公开(公告)号：US10074251B2

公开(公告)日：2018-09-11

申请号：US15197214

申请日：2016-06-29

Applicant: Honeywell International Inc.

Inventor： Eric Oh ,  David S. Zakrewski ,  Mi Suen Lee

IPC: G08B13/196 ,  G06K9/00 ,  H04N7/18

CPC classification number: G08B13/19608 ,  G06K9/00389 ,  G06K9/00771 ,  G06K9/00885 ,  G08B13/19613 ,  H04N7/188

Abstract: Systems and methods for detecting personal distress and gesture commands in security video are provided. Some methods can include receiving a sequence of images from a video device monitoring a secured area, analyzing the sequence of images to detect a presence of a human in the sequence of images, when the presence of the human is detected in the sequence of images, analyzing the sequence of images to detect one of a plurality of contexts that requires an action in the sequence of images, and when the one of the plurality of contexts that requires the action is detected in the sequence of images, transmitting a signal to execute the action. Each of the plurality of contexts can include a respective act performed by the human or a respective condition of the human irrespective of the human being in motion or being stationary.

公开(公告)号：US20160314666A1

公开(公告)日：2016-10-27

申请号：US15197214

申请日：2016-06-29

Applicant: Honeywell International Inc.

Inventor： Eric Oh ,  David S. Zakrewski ,  Mi Suen Lee

IPC: G08B13/196 ,  G06K9/00

CPC classification number: G08B13/19608 ,  G06K9/00389 ,  G06K9/00771 ,  G06K9/00885 ,  G08B13/19613 ,  H04N7/188

Abstract: Systems and methods for detecting personal distress and gesture commands in security video are provided. Some methods can include receiving a sequence of images from a video device monitoring a secured area, analyzing the sequence of images to detect a presence of a human in the sequence of images, when the presence of the human is detected in the sequence of images, analyzing the sequence of images to detect one of a plurality of contexts that requires action in the sequence of images, and when the one of the plurality of contexts that requires the action is detected in the sequence of images, transmitting a signal to execute the action. Each of the plurality of contexts can include an act performed by the human or a condition of the human irrespective of the human being in motion or being stationary.

Abstract translation: 提供了在安全视频中检测个人遇险和手势命令的系统和方法。 一些方法可以包括：当在图像序列中检测到人的存在时，从监视安全区域的视频设备接收图像序列，分析图像序列以检测人类在图像序列中的存在， 分析图像序列以检测需要图像序列中的动作的多个上下文之一，并且当在图像序列中检测到需要动作的多个上下文中的一个时，发送信号以执行动作 。 多个上下文中的每一个可以包括由人执行的动作或人的状态，而不管人在运动或静止。