光学文字识别(OCR)

左手的ㄟ右手 2023-06-16 15:59 386阅读 0赞

## 一、什么是光学文字识别？ ##

简单来说，就是将图片上的文字给识别出来，称为光学文字识别(Optical Character Recognition),简写为OCR。目前市面上这种库并不多，因为做这个需要很强的深度学习功底，而大多数人又不懂这个方面的内容，所以开源的库就更少了，公认的最优秀的开源的OCR库为谷歌赞助开发的Tesseract。

Tesseract是一个开源的OCR引擎，大家均可以免费使用，惠普公司的布里斯托尔实验室在1984-1994年开发完成，起初作为惠普的平板扫描仪的文字识别引擎。Tesseract在1995年UNLV OCR字符识别准确性测试中拔得头筹，受到广泛关注，后来HP放弃了OCR市场，在1994年以后，Tesseract的开发就停止了。

在2005年，HP将Tesseract贡献给开源社区。美国内华达州信息技术研究所获得该源码，同时，Google开始对Tesseract进行功能扩展及优化。目前，Tesseract作为开源项目发布在Google Project上，重获新生，它支持60种以上的语言，提供一个引擎和一个命令行工具。

### 二、安装tesseract-ocr ###

**1.双击可执行文件。**

![20191127110350408.png][]

**2.按照步骤安装即可。**

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 1][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 2][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 3][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 4][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 5][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 6][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 7][]

**3.设置环境变量**

(1)训练数据文件环境变量【环境变量必须为这个名字】

TESSDATA\_PREFIX=D:\\study\\Tesseract-OCR\\tessdata

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 8][]

(2)tesseract.exe环境变量

在Path中直接添加即可。

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 9][]

## 三、在命令行使用tesseract-ocr ##

1.将【tesseract.exe】所在的文件夹添加到环境变量里面。

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 10][]

2.使用cmd进入到想识别的图片的所在的文件夹内。

使用方法：tesseract   eg.png(文件名)    eg(识别结果存放的文件)

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 11][]

![2019112713120991.png][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 12][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 13][]

**3.识别中文**

使用方法：tesseract   eg.png(文件名)    eg(识别结果存放的文件)  -l   chi\_sim

![20191127131854874.png][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 14][]

## 四、在Python中使用tesseract-ocr ##

**1.Python需要安装对应的库**

(1)在线安装：

pip install pytesseract

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 15][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 16][]

(2)离线安装：

下载对应的.whl文件即可安装。

示例博客地址：[https://blog.csdn.net/weixin\_42830697/article/details/102492605][https_blog.csdn.net_weixin_42830697_article_details_102492605]

**2.查看对应的版本**

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 17][]

**3.测试是否安装成功**

执行导入操作未报错即表示安装成功！！

![20191127114046386.png][]

[20191127110350408.png]: https://img-blog.csdnimg.cn/20191127110350408.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70]: https://img-blog.csdnimg.cn/20191127110418229.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 1]: https://img-blog.csdnimg.cn/20191127110443901.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 2]: https://img-blog.csdnimg.cn/20191127110505957.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 3]: https://img-blog.csdnimg.cn/20191127110654849.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 4]: https://img-blog.csdnimg.cn/20191127110731607.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 5]: https://img-blog.csdnimg.cn/20191127110915226.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 6]: https://img-blog.csdnimg.cn/20191127110943399.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 7]: https://img-blog.csdnimg.cn/20191127111014860.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 8]: https://img-blog.csdnimg.cn/20191127113846189.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 9]: https://img-blog.csdnimg.cn/20191127111811554.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 10]: https://img-blog.csdnimg.cn/20191127130809818.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 11]: https://img-blog.csdnimg.cn/20191127131028179.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw==,size_16,color_FFFFFF,t_70
[2019112713120991.png]: https://img-blog.csdnimg.cn/2019112713120991.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 12]: https://img-blog.csdnimg.cn/20191127131251826.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 13]: https://img-blog.csdnimg.cn/20191127131333506.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw==,size_16,color_FFFFFF,t_70
[20191127131854874.png]: https://img-blog.csdnimg.cn/20191127131854874.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 14]: https://img-blog.csdnimg.cn/20191127131942941.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 15]: https://img-blog.csdnimg.cn/20191127112602219.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 16]: https://img-blog.csdnimg.cn/20191127112630947.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw==,size_16,color_FFFFFF,t_70
[https_blog.csdn.net_weixin_42830697_article_details_102492605]: https://blog.csdn.net/weixin_42830697/article/details/102492605
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw_size_16_color_FFFFFF_t_70 17]: https://img-blog.csdnimg.cn/20191127112733836.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MjgzMDY5Nw==,size_16,color_FFFFFF,t_70
[20191127114046386.png]: https://img-blog.csdnimg.cn/20191127114046386.png