之前提及到的C++与Python之间数据交互时花费时间较大，暂定使用tesseract实现字符的识别。但是由于tesseract对中文的识别效果不好，一直在研究字符识别准确率的问题，现在中文数字字母的识别准确率可以到95%以上，基本可以媲美深度学习的识别准确率了。之前使用Python的一个流程时间是16s，现在一个流程5s，满足需求。值得继续研究。
- As mentioned before, data interaction between C++ and Python takes a lot of time, so tesseract is tentatively used to realize character recognition.However, tesseract has been studying the problem of character recognition accuracy due to its poor recognition effect on Chinese, and now the recognition accuracy of Chinese numerals and letters can reach more than 95%, basically comparable to the recognition accuracy of deep learning.One process using Python used to take 16s, but now one process is 5S, which meets the requirements.Further research is warranted.
//----------------------------------------------------------------------------------------------------------------------------------
//----------------------------------------------------------------------------------------------------------------------------------
1.测试系统环境：WIN7专业版64位操作系统 AMD A10-5757M APU 2.5GHZ
2.测试图像大小：像素：2431*3444 ，大小：2.19MB
3.测试软件版本：QT5.9.8+Python3.7.4+Opencv3.1
3.算法测试时间：
（0）程序启动 ---->打开文件夹时间： 0s
（1）C++代码 ---->表格检测时间：2.666s
（2）C++代码 ---->表格单元格分割时间：4.94s
（3）C++代码 ---->CRNN启动时间：29.946s CRNN启动时间：25.006s
（4）CRNN ---->CRNN识别时间：180.382s CRNN识别时间：150.436s
（5）C++代码 ---->对CRNN识别结果处理时间：180.458s
//----------------------------------------------------------------------------------------------------------------------------------
//----------------------------------------------------------------------------------------------------------------------------------
C++ and tesseract：
C++ and CRNN：
I hope I can help you,If you have any questions, please comment on this blog or send me a private message. I will reply in my free time.