Detect words in the picture using the baidu's api rather than tesseract

最新推荐文章于 2021-12-20 16:42:52 发布

小白笑苍

最新推荐文章于 2021-12-20 16:42:52 发布

阅读量169

点赞数

CC 4.0 BY-SA版权

分类专栏： Python

本文链接：https://blog.youkuaiyun.com/toyijiu/article/details/86502846

Python 专栏收录该内容

65 篇文章

订阅专栏

本文介绍了一种使用Baidu OCR API进行图片文字识别的方法，对比了Tesseract的不足，详细展示了如何注册获取API密钥、安装必要的库以及使用Python代码实现文字检测的过程。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

Sometimes tesseract does not work well on detecting the words in the picture,especially the colorful words or blurry ones. So I try to find a new way to realize it and there has an baidu’s API which works well.

The official document is here:word detection document

After you sign in the website, you can get three values about the api that will be used in later’s register progress:

APP_ID = '10xxxx57'
API_KEY = 'vxxxxxxxxxxxxxxxxxsZyuwz9yKS2EghBs'
SECRET_KEY = 'm7pjnSNCKZxxxxxxxxxxxxxxxswGmIO35zsi'

then install the Lib:

pip install aip

and you can use the api to detect the words,for example:

from aip import AipOcr
APP_ID = '10xxxx57'
API_KEY = 'vxxxxxxxxxxxxxxxxxsZyuwz9yKS2EghBs'
SECRET_KEY = 'm7pjnSNCKZxxxxxxxxxxxxxxxswGmIO35zsi'
client = AipOcr(APP_ID, API_KEY, SECRET_KEY)
with open('/Users/zhaoluyang/Desktop/test2.png','rb') as f:
    img = f.read()
    msg = client.basicGeneral(img)
    for i in msg.get('words_result'):
        print(i.get('words'))