首页> 外国专利> AUTOMATIC RECOGNITION SYSTEM OF TEXT INFORMATION BASED ON DEEP LEARNING MODEL AND AUTOMATIC RECOGNITION METHOD

AUTOMATIC RECOGNITION SYSTEM OF TEXT INFORMATION BASED ON DEEP LEARNING MODEL AND AUTOMATIC RECOGNITION METHOD

机译:基于深度学习模型和自动识别方法的文本信息自动识别系统

摘要

The present invention relates to a deep learning model-based automatic character information recognition apparatus and an automatic recognition method, and the deep learning model-based automatic character information recognition system according to an embodiment of the present invention, a process piping diagram (P&ID) image is input input module; a character detection module for extracting coordinates for positions of characters included in the image of the process piping diagram (P&ID) and a character region image disposed at the character position coordinates; a character recognition module for receiving the character region image from the character detection module and predicting a character included in the character region image; and a character output module for outputting the character position coordinates extracted from the character detection module and characters predicted by the character recognition module. According to the present invention, various character information is automatically recognized from the image-type process piping diagram (P&ID) received based on the deep learning model, and the recognized character information is listed, which occurs when an existing engineer works manually. errors can be minimized.
机译:本发明涉及一种基于深度学习模型的自动字符信息识别设备和自动识别方法,以及根据本发明实施例的基于深度学习模型的自动字符信息识别系统,过程管道图(P&ID)图像是输入模块;字符检测模块,用于提取工艺管道图(P&ID)图像中包含的字符位置的坐标,以及在字符位置坐标处处理的字符区域图像;字符识别模块,用于从字符检测模块接收字符区域图像,并预测包括在字符区域图像中的字符;以及字符输出模块,用于输出从字符检测模块提取的字符位置坐标和由字符识别模块预测的字符。根据本发明,从基于深度学习模型接收的图像类型工艺管道图(P&ID)中自动识别各种字符信息,并列出当现有工程师手动工作时发生的识别字符信息。可以最小化错误。

著录项

  • 公开/公告号KR20220060162A

    专利类型

  • 公开/公告日2022-05-11

    原文格式PDF

  • 申请/专利权人 고등기술연구원연구조합;

    申请/专利号KR20200145809

  • 发明设计人 이태경;김준영;

    申请日2020-11-04

  • 分类号G06K9;G06N3/08;

  • 国家 KR

  • 入库时间 2022-08-25 00:54:56

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号