python下OCR识别-baiduOCR

登录百度OCR，创建应用

https://console.bce.baidu.com/ai/?fromai=1#/ai/ocr/overview/index

创建后，查看：

根据API文档调试：

先获取Access Token：

如：https://aip.baidubce.com/oauth/2.0/token?grant_type=client_credentials&client_id=Va5yQRHlA4Fq5eR3LT0vuXV4&client_secret=0rDSjzQ20XUj5itV6WRtznPQSzr5pVw2&

API调用：

测试代码1：

#coding:utf-8
__author__='vulsee.com'
import requests
import base64

def getToken():
	html=requests.post(getTokenurl).json()
	access_token= html['access_token']
	return access_token

def getImgbs64(imgfile):
	f = open(imgfile, 'rb')
	img = base64.b64encode(f.read())
	return img

def getResult(imgfile):
	access_token=getToken()
	url='https://aip.baidubce.com/rest/2.0/ocr/v1/accurate_basic?access_token=%s' % access_token

	headers = {'content-type': 'application/x-www-form-urlencoded'}
	image=getImgbs64(imgfile)
	params = {"image":image}
	#以下非必要信息，可以删除
	'''
	language_type='auto_detect'
	detect_direction='false'
	paragraph='false'
	probability='false'
	'''
	html=requests.post(url,headers=headers,data=params).json()
	res= html['words_result']
	rs=''
	for i in res:
		#print i['words']
		r=i['words']
		rs+=i['words']
	print rs
def start():
	getResult('105010.jpg')

def main():
	start()

if __name__ == '__main__':
	global getTokenurl,grant_type,client_id,client_secret 
	grant_type='client_credentials'
	client_id='*****'
	client_secret='***'
	getTokenurl='https://aip.baidubce.com/oauth/2.0/token?grant_type=client_credentials&client_id=%s&client_secret=%s&' %(client_id,client_secret)
	main()

处理结果：

测试代码2：

如果flag为file，则从文件读取；如果flag为url，则直接从web获取

#coding:utf-8
__author__='vulsee.com'
import requests
import base64

def getToken():
	html=requests.post(getTokenurl).json()
	access_token= html['access_token']
	return access_token

def getImgbs64(imgfile):
	f = open(imgfile, 'rb')
	img = base64.b64encode(f.read())
	return img

def getImgbs64Online(url):
	html=requests.get(url).content
	img = base64.b64encode(html)
	return img

def getResult(flag,imgfile,access_token):

	url='https://aip.baidubce.com/rest/2.0/ocr/v1/accurate_basic?access_token=%s' % access_token
	headers = {'content-type': 'application/x-www-form-urlencoded'}
	if flag=='file':
		image=getImgbs64(imgfile)
		params = {"image":image}
		html=requests.post(url,headers=headers,data=params).json()
		res= html['words_result']
		rs=''
		for i in res:
			r=i['words']
			rs+=i['words']
		print 'source:%s' % flag,rs
	else:
		image=getImgbs64Online(imgfile)
		params = {"image":image}
		html=requests.post(url,headers=headers,data=params).json()
		res= html['words_result']
		rs=''
		for i in res:
			#print i['words']
			r=i['words']
			rs+=i['words']
		print 'source:%s' % flag,rs
def start(access_token,flag):

	if flag=='file':
		filename='105007.jpg'
		getResult(flag,filename,access_token)
	elif flag=='url':
		url='https://****:***/api/api/*****?id=***&***=159980316534726'
		getResult('url',url,access_token)
	else:
		print 'error'

def main():
	access_token=getToken()
	#如果flag为file，则从文件读取；如果flag为url，则直接从web获取
	flag='url'
	start(access_token,flag)

if __name__ == '__main__':
	global getTokenurl,grant_type,client_id,client_secret
	grant_type='client_credentials'
	client_id='****'
	client_secret='******'
	getTokenurl='https://aip.baidubce.com/oauth/2.0/token?grant_type=client_credentials&client_id=%s&client_secret=%s&' %(client_id,client_secret)
	main()