devbox
你好赵伟
这个屌丝很懒，什么也没留下！
关注作者
热门标签
热门文章
当前位置:   article > 正文

ocr大pk（阿里读光，百度，face++，有道）_读光ocr

作者：你好赵伟 | 2024-02-27 20:31:59

赞

踩

读光ocr

当下视角：为客户找到你的价值

​​​​​​​未来视角：为你的价值找到客户

 

阿里读光

  1. #https://market.aliyun.com/products/57124001/cmapi023866.html?spm=a2c0j.8222507.973837.1.4dd8e532YZ2wUE#sku=yuncode1786600000
  2. #https://market.aliyun.com/products/57124001/cmapi028554.html?spm=a2c0j.8222507.1092900.1.4dd8e532YZ2wUE#sku=yuncode2255400000
  3.  
  4. #图像数据：base64编码，要求base64编码后大小不超过4M，最短边至少15px，最长边最大4096px，支持jpg/png/bmp格式，和url参数只能同时存在一个\"img\":\"\",
  5. #图像url地址：图片完整URL，URL长度不超过1024字节，URL对应的图片base64编码后大小不超过4M，最短边至少15px，最长边最大4096px，支持jpg/png/bmp格式，和img参数只能同时存在一个\"url\":\"\",
  6. #是否需要识别结果中每一行的置信度，默认不需要。true：需要false：不需要\"prob\":false,
  7. #是否需要单字识别功能，默认不需要。true：需要false：不需要\"charInfo\":false,
  8. #是否需要自动旋转功能，默认不需要。true：需要false：不需要\"rotate\":false,
  9. #是否需要表格识别功能，默认不需要。true：需要false：不需要\"table\":false
  10. import urllib
  11. import urllib.request
  12. import urllib.parse
  13. import json
  14. import ssl
  15. import base64
  16. import cv2
  17.  
  18. class ALI_OCR():
  19.     def __init__(self):
  20.         """ APPID AK SK """
  21.         self.APPCODE = '47fcde151a52426db0f61fede1c69507'
  22.         #self.APPCODE = 'cb1f7559255140ab989d319be313c1b7'
  23.  
  24.         #印刷文字识别－文档小说图片文字识别
  25.         self.url="https://ocrapi-document.taobao.com/ocrservice/document"
  26.         #通用文字识别－高精版
  27.         #self.url="https://ocrapi-advanced.taobao.com/ocrservice/advanced"
  28.  
  29.         self.headers = {
  30.          'Authorization': 'APPCODE '+self.APPCODE,
  31.          'Content-Type': 'application/json; charset=UTF-8'
  32.     }
  33.     """ 读取图片 """
  34.     def get_file_content(self,filePath):
  35.         with open(filePath, 'rb') as fp:
  36.             return base64.b64encode(fp.read())
  37.     def draw_rectangle(self,img,coordinate_list,color):
  38.         cv2.line(img,(int(coordinate_list[0]["x"]),int(coordinate_list[0]["y"])),(int(coordinate_list[1]["x"]),int(coordinate_list[1]["y"])),color,thickness=2)
  39.         cv2.line(img,(int(coordinate_list[1]["x"]),int(coordinate_list[1]["y"])),(int(coordinate_list[2]["x"]),int(coordinate_list[2]["y"])),color,thickness=2)
  40.         cv2.line(img,(int(coordinate_list[2]["x"]),int(coordinate_list[2]["y"])),(int(coordinate_list[3]["x"]),int(coordinate_list[3]["y"])),color,thickness=2)
  41.         cv2.line(img,(int(coordinate_list[3]["x"]),int(coordinate_list[3]["y"])),(int(coordinate_list[0]["x"]),int(coordinate_list[0]["y"])),color,thickness=2)
  42.  
  43.         return img
  44.  
  45.     def process(self,image_name):
  46.         try:
  47.             img = self.get_file_content(image_name).decode()
  48.             dict = {'img': img,'prob': 'false'}
  49.             params=json.dumps(dict).encode(encoding='UTF8')
  50.             req = urllib.request.Request(self.url, params,self.headers)
  51.             
  52.             ctx = ssl.create_default_context()
  53.             ctx.check_hostname = False
  54.             ctx.verify_mode = ssl.CERT_NONE
  55.  
  56.             r = urllib.request.urlopen(req, context=ctx)
  57.             html =r.read()
  58.             return html.decode("utf8")
  59.         except urllib.error.HTTPError as e:
  60.             print(e.code)
  61.             print(e.read().decode("utf8"))
  62.  
  63.     def save_result_pic_txt(self,input_img,input_json,save_pic=True,save_txt=True,save_dir="./result/"):
  64.         result_json=json.loads(input_json)
  65.         img=cv2.imread(input_img)
  66.         if save_txt:
  67.             txt_name=save_dir+input_img.split("/")[-1].split(".")[0]+".txt"
  68.             file_writer=open(txt_name,"w",encoding="utf-8")
  69.             for lines in result_json["prism_wordsInfo"]:
  70.                 coordinate_list=lines["pos"]
  71.                 img=scan_ocr.draw_rectangle(img,coordinate_list,color=(0,0,255))
  72.                 print(lines['word'])
  73.                 file_writer.write(lines['word']+"\n")
  74.             file_writer.close()
  75.         if save_pic:
  76.             #cv2.imshow("result",img)
  77.             #cv2.waitKey()
  78.             cv2.imwrite(save_dir+input_img.split("/")[-1],img)
  79.  
  80.  
  81. if __name__=="__main__":
  82.     scan_ocr=ALI_OCR()
  83.     image_name='test.png'
  84.     result_str=scan_ocr.process(image_name)
  85.     scan_ocr.save_result_pic_txt(image_name,result_str,save_pic=True,save_txt=True,save_dir="./result_ali/")

 

百度

  1. # coding=utf-8
  2. ##https://cloud.baidu.com/doc/OCR/OCR-Python-SDK.html
  3. import os
  4. import cv2
  5. from aip import AipOcr
  6.  
  7. #pip3 install baidu-aip
  8.  
  9.  
  10. class BAIDU_OCR():
  11.     def __init__(self):
  12.         """ APPID AK SK """
  13.         """
  14.         self.APP_ID = '10122282'
  15.         self.API_KEY = 'weGcK2iHjzOA80VnwT14d1yZ'
  16.         self.SECRET_KEY = '8z7YkTPNF6rS4a6l6F98Fc5a06cvLM44'
  17.         """
  18.         self.APP_ID = '10122118'
  19.         self.API_KEY = 'LI8ZEszLmscbLwYN38azHGgG'
  20.         self.SECRET_KEY = '2aWkfvdHXNUVrXzdVvA6nBQxoeIgbvbG'
  21.  
  22.         self.client = AipOcr(self.APP_ID, self.API_KEY, self.SECRET_KEY)
  23.  
  24.     """ 读取图片 """
  25.     def get_file_content(self,filePath):
  26.         with open(filePath, 'rb') as fp:
  27.             return fp.read()
  28.  
  29.     def process(self,image_name):
  30.         image = self.get_file_content(image_name)
  31.         """ 调用通用文字识别（含位置信息版）, 图片参数为本地图片 """
  32.         result=self.client.general(image)
  33.         """ 调用通用文字识别（含位置高精度版） """
  34.         #result=self.client.accurate(image)
  35.         return result
  36.  
  37.     def save_result_pic_txt(self,input_img,input_json,save_pic=True,save_txt=True,save_dir="./result/"):
  38.         img=cv2.imread(input_img)
  39.         if save_txt:
  40.             txt_name=save_dir+input_img.split("/")[-1].split(".")[0]+".txt"
  41.             file_writer=open(txt_name,"w",encoding="utf-8")
  42.             for word_result in result["words_result"]:
  43.                 x1=word_result["location"]["left"]
  44.                 y1=word_result["location"]["top"]
  45.                 x2=x1+word_result["location"]["width"]
  46.                 y2=y1+word_result["location"]["height"]
  47.                 cv2.rectangle(img,(x1,y1),(x2,y2),color=(0,0,255),thickness=2)
  48.                 file_writer.write(word_result["words"]+"\n")
  49.                 print(word_result["words"])
  50.             file_writer.close()
  51.         if save_pic:
  52.             #cv2.imshow("result",img)
  53.             #cv2.waitKey()
  54.             cv2.imwrite(save_dir+input_img.split("/")[-1],img)
  55.  
  56. if __name__=="__main__":
  57.     scan_ocr=BAIDU_OCR()
  58.     image_name='test.png'
  59.     result=scan_ocr.process(image_name)
  60.     scan_ocr.save_result_pic_txt(image_name,result,save_pic=True,save_txt=True,save_dir="./result_baidu/")
  61.

face++

  1. #https://console.faceplusplus.com.cn/documents/7776484
  2. #图片要求：
  3. #图片格式：JPG(JPEG)，PNG
  4. #图片像素尺寸：最小48*48像素，最大800*800像素
  5. #图片文件大小：2MB
  6. import requests
  7. from json import JSONDecoder
  8. import json
  9. import cv2
  10.  
  11.  
  12. class FACEPLUSPLUS_OCR():
  13.     def __init__(self):
  14.         """ APPID AK SK """
  15.         self.api_key = "1mCONgxd9oRMjezN9PJd0l-hjdqVT5xS"
  16.         self.api_secret="FrX9rbKZi6CNsiN8ee-ISaVnRhRVkPTc"
  17.  
  18.         self.url="https://api-cn.faceplusplus.com/imagepp/v1/recognizetext"
  19.         self.data = {"api_key": self.api_key, "api_secret": self.api_secret}
  20.  
  21.     def draw_rectangle(self,img,coordinate_list,color):
  22.         cv2.line(img,(int(coordinate_list[0]["x"]),int(coordinate_list[0]["y"])),(int(coordinate_list[1]["x"]),int(coordinate_list[1]["y"])),color,thickness=2)
  23.         cv2.line(img,(int(coordinate_list[1]["x"]),int(coordinate_list[1]["y"])),(int(coordinate_list[2]["x"]),int(coordinate_list[2]["y"])),color,thickness=2)
  24.         cv2.line(img,(int(coordinate_list[2]["x"]),int(coordinate_list[2]["y"])),(int(coordinate_list[3]["x"]),int(coordinate_list[3]["y"])),color,thickness=2)
  25.         cv2.line(img,(int(coordinate_list[3]["x"]),int(coordinate_list[3]["y"])),(int(coordinate_list[0]["x"]),int(coordinate_list[0]["y"])),color,thickness=2)
  26.  
  27.         return img
  28.  
  29.     def process(self,image_name):
  30.         try:
  31.             files = {"image_file": open(image_name, "rb")}
  32.  
  33.             response = requests.post(self.url, data=self.data, files=files)
  34.             req_con = response.content.decode('utf-8')
  35.             req_dict = JSONDecoder().decode(req_con)
  36.  
  37.             return req_dict
  38.         except Exception as e:
  39.             print(e)
  40.  
  41.     def save_result_pic_txt(self,input_img,input_json,save_pic=True,save_txt=True,save_dir="./result/"):
  42.         img=cv2.imread(input_img)
  43.         if save_txt:
  44.             txt_name=save_dir+input_img.split("/")[-1].split(".")[0]+".txt"
  45.             file_writer=open(txt_name,"w",encoding="utf-8")
  46.             for objects in input_json["result"]:
  47.                 print(objects["value"])
  48.                 file_writer.write(objects["value"]+"\n")
  49.                 for lines in objects["child-objects"]:
  50.                     coordinate_list=lines["position"]
  51.                     img=scan_ocr.draw_rectangle(img,coordinate_list,color=(0,0,255))
  52.                     #print(lines['value'])
  53.  
  54.             file_writer.close()
  55.         if save_pic:
  56.             #cv2.imshow("result",img)
  57.             #cv2.waitKey()
  58.             cv2.imwrite(save_dir+input_img.split("/")[-1],img)
  59.  
  60.  
  61. if __name__=="__main__":
  62.     scan_ocr=FACEPLUSPLUS_OCR()
  63.     image_name='test.png'
  64.     result_json=scan_ocr.process(image_name)
  65.     scan_ocr.save_result_pic_txt(image_name,result_json,save_pic=True,save_txt=True,save_dir="./result_faceplusplus/")

有道

  1. #/usr/bin/env python
  2. #coding=utf8
  3. ##http://ai.youdao.com/docs/doc-ocr-api.s#p01
  4.  
  5.  
  6. import hashlib
  7. import urllib
  8. import urllib.request
  9. import random
  10. import json
  11. import base64
  12. import cv2
  13.  
  14.  
  15.  
  16.  
  17. httpClient = None
  18.  
  19.  
  20.  
  21. class YOUDAO_OCR():
  22.     def __init__(self):
  23.         """ APPID AK SK """
  24.  
  25.         #self.appKey = '7b02eb435b51584d'
  26.         #self.secretKey = '5nr7r2QO0fS6J4fNRwVgiqNVOhjA6ZBw'
  27.         
  28.         self.appKey = '054adec25ae90a75'
  29.         self.secretKey = 'Fe894uffWfQv5KMy77olO8wODVlt0dRw'
  30.         
  31.         #按字识别:10011,按行识别:10012
  32.         self.detectType = '10012'
  33.         #1,表示base64
  34.         self.imageType = '1'
  35.         #中英混合:zh-en,英文:en
  36.         self.langType = 'zh-en'
  37.         self.salt = random.randint(1, 65536)
  38.         self.url='http://openapi.youdao.com/ocrapi'
  39.  
  40.     """ 读取图片 """
  41.     def get_file_content(self,filePath):
  42.         with open(filePath, 'rb') as fp:
  43.             return base64.b64encode(fp.read())
  44.     def draw_rectangle(self,img,coordinate_list,color):
  45.         cv2.line(img,(int(coordinate_list[0]),int(coordinate_list[1])),(int(coordinate_list[2]),int(coordinate_list[3])),color,thickness=2)
  46.         cv2.line(img,(int(coordinate_list[2]),int(coordinate_list[3])),(int(coordinate_list[4]),int(coordinate_list[5])),color,thickness=2)
  47.         cv2.line(img,(int(coordinate_list[4]),int(coordinate_list[5])),(int(coordinate_list[6]),int(coordinate_list[7])),color,thickness=2)
  48.         cv2.line(img,(int(coordinate_list[6]),int(coordinate_list[7])),(int(coordinate_list[0]),int(coordinate_list[1])),color,thickness=2)
  49.         return img
  50.  
  51.     def process(self,image_name):
  52.         try:
  53.             img = self.get_file_content(image_name)
  54.             sign = self.appKey+img.decode()+str(self.salt)+self.secretKey
  55.             m1 = hashlib.md5()
  56.             m1.update(sign.encode("utf8"))
  57.             sign = m1.hexdigest()
  58.             data = {'appKey':self.appKey,'img':img,'detectType':self.detectType,'imageType':self.imageType,'langType':self.langType,'salt':str(self.salt),'sign':sign}
  59.             data = urllib.parse.urlencode(data).encode(encoding='utf-8')
  60.             req = urllib.request.Request(self.url,data)
  61.             #response是HTTPResponse对象
  62.             response = urllib.request.urlopen(req)
  63.             output_str=response.read().decode('utf-8')
  64.             return output_str
  65.         except Exception as e:
  66.             print ("except:",e)
  67.         finally:
  68.             if httpClient:
  69.                 httpClient.close()
  70.                 return ""
  71.  
  72.     def save_result_pic_txt(self,input_img,input_json,save_pic=True,save_txt=True,save_dir="./result/"):
  73.         result_json=json.loads(input_json)
  74.         img=cv2.imread(input_img)
  75.         if save_txt:
  76.             txt_name=save_dir+input_img.split("/")[-1].split(".")[0]+".txt"
  77.             file_writer=open(txt_name,"w",encoding="utf-8")
  78.             for regions in result_json["Result"]["regions"]:
  79.                 img=scan_ocr.draw_rectangle(img,regions["boundingBox"].split(','),color=(0,255,0))
  80.                 for lines in regions['lines']:
  81.                     coordinate_list=lines["boundingBox"]
  82.                     img=scan_ocr.draw_rectangle(img,coordinate_list.split(','),color=(0,0,255))
  83.                     print(lines['text'])
  84.                     file_writer.write(lines['text']+"\n")
  85.             file_writer.close()
  86.         if save_pic:
  87.             #cv2.imshow("result",img)
  88.             #cv2.waitKey()
  89.             cv2.imwrite(save_dir+input_img.split("/")[-1],img)
  90.  
  91.  
  92. if __name__=="__main__":
  93.     scan_ocr=YOUDAO_OCR()
  94.     image_name='test.png'
  95.     result_str=scan_ocr.process(image_name)
  96.     scan_ocr.save_result_pic_txt(image_name,result_str,save_pic=True,save_txt=True,save_dir="./result_youdao/")

 

 

 

 

声明：本文内容由网友自发贡献，不代表【wpsshop博客】立场，版权归原作者所有，本站不承担相应法律责任。如您发现有侵权的内容，请联系我们。转载请注明出处：https://www.wpsshop.cn/w/你好赵伟/article/detail/155780
推荐阅读
相关标签

Copyright © 2003-2013 www.wpsshop.cn 版权所有，并保留所有权利。

  

闽ICP备14008679号