• 主题
高级搜索  >
历史搜索 清空历史
首页> 中文期刊> 《小型微型计算机系统》 >基于深度学习模型的非法流动摊贩检测方法研究

基于深度学习模型的非法流动摊贩检测方法研究

         
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Illegal floating vendors bring certain conveniences to the lives of residents,but due to their affect to traffic block,social or-der,and food safety problems in the goods they sell,they have become the target that the city management department needs to moni-tor and manage. The difficulty in managing illegal floating vendors is their liquidity. The city management department cannot determine the time and place of its operation,which brings high management costs. This paper applies the advanced deep learning technology and the machine vision technology to the urban management task to propose an illegal floating vendor detection method based on the deep learning model. The method is trained by manually annotating the data set and using the attention mechanism and the Inception Resnet-v2 model improved Faster R-CNN to obtain the target detection model,named Faster R-CNN Inception Resnet-v2 attention model (FRIRAM). So that it can identify both illegal floating vendors,pedestrians and obtain their location information. Then collects evi-dence and notify the management department for fixed point monitoring and management. The method is verified on the test data and the results show that it has a good detection effect which also prove the application value.%流动摊贩虽然为居民生活带来一定便利,但由于其阻碍交通,影响社会秩序,出售的产品存在食品安全问题,是城市管理部门需要重点监控、管理的对象.非法流动摊贩管理的难点在于其流动性.城市管理部门无法确定其经营的时间、地点,监管难度大,管理成本高.本文将先进的深度学习和机器视觉技术应用到城市管理任务,针对实际需求提出基于深度学习模型的非法流动摊贩检测方法.该方法通过人工标注数据集训练并使用注意力机制( attention mechanism)和Inception Resnet-v2模型改进的Faster R-CNN,得到目标检测模型Faster R-CNN Inception Resnet-v2 attention模型(FRIRAM),对非法流动摊位与行人进行检测,获取位置信息.进一步进行取证,通知管理人员定点监控、管理.本文在测试数据集上进行验证,结果显示出较好的检测效果,证明本方法的应用价值.

著录项

相似文献

  • 中文文献
  • 外文文献
  • 专利
获取原文

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