首页> 外国专利> Processing and system for effective detection of events in a large number of concurrent image sequences.

Processing and system for effective detection of events in a large number of concurrent image sequences.

机译:用于有效检测大量并发图像序列中事件的处理和系统。

摘要

Procedure for use in realizacióProcedimiento for use in performing event detection and tracking objetosn event detection and object tracking in image streams, using a system, is in image streams, using a system, is formed through a series of devices I adquiá formed by a number of image acquisition devices, each installed in camposición images, each installed in field where it has detected an event and connected au where be detected an event and connected a central station through a data network, na central station through a data network, the method comprising: processing a formatcomprendiendo procedure: processing a digital format of an image stream, consisting digital ene flow image, consisting of one or more images, to detect features one or more images, for detect features, being parameters related to attributes which are parameters related to attributes of objects in said image stream, and, bas of objects in said image stream, and, based on these features, performing deteccisado in these features, performing the event detection and object tracking, ng events and tracking objects, the method being characterized by the fact the method being characterized by the fact that: each of the acquisition devices that: each image acquisition devices is preprogrammed to work with images is pre-programmed to work with a server image processing such estservidor image processing at said central station on the basis of distribution center algación on the basis distribution of image processing algorithms between eldispoori tmos image processing between dispsitivo image acquisition and server eositivo image acquisition and the server, stando controlled acquisition device deestando controlled device acquisition d images by the server to apply a exe images by the server to apply a etracción of low-level features to formaxtracción of low-level features to digital formto image stream, comprising digital diate image stream, comprising Right feature extraction under elicha level extracting low level features and analysis a frame and identifying dichol analysis frame and identifying dich frame as an event generating, from said frame as an event generating, from doctor image flow, flow cho characteristics image stream , flow characteristics that sea orrespondiente co, and transmission corresponding to said frame, and transmission dedicho flow characteristics server proc said feature stream to the server proesamiento image through the data network, cessing image through the data network provided that the characteristics exceed a limit, provided that the features exceed a corresponding limit, corresponding determined dynamically me, hrough determined dynamically said processing server imágenemediante said IMAGENS processing server; whereby the flow characteristics Indicates; whereby the flow indicaivo characteristics frame identified as containing posibtivo frame identified as containing posilemente an event is configured for procesamienblemente an event is configured for further procesamieto high level by said additional servidornto high level by said served image processing, thus making lr image processing in order to perform a event detection and tracking objetola event detection and tracking objet flow indicative characteristics of said flow indicative characteristics Physicians image streams. hos image streams.
机译:用于实现事件检测和跟踪的过程用于在系统中使用系统通过图像设备进行事件检测和跟踪对象事件检测和对象跟踪的过程是由一系列设备构成的,这些设备由多个设备组成。图像采集设备,每个都安装在Camposición图像中,每个图像采集设备都安装在已检测到事件的现场,并已连接到au,其中已检测到事件并通过数据网络连接到中央站,通过数据网络连接到中央站,该方法包括:处理格式程序:处理由数字烯流图像组成的图像流的数字格式,该图像由一个或多个图像组成,以检测特征一个或多个图像,以检测特征,这些特征是与属性有关的参数,这些参数是与属性有关的参数所述图像流中的对象的数量,以及所述图像流中的对象的数量,并基于这些特征在这些特征,执行事件检测和对象跟踪,ng事件和跟踪对象,该方法的特征在于以下事实:该方法的特征在于:每个采集设备:每个图像采集设备都经过预编程以处理图像预先编程为可与服务器图像处理配合使用,例如,在配送中心algación的基础上,在所述中央站使用雌虫图像处理,基于disdistivo图像获取和服务器eositivo图像获取之间的disdispoori tmos图像处理之间的图像处理算法分布。服务器,由服务器控制的站控制设备,由服务器控制的d图像,由服务器应用exe图像,以将低级特征的理论应用到数字形式到图像流的低级特征的最大化,包括数字图像图像流,包括elicha级提取下的Right特征提取对低级特征进行分析并确定帧,并确定dichol分析帧,并将dich帧识别为事件,从所述帧作为事件,从医生图像流,流cho特征图像流,sea orrespondiente co的流特征以及传输对应于所述帧,并且传输去中心流特征服务器将所述特征流通过数据网络传输到服务器proesamiento图像,前提是特征超过限制,则通过数据网络停止图像,前提是特征超过相应限制,则相应确定通过动态地确定所述处理服务器,动态地确定所述图像处理服务器;流动特性表示;从而将被识别为包含的流动指示特征帧,被识别为包含潜在的特征帧的事件配置为procesamienblement,将事件通过所述另外的servidornto高级配置为进一步procesamieto_high级别,从而通过服务图像处理来进行,从而进行图像处理以便执行事件检测和跟踪对象喷射事件检测和跟踪对象喷射指示所述流指示特征医师图像流。居室图像流。

著录项

  • 公开/公告号ES2288610T3

    专利类型

  • 公开/公告日2008-01-16

    原文格式PDF

  • 申请/专利权人 AGENT VIDEO INTELLIGENCE LTD;

    申请/专利号ES20030735950T

  • 发明设计人 TALMON GAD;ASHANI ZVI;

    申请日2003-07-03

  • 分类号G06T7;H04N7/18;G06K9;G06K9/36;G06T7/20;G08B13/196;

  • 国家 ES

  • 入库时间 2022-08-21 20:04:25

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