python图片识别easyocr：Downloading detection model, please wait. This may take several minutes depending

python图片识别easyocr：Downloading detection model, please wait. This may take several minutes depending upon your network connection.

初次认识easyocr，初衷是为了解决网站因为有验证码所以需要图片识别功能

easyocr安装

pip3 install easyocr

easyocr使用

"""
@File    :   MyOCR.py    
@Modify Time      @Author    @Version
------------      -------    --------    
"""
import os.path
import easyocr
# 设定识别的文字类别，en为english，也可以设置成 easyocr.Reader(['ch_sim', 'en'])，即为简体中文或者英文
reader = easyocr.Reader(['en'])
# 获取图片地址
file = os.path.join(os.path.dirname(__file__), "createImage.jpg")
result = reader.readtext(file)
for res in result:print(res)print(type(res))print(res[1])print(type(res[1]))

报错

但是直接执行是报错的，会在打印下面的信息后报错

Downloading detection model, please wait. This may take several minutes depending upon your network connection.

是因为我们的网络太慢了，导致并没有办法正常下载模型，也就没办法正常识别图片文件

我们可以通过手动下载的办法来解决

easyocr模型下载

我们点击下面这三个链接就会自动下载文件

但是我自己下载的时候很慢，没办法翻墙了才下载下来，然后上传到百度云了，需要的自取,里面有三个压缩包，也有三个pth文件，就是解压后的文件，下载pth的即可

链接: https://pan.baidu.com/s/1xyPsKORak447jS82Pt-JHg 提取码: onvl

下载到本地后，复制到当前用户的目录下

.EasyOCR\model

比如我的是mac电脑（同linux）,如果是windows电脑的话就是，C:\Users\yourname.EasyOCR\model，其中yourname是登录用户名。

ZHR:model zc$ pwd
/Users/zc/.EasyOCR/model
ZHR:model zc$ open /Users/zc/.EasyOCR/model

将文件复制过来即可

ZHR:model zc$ ls
craft_mlt_25k.pth	english_g2.pth		zh_sim_g2.pth
ZHR:model zc$ pwd
/Users/zc/.EasyOCR/model

再次执行代码就不报错了

/Users/zc/PycharmProjects/fastapitest/venv/bin/python /Users/zc/PycharmProjects/fastapitest/One/MyOCR.py
CUDA not available - defaulting to CPU. Note: This module is much faster with a GPU.
([[18, 6], [82, 6], [82, 30], [18, 30]], '47HH', 0.87657231092453)
<class 'tuple'>
47HH
<class 'str'>

识别网络图片

其实就是将图片先下载到本地而已

"""
@File    :   MyOCR.py    
@Modify Time      @Author    @Version
------------      -------    --------    
"""
import os.path
import easyocr
import requestsurl = 'https://img-blog.csdnimg.cn/e9f0a6ac9826499fadd19b7d5aa1e1d4.png'
# verify=False：是否验证=否
rsp = requests.get(url, verify=False)
with open('1.jpg', 'wb') as f:f.write(rsp.content)print(easyocr.__version__)
reader = easyocr.Reader(['en'])
file = os.path.join(os.path.dirname(__file__), "1.jpg")
result = reader.readtext(file)
for res in result:print(res)print(type(res))print(res[1])print(type(res[1]))

结果

/Users/zc/PycharmProjects/fastapitest/venv/lib/python3.8/site-packages/urllib3/connectionpool.py:1013: InsecureRequestWarning: Unverified HTTPS request is being made to host '127.0.0.1'. Adding certificate verification is strongly advised. See: https://urllib3.readthedocs.io/en/1.26.x/advanced-usage.html#ssl-warningswarnings.warn(
1.4.1
CUDA not available - defaulting to CPU. Note: This module is much faster with a GPU.
([[34, 11], [157, 11], [157, 53], [34, 53]], '6Ty 6', 0.504037462923347)
<class 'tuple'>
6Ty 6
<class 'str'>

大家也看到了，这个识别的不是那么的准，比如上面的图片会有结果中存在空格的情况，需要在优化，甚至会有识别不全的情况，要按照情况特别优化才可以