首页> 外国专利> FALL DETECTION DEVICE, FALL DETECTION METHOD, FALL DETECTION CAMERA, AND COMPUTER PROGRAM

FALL DETECTION DEVICE, FALL DETECTION METHOD, FALL DETECTION CAMERA, AND COMPUTER PROGRAM

机译:坠落检测装置,坠落检测方法,坠落检测相机和计算机程序

摘要

Disclosed is a fall detection device and the like capable of detecting a fall of a human body in a highly accurate manner, even when a low-frame-rate image is used and a low-performance processing unit is employed. A fall detection device (1), characterized in being provided with: a height detection unit (2) for detecting, in response to the acquisition of input data (101), whether or not the height of the person is decreasing, on the basis of the input data (101); an appearance feature value extraction unit (3) for extracting the appearance feature value on the basis of the appearance of a human body image if the height is decreasing; an appearance detection unit (5) for establishing whether or not the human body is in a collapsed state on the basis of the result of referencing a first appearance dictionary using the appearance feature value as a basis; and a movement distance detection unit (4) for detecting, if it has been established that the human body is in a collapsed state, whether or not the movement distance of the human body included in the body image is small on the basis of the input data (101), and detecting that the human body being detected has fallen down if the movement distance is small.
机译:公开了一种跌倒检测装置等,即使在使用低帧率图像并且使用低性能处理单元的情况下,也能够以高精度方式检测人体的跌倒。跌倒检测装置(1),其特征在于,具有:高度检测单元(2),用于基于输入数据(101)的获取来检测人的身高是否在下降。输入数据(101);外观特征值提取单元(3),如果高度降低,则根据人体图像的外观提取外观特征值;外观检测单元(5),用于基于以外观特征值为基础的第一外观字典的引用结果,来确定人体是否处于塌陷状态;移动距离检测单元(4),用于在确定人体处于塌陷状态的情况下,基于输入来检测人体图像中包括的人体的移动距离是否小数据(101),并且如果移动距离小,则检测到被检测的人体已经倒下。

著录项

  • 公开/公告号WO2014010203A1

    专利类型

  • 公开/公告日2014-01-16

    原文格式PDF

  • 申请/专利权人 NEC CORPORATION;

    申请/专利号WO2013JP04119

  • 发明设计人 HOSOI TOSHINORI;

    申请日2013-07-03

  • 分类号G06T7/20;G08B25;G08B25/04;H04N5/225;H04N5/232;

  • 国家 WO

  • 入库时间 2022-08-21 15:52:14

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