首页> 外国专利> Method and system for automobile traction type classification and automobile number recognition based on deep-learning

Method and system for automobile traction type classification and automobile number recognition based on deep-learning

机译:基于深度学习的汽车牵引式分类和汽车数识别方法和系统

摘要

A vehicle towing type classification and vehicle number recognition method and system are provided. According to an embodiment of the present invention, the vehicle towing type classification and car number recognition method is a vehicle towing type classification and car number recognition system, using a towing image of a car as learning data, a Deep Neural Network (DNN) based a learning step of learning the learning model; And when the towing image is input, the car towing type classification and car number recognition system performs a classification task of classifying the towing type of the car in the input towing image based on the learned learning model and a recognition task of recognizing the car number service processing step; Thereby, by classifying the tow image based on the DNN and recognizing the number of the vehicle in the tow picture, it is verified whether the towed vehicle in the towing picture matches the actual vehicle, even if the insurance company employee or the owner does not directly check it. Furthermore, parallel processing of multiple towing images is possible, reducing human time/time, and improving the efficiency of towing service processing by building automated services in insurance companies.
机译:提供了一种车辆牵引式分类和车辆号识别方法和系统。根据本发明的一个实施例,车辆牵引型分类和汽车数识别方法是车辆牵引型分类和汽车数识别系统,使用汽车的牵引图像作为学习数据,基于深度神经网络(DNN)学习学习模式的学习步骤;并且当输入牵引图像时,汽车牵引型分类和汽车数识别系统的分类任务基于所学的学习模型和识别汽车数的识别任务来分类输入牵引图像中的牵引类型的牵引类型。服务处理步骤;因此,通过基于DNN对丝网图像进行分类并识别牵引图像中的车辆的数量,验证牵引车中的牵引车是否与实际车辆匹配,即使保险公司员工或所有者没有直接检查一下。此外,可以通过在保险公司中建立自动化服务来提高人类时间/时间,降低人类的时间/时间并提高牵引服务处理的效率。

著录项

  • 公开/公告号KR20210098109A

    专利类型

  • 公开/公告日2021-08-10

    原文格式PDF

  • 申请/专利权人 에스케이 주식회사;

    申请/专利号KR1020200011890

  • 发明设计人 안호정;

    申请日2020-01-31

  • 分类号G06K9/62;G06K9/32;G06N3/08;G06T3/40;G06T5;

  • 国家 KR

  • 入库时间 2022-08-24 20:30:57

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