ocr大pk（阿里读光，百度，face++，有道）

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

#https://market.aliyun.com/products/57124001/cmapi023866.html?spm=a2c0j.8222507.973837.1.4dd8e532YZ2wUE#sku=yuncode1786600000
#https://market.aliyun.com/products/57124001/cmapi028554.html?spm=a2c0j.8222507.1092900.1.4dd8e532YZ2wUE#sku=yuncode2255400000

#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\":\"\",
#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\":\"\",
#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,
#是否需要单字识别功能，默认不需要。true：需要false：不需要\"charInfo\":false,
#是否需要自动旋转功能，默认不需要。true：需要false：不需要\"rotate\":false,
#是否需要表格识别功能，默认不需要。true：需要false：不需要\"table\":false
import urllib
import urllib.request
import urllib.parse
import json
import ssl
import base64
import cv2

class ALI_OCR():
    def __init__(self):
        """ APPID AK SK """
        self.APPCODE = '47fcde151a52426db0f61fede1c69507'
        #self.APPCODE = 'cb1f7559255140ab989d319be313c1b7'

        #印刷文字识别－文档小说图片文字识别
        self.url="https://ocrapi-document.taobao.com/ocrservice/document"
        #通用文字识别－高精版
        #self.url="https://ocrapi-advanced.taobao.com/ocrservice/advanced"

        self.headers = {
         'Authorization': 'APPCODE '+self.APPCODE,
         'Content-Type': 'application/json; charset=UTF-8'
    }
    """ 读取图片 """
    def get_file_content(self,filePath):
        with open(filePath, 'rb') as fp:
            return base64.b64encode(fp.read())
    def draw_rectangle(self,img,coordinate_list,color):
        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)
        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)
        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)
        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)

        return img

    def process(self,image_name):
        try:
            img = self.get_file_content(image_name).decode()
            dict = {'img': img,'prob': 'false'}
            params=json.dumps(dict).encode(encoding='UTF8')
            req = urllib.request.Request(self.url, params,self.headers)
            
            ctx = ssl.create_default_context()
            ctx.check_hostname = False
            ctx.verify_mode = ssl.CERT_NONE

            r = urllib.request.urlopen(req, context=ctx)
            html =r.read()
            return html.decode("utf8")
        except urllib.error.HTTPError as e:
            print(e.code)
            print(e.read().decode("utf8"))

    def save_result_pic_txt(self,input_img,input_json,save_pic=True,save_txt=True,save_dir="./result/"):
        result_json=json.loads(input_json)
        img=cv2.imread(input_img)
        if save_txt:
            txt_name=save_dir+input_img.split("/")[-1].split(".")[0]+".txt"
            file_writer=open(txt_name,"w",encoding="utf-8")
            for lines in result_json["prism_wordsInfo"]:
                coordinate_list=lines["pos"]
                img=scan_ocr.draw_rectangle(img,coordinate_list,color=(0,0,255))
                print(lines['word'])
                file_writer.write(lines['word']+"\n")
            file_writer.close()
        if save_pic:
            #cv2.imshow("result",img)
            #cv2.waitKey()
            cv2.imwrite(save_dir+input_img.split("/")[-1],img)


if __name__=="__main__":
    scan_ocr=ALI_OCR()
    image_name='test.png'
    result_str=scan_ocr.process(image_name)
    scan_ocr.save_result_pic_txt(image_name,result_str,save_pic=True,save_txt=True,save_dir="./result_ali/")

 

百度

# coding=utf-8
##https://cloud.baidu.com/doc/OCR/OCR-Python-SDK.html
import os
import cv2
from aip import AipOcr

#pip3 install baidu-aip


class BAIDU_OCR():
    def __init__(self):
        """ APPID AK SK """
        """
        self.APP_ID = '10122282'
        self.API_KEY = 'weGcK2iHjzOA80VnwT14d1yZ'
        self.SECRET_KEY = '8z7YkTPNF6rS4a6l6F98Fc5a06cvLM44'
        """
        self.APP_ID = '10122118'
        self.API_KEY = 'LI8ZEszLmscbLwYN38azHGgG'
        self.SECRET_KEY = '2aWkfvdHXNUVrXzdVvA6nBQxoeIgbvbG'

        self.client = AipOcr(self.APP_ID, self.API_KEY, self.SECRET_KEY)

    """ 读取图片 """
    def get_file_content(self,filePath):
        with open(filePath, 'rb') as fp:
            return fp.read()

    def process(self,image_name):
        image = self.get_file_content(image_name)
        """ 调用通用文字识别（含位置信息版）, 图片参数为本地图片 """
        result=self.client.general(image)
        """ 调用通用文字识别（含位置高精度版） """
        #result=self.client.accurate(image)
        return result

    def save_result_pic_txt(self,input_img,input_json,save_pic=True,save_txt=True,save_dir="./result/"):
        img=cv2.imread(input_img)
        if save_txt:
            txt_name=save_dir+input_img.split("/")[-1].split(".")[0]+".txt"
            file_writer=open(txt_name,"w",encoding="utf-8")
            for word_result in result["words_result"]:
                x1=word_result["location"]["left"]
                y1=word_result["location"]["top"]
                x2=x1+word_result["location"]["width"]
                y2=y1+word_result["location"]["height"]
                cv2.rectangle(img,(x1,y1),(x2,y2),color=(0,0,255),thickness=2)
                file_writer.write(word_result["words"]+"\n")
                print(word_result["words"])
            file_writer.close()
        if save_pic:
            #cv2.imshow("result",img)
            #cv2.waitKey()
            cv2.imwrite(save_dir+input_img.split("/")[-1],img)

if __name__=="__main__":
    scan_ocr=BAIDU_OCR()
    image_name='test.png'
    result=scan_ocr.process(image_name)
    scan_ocr.save_result_pic_txt(image_name,result,save_pic=True,save_txt=True,save_dir="./result_baidu/")

face++

#https://console.faceplusplus.com.cn/documents/7776484
#图片要求：
#图片格式：JPG(JPEG)，PNG
#图片像素尺寸：最小48*48像素，最大800*800像素
#图片文件大小：2MB
import requests
from json import JSONDecoder
import json
import cv2


class FACEPLUSPLUS_OCR():
    def __init__(self):
        """ APPID AK SK """
        self.api_key = "1mCONgxd9oRMjezN9PJd0l-hjdqVT5xS"
        self.api_secret="FrX9rbKZi6CNsiN8ee-ISaVnRhRVkPTc"

        self.url="https://api-cn.faceplusplus.com/imagepp/v1/recognizetext"
        self.data = {"api_key": self.api_key, "api_secret": self.api_secret}

    def draw_rectangle(self,img,coordinate_list,color):
        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)
        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)
        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)
        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)

        return img

    def process(self,image_name):
        try:
            files = {"image_file": open(image_name, "rb")}

            response = requests.post(self.url, data=self.data, files=files)
            req_con = response.content.decode('utf-8')
            req_dict = JSONDecoder().decode(req_con)

            return req_dict
        except Exception as e:
            print(e)

    def save_result_pic_txt(self,input_img,input_json,save_pic=True,save_txt=True,save_dir="./result/"):
        img=cv2.imread(input_img)
        if save_txt:
            txt_name=save_dir+input_img.split("/")[-1].split(".")[0]+".txt"
            file_writer=open(txt_name,"w",encoding="utf-8")
            for objects in input_json["result"]:
                print(objects["value"])
                file_writer.write(objects["value"]+"\n")
                for lines in objects["child-objects"]:
                    coordinate_list=lines["position"]
                    img=scan_ocr.draw_rectangle(img,coordinate_list,color=(0,0,255))
                    #print(lines['value'])

            file_writer.close()
        if save_pic:
            #cv2.imshow("result",img)
            #cv2.waitKey()
            cv2.imwrite(save_dir+input_img.split("/")[-1],img)


if __name__=="__main__":
    scan_ocr=FACEPLUSPLUS_OCR()
    image_name='test.png'
    result_json=scan_ocr.process(image_name)
    scan_ocr.save_result_pic_txt(image_name,result_json,save_pic=True,save_txt=True,save_dir="./result_faceplusplus/")

有道

#/usr/bin/env python
#coding=utf8
##http://ai.youdao.com/docs/doc-ocr-api.s#p01


import hashlib
import urllib
import urllib.request
import random
import json
import base64
import cv2




httpClient = None



class YOUDAO_OCR():
    def __init__(self):
        """ APPID AK SK """

        #self.appKey = '7b02eb435b51584d'
        #self.secretKey = '5nr7r2QO0fS6J4fNRwVgiqNVOhjA6ZBw'
        
        self.appKey = '054adec25ae90a75'
        self.secretKey = 'Fe894uffWfQv5KMy77olO8wODVlt0dRw'
        
        #按字识别:10011,按行识别:10012
        self.detectType = '10012'
        #1,表示base64
        self.imageType = '1'
        #中英混合:zh-en,英文:en
        self.langType = 'zh-en'
        self.salt = random.randint(1, 65536)
        self.url='http://openapi.youdao.com/ocrapi'

    """ 读取图片 """
    def get_file_content(self,filePath):
        with open(filePath, 'rb') as fp:
            return base64.b64encode(fp.read())
    def draw_rectangle(self,img,coordinate_list,color):
        cv2.line(img,(int(coordinate_list[0]),int(coordinate_list[1])),(int(coordinate_list[2]),int(coordinate_list[3])),color,thickness=2)
        cv2.line(img,(int(coordinate_list[2]),int(coordinate_list[3])),(int(coordinate_list[4]),int(coordinate_list[5])),color,thickness=2)
        cv2.line(img,(int(coordinate_list[4]),int(coordinate_list[5])),(int(coordinate_list[6]),int(coordinate_list[7])),color,thickness=2)
        cv2.line(img,(int(coordinate_list[6]),int(coordinate_list[7])),(int(coordinate_list[0]),int(coordinate_list[1])),color,thickness=2)
        return img

    def process(self,image_name):
        try:
            img = self.get_file_content(image_name)
            sign = self.appKey+img.decode()+str(self.salt)+self.secretKey
            m1 = hashlib.md5()
            m1.update(sign.encode("utf8"))
            sign = m1.hexdigest()
            data = {'appKey':self.appKey,'img':img,'detectType':self.detectType,'imageType':self.imageType,'langType':self.langType,'salt':str(self.salt),'sign':sign}
            data = urllib.parse.urlencode(data).encode(encoding='utf-8')
            req = urllib.request.Request(self.url,data)
            #response是HTTPResponse对象
            response = urllib.request.urlopen(req)
            output_str=response.read().decode('utf-8')
            return output_str
        except Exception as e:
            print ("except:",e)
        finally:
            if httpClient:
                httpClient.close()
                return ""

    def save_result_pic_txt(self,input_img,input_json,save_pic=True,save_txt=True,save_dir="./result/"):
        result_json=json.loads(input_json)
        img=cv2.imread(input_img)
        if save_txt:
            txt_name=save_dir+input_img.split("/")[-1].split(".")[0]+".txt"
            file_writer=open(txt_name,"w",encoding="utf-8")
            for regions in result_json["Result"]["regions"]:
                img=scan_ocr.draw_rectangle(img,regions["boundingBox"].split(','),color=(0,255,0))
                for lines in regions['lines']:
                    coordinate_list=lines["boundingBox"]
                    img=scan_ocr.draw_rectangle(img,coordinate_list.split(','),color=(0,0,255))
                    print(lines['text'])
                    file_writer.write(lines['text']+"\n")
            file_writer.close()
        if save_pic:
            #cv2.imshow("result",img)
            #cv2.waitKey()
            cv2.imwrite(save_dir+input_img.split("/")[-1],img)


if __name__=="__main__":
    scan_ocr=YOUDAO_OCR()
    image_name='test.png'
    result_str=scan_ocr.process(image_name)
    scan_ocr.save_result_pic_txt(image_name,result_str,save_pic=True,save_txt=True,save_dir="./result_youdao/")