百度OCR api 使用 (下)

上一篇写的,OCR调用是失败的
后来仔细研究过,用base64编码,最后成功了.

另外一个问题,百度原来提供两种方式传递原图片,但是只有第一种,即图片base64编码后上传才有效,第二种上传原文件,怎么都不成功.后来想到把在线的图片读到缓存再base64编码,用第一种方式传递图片,结果可行.

另 安装 simplejson ,然后 import simplejson as json 也直接可用 ,而且,.decode()都可以不需要.比原本的json好用.

---

#!/usr/bin/env python
# -*- coding: utf-8 -*-

#@Author           :  BigBro
#@DateTime         :  2015-11-21 16:41:31
#@Filename         :  baidu_ocr_ver1.0.py
#@Description      :  百度 图像 ocr

import sys, urllib, json
import urllib.request
import urllib.parse
import base64
url = 'http://apis.baidu.com/apistore/idlocr/ocr'

data = {}
data['fromdevice'] = "pc"
data['clientip'] = "10.10.10.0"
data['detecttype'] = "LocateRecognize"
data['languagetype'] = "CHN_ENG"
data['imagetype'] = "1"
#图片在本地
with open(r'C:\Users\bigbrolee\Desktop\abc.jpg',mode = 'rb') as f: 
    img = f.read()

#图片在网络
# img_url = 'https://pic4.zhimg.com/b8189c1ac3a2d5b790d8bc9bf97f41d7_l.jpg'
# try:
#   resp_img=urllib.request.urlopen(img_url, timeout=5)
#   img=resp_img.read()
# except socket.timeout:
#   raise socket.timeout

data['image'] =base64.b64encode(img)


decoded_data = urllib.parse.urlencode(data)
decoded_data = decoded_data.encode('utf-8')
# #print(decoded_data)


req = urllib.request.Request(url,decoded_data)

req.add_header("Content-Type", "application/x-www-form-urlencoded")
req.add_header("apikey", "f46299494a15949102e3d040fdbd5bbc") #a63fd60067141f18cffa9af6a1563b4e


resp = urllib.request.urlopen(req)#, data = decoded_data)
content = resp.read()
if(content):
    content = json.loads(content.decode())
    print(content) 

---