) print(txt3) return txt3
提取付款行
def text4(new_img): #提取付款行 left = 177 top = 274 right = 492 bottom = 311 image_text4 = new_img.crop((left, top, right, bottom)) #展示图片 #image_text4.show() image_text4.save("img/tmp.png") temp = ocr.ocr("img/tmp.png") txt4="".join(temp[0]) print(txt4) return txt4
提取汇票到账日期
def text5(new_img): #提取汇票到日期 left = 92 top = 166 right = 176 bottom = 184 image_text5 = new_img.crop((left, top, right, bottom)) #展示图片 #image_text5.show() txt5 = tool.image_to_string(image_text5) print(txt5) return txt5
提取票据单据
def text6(new_img): #提取票据号码 left = 598 top = 166 right = 870 bottom = 182 image_text6 = new_img.crop((left, top, right, bottom)) #展示图片 #image_text6.show() txt6 = tool.image_to_string(image_text6) print(txt6) return txt6
在将数据全部提取完成之后,即进入设置环节,我们需要首先将所有账单文件进行提取,获取他们的文件名和路径。
ocr=CnOcr() tool = pyocr.get_available_tools()[0] filePath='img' img_name=[] for i,j,name in os.walk(filePath): img_name=name
在获取完整后,即可进行数据导入Excel的操作。
count=1 book = xlwt.Workbook(encoding='utf-8',style_compression=0) sheet = book.add_sheet('test',cell_overwrite_ok=True) for i in img_name: img_url = filePath+"/"+i with open(img_url, 'rb') as f: a = f.read() new_img = PI.open(io.BytesIO(a)) ## 写入csv col = ('年份','出票日期','金额','出票人','付款行全称','汇票到日期','备注') for j in range(0,7): sheet.write(0,j,col[j]) book.save('1.csv') shijian=text1(new_img) sheet.write(count,0,shijian[0:4]) sheet.write(count,1,shijian[5:]) sheet.write(count,2,text2(new_img)) sheet.write(count,3,text3(new_img)) sheet.write(count,4,text4(new_img)) sheet.write(count,5,text5(new_img)) sheet.write(count,6,text6(new_img)) count = count + 1
至此,完整流程结束。
附上源码全部
from wand.image import Image from PIL import Image as PI import pyocr import io import re import os import shutil import pyocr.builders from cnocr import CnOcr import requests import xlrd import xlwt from openpyxl import load_workbook def chineseNumber2Int(strNum: str): result = 0 temp = 1 # 存放一个单位的数字如:十万 count = 0 # 判断是否有chArr cnArr = ['壹', '贰', '叁', '肆', '伍', '陆', '柒', '捌', '玖'] chArr = ['拾', '佰', '仟', '万', '亿'] for i in range(len(strNum)): b = True c = strNum[i] for j in range(len(cnArr)): if c == cnArr[j]: if count != 0: result += temp count = 0 temp = j + 1 b = False break if b: for j in range(len(chArr)): if c == chArr[j]: if j == 0: temp *= 10 elif j == 1: temp *= 100 elif j == 2: temp *= 1000 elif j == 3: temp *= 10000 elif j == 4: temp *= 100000000 count += 1 if i == len(strNum) - 1: result += temp return result def text1(new_img): #提取出票日期 left = 80 top = 143 right = 162 bottom = 162 image_text1 = new_img.crop((left, top, right, bottom)) #展示图片 #image_text1.show() txt1 = tool.image_to_string(image_text1) print(txt1) return str(txt1) def text2(new_img): #提取金额 left = 224 top = 355 right = 585 bottom = 380 image_text2 = new_img.crop((left, top, right, bottom)) #展示图片 #image_text2.show() image_text2.save("img/tmp.png") temp = ocr.ocr("img/tmp.png") temp="".join(temp[0]) txt2=chineseNumber2Int(temp) print(txt2) return txt2 def text3(new_img): #提取出票人 left = 177 top = 207 right = 506 bottom = 231 image_text3 = new_img.crop((left, top, right, bottom)) #展示图片 #image_text3.show() image_text3.save("img/tmp.png") temp = ocr.ocr("img/tmp.png") txt3="".join(temp[0]) print(txt3) return txt3 def text4(new_img): #提取付款行 left = 177 top = 274 right = 492 bottom = 311 image_text4 = new_img.crop((left, top, right, bottom)) #展示图片 #image_text4.show() image_text4.save("img/tmp.png") temp = ocr.ocr("img/tmp.png") txt4="".join(temp[0]) print(txt4) return txt4 def text5(new_img): #提取汇票到日期 left = 92 top = 166 right = 176 bottom = 184 image_text5 = new_img.crop((left, top, right, bottom)) #展示图片 #image_text5.show() txt5 = tool.image_to_string(image_text5) print(txt5) return txt5 def text6(new_img): #提取票据号码 left = 598 top = 166 right = 870 bottom = 182 image_text6 = new_img.crop((left, top, right, bottom)) #展示图片 #image_text6.show() txt6 = tool.image_to_string(image_text6) print(txt6) return txt6 ocr=CnOcr() tool = pyocr.get_available_tools()[0] filePath='img' img_name=[] for i,j,name in os.walk(filePath): img_name=name count=1 book = xlwt.Workbook(encoding='utf-8',style_compression=0) sheet = book.add_sheet('test',cell_overwrite_ok=True) for i in img_name: img_url = filePath+"/"+i with open(img_url, 'rb') as f: a = f.read() new_img = PI.open(io.BytesIO(a)) ## 写入csv col = ('年份','出票日期','金额','出票人','付款行全称','汇票到日期','备注') for j in range(0,7): sheet.write(0,j,col[j]) book.save('1.csv') shijian=text1(new_img) sheet.write(count,0,shijian[0:4]) sheet.write(count,1,shijian[5:]) sheet.write(count,2,text2(new_img)) sheet.write(count,3,text3(new_img)) sheet.write(count,4,text4(new_img)) sheet.write(count,5,text5(new_img)) sheet.write(count,6,text6(new_img)) count = count + 1
以上就是详解如何通过Python实现批量数据提取的详细内容,更多关于Python批量数据提取的资料请关注我们其它相关文章!
开发者_Python学习
精彩评论