WO2012148258A1 - Système et procédé de détection de mouvements brusques d'objets - Google Patents

Système et procédé de détection de mouvements brusques d'objets Download PDF

Info

Publication number
WO2012148258A1
WO2012148258A1 PCT/MY2012/000094 MY2012000094W WO2012148258A1 WO 2012148258 A1 WO2012148258 A1 WO 2012148258A1 MY 2012000094 W MY2012000094 W MY 2012000094W WO 2012148258 A1 WO2012148258 A1 WO 2012148258A1
Authority
WO
WIPO (PCT)
Prior art keywords
frame rate
abrupt
displacement information
object movement
detection system
Prior art date
Application number
PCT/MY2012/000094
Other languages
English (en)
Inventor
Zulaikha Kadim
Kim Meng Liang
Norshuhada Samudin
Original Assignee
Mimos Berhad
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Mimos Berhad filed Critical Mimos Berhad
Publication of WO2012148258A1 publication Critical patent/WO2012148258A1/fr

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/20—Analysis of motion
    • G06T7/254—Analysis of motion involving subtraction of images
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10016—Video; Image sequence
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30196—Human being; Person
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30232—Surveillance
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30236—Traffic on road, railway or crossing

Definitions

  • the present invention relates generally to a detection system, more particularly to a detection system for detecting abrupt object movement in inconsistent frame rate and the method thereof .
  • Video surveillance systems are widely used apparatuses to detect and monitor objects within a monitoring area. Usually those video surveillance systems are used to detect and track individuals or vehicles entering or leaving a building facility or to monitor individuals within the monitoring area where the safety of the occupants may be of concern. And another example of application for the surveillance system is to monitor crowds as surveillance cameras are low cost and require no physical contact with the object being monitored.
  • the inconsistent frame rate includes low frame rate that resulting in lagging issue. Therefore when the video input containing lagging problem is converted to image frame, there will be some missing frame of images due to lagging video as shown in Figures la and lb. This situation gives large error because the movement of moving object is measured by measuring the different locations of the moving object between every two frame of images. The higher of the differences will be interpreted as higher movement. Therefore, if there is some missing frame of images, the different of location of moving object between two frames will be higher. Then the system will interpret it as high movement even though there is no high movement in the sample data as shown in Figure lc. Hence, false positive happened which is when no any abnormalities in the monitoring area but the system triggered the detection.
  • the present invention aims to provide a system and method to detect abrupt object movement with the ability to work in unstable network is needed.
  • an abrupt object movement detection system comprises at least one capturing device for capturing a sequence of image frames at a monitoring area, means for registering frame of images having different types of frame rates, wherein motion blobs of the images are extracted to obtain object's displacement information for categorizing into predefined categories of frame rates and the maximum displacement information for each frame rate is chosen as a filtering threshold for the respective frame rate categories; and means for detecting abrupt object movement by filtering a scene displacement information with the filtering threshold based on a corresponding predefined frame rate category and determining abrupt movement from total density of the filtered displacement information.
  • a method for detecting abrupt object movement with a detection system comprising the steps of capturing a sequence of image frames at a monitoring area as input images for processing, registering frame of images having different types of frame rates by extracting motion blobs of the images to obtain object's displacement information for categorizing into predefined categories of frame rates and determining the maximum displacement information as a filtering threshold for the respective frame rate categories, and detecting an abrupt object movement event by filtering a scene displacement information with the filtering threshold based on a corresponding predefined frame rate category and determining abrupt movement from total density of the filtered displacement information.
  • Figure la shows a series of frame of image converted from stream of video for a similar scene at different frame rate of an existing detection system
  • Figure lb shows a sequence of frame of image having some of the frames of image missing due to unstable network issuer-
  • Figure lc illustrates two frames of images which are interpreted as high movement due to the issue shown in figure lb;
  • Figure 2 is a flowchart showing the method for the registration process of the abrupt object movement detection system of the present invention
  • Figure 3a depicts an original input image
  • Figure 3b shows the image of figure 3a with extracted motion blobs
  • Figure 3c is the image of figure 3b showing the displacement information
  • Figure 4 is a schematic diagram showing the step of buffering the displacement information and examining the information to determine the frame rate category;
  • Figure 5 is a flowchart showing the method for the step of detecting process of the of the abrupt object movement detection system of the present invention
  • Figure 6 shows a schematic diagram showing the displacement information within a sliding window filtered based on the selected filtering threshold and the remaining displacement information
  • Figure 7 illustrates an example of connected movement information .
  • the abrupt object movement detection system (10) includes at least one capturing device (11) for capturing a sequence of image frames at a monitoring area, means for registering (12) frame of images having different types of frame rates for acquiring optimal filtering threshold, means for detecting (13) abrupt object movement by filtering the displacement information with the filtering threshold based on the frame rate and determining from the total density of the filtered movements, and a displaying device (40) for alerting the user if the abrupt object movement event is detected.
  • frame rates There are three different types of frame rates which include the low, medium and high frame rate.
  • the means for registering (12) performs the registration process as shown in Figure 2 and the means for detecting (13) performs the online detection process as shown in Figure 4.
  • the system (10) will receive a sequence of image frames which is captured by the capturing device (11). These image frames will then be processed (14) by extracting all moving pixels in the current frame and in this case, current frame is compared to the background image for detecting motion blobs (15).
  • the moving pixels are called motion pixels and they represent the moving object of interest within the current frame. Group of connected motion pixels is called motion blobs.
  • Figure 3a shows an original input image, follows by the example of extracted motion blobs (21) shown in Figure 3b and the displacement information which is the optical flow vector shown in Figure 3c.
  • the extracted motion blobs With the extracted motion blobs, the current motion blob of the same object and its previous motion blob in previous frame are compared and the optical flow is computed to extract the displacement information (16) .
  • the extracted motion blobs are used as the mask in computing the optical flow, whereby the stray motion vector resultant from the background pixels will be ignored.
  • each object's displacement information is computed, the cumulative or average of the data within current frame is then computed and the value is stored into a buffer.
  • the registration process followed by the step of buffering the displacement information (17) and examining in a sliding window to classify (18) them into three different types of frame rates namely the low, medium and high frame rate as shown in Figure .
  • each sliding window it contains the same number of image frame but the time duration for each window is different and it is depending on the network.
  • the frame rate for each sliding window is determined by dividing the number of frames to time duration in second of the window. Then the sliding window is classified into low frame rate category if the computed frame rate is within 0 to 6 fps (frame per second), or into medium frame rate category if the computed frame rate is within 7 to 12 fps or high frame rate category if the computed frame rate within 13 to 25 fps.
  • the maximum displacement data in each sliding window is extracted (19) .
  • all maximum displacement data for each frame rate categories is compared (20) and the maximum value is chosen as the filtering threshold for that frame rate category. Therefore, three optimal filtering thresholds for three different frame rate categories are obtained .
  • the initial processes are the same as in the registration process.
  • the moving objects are extracted from the scene (31) which is the same as in step (15) in the registration process.
  • the displacement information of each object in the scene is extracted (32).
  • the scene displacement information which is the average displacement information in current frame is computed.
  • the scene displacement information is stored into a buffer.
  • the system (10) then performs the step of buffering the displacement (33) wherein if the buffer size is more than the minimum window size, the window is examined to determine its frame rate category (34).
  • the filtering threshold is selected accordingly.
  • the displacement information within the sliding window is filtered based on the selected filtering threshold. If it is lower than the selected filtering threshold, the displacement information will be removed (35) as shown in Figure 6.
  • the detection process is followed by grouping the remaining displacement information (36) . In this step, every single magnitude will be linking to each other with their connected neighbor and the example of connected displacement information is shown in Figure 7.
  • the total density of each group of displacement information is then calculated (37) .
  • the total density of each group is then compared to the sensitivity threshold to determine (38) the event of an abrupt object movement. If an event is detected, the system will alert the user by displaying the information on the displaying device (40) .

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Image Analysis (AREA)

Abstract

L'invention concerne un système (10) de détection de mouvements brusques d'objets comportant au moins un dispositif (11) de capture servant à capturer une séquence de vues d'image dans une zone de surveillance, un moyen servant à aligner (12) les vues d'images caractérisées par des types différents de cadences d'images, des taches de mouvement des images étant extraites pour obtenir des informations de déplacement d'un objet en vue d'une classification en catégories prédéfinies de cadence d'images et les informations de déplacement maximal pour chaque vue étant choisies comme seuil de filtrage pour les catégories respectives de cadence d'images ; et un moyen servant à détecter (13) un mouvement brusque d'objet en filtrant les informations de déplacement d'une scène avec le seuil de filtrage basé sur une catégorie correspondante prédéfinie de cadence d'images et à déterminer un mouvement brusque à partir de la densité totale des informations de déplacement filtrées.
PCT/MY2012/000094 2011-04-29 2012-04-27 Système et procédé de détection de mouvements brusques d'objets WO2012148258A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
MYPI2011001929A MY178641A (en) 2011-04-29 2011-04-29 Abrupt object movement detection system and method thereof
MYPI2011001929 2011-04-29

Publications (1)

Publication Number Publication Date
WO2012148258A1 true WO2012148258A1 (fr) 2012-11-01

Family

ID=47072573

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/MY2012/000094 WO2012148258A1 (fr) 2011-04-29 2012-04-27 Système et procédé de détection de mouvements brusques d'objets

Country Status (2)

Country Link
MY (1) MY178641A (fr)
WO (1) WO2012148258A1 (fr)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105308876A (zh) * 2012-11-29 2016-02-03 康宁光电通信有限责任公司 在多输入、多输出(mimo)分布式天线系统(das)中的混合式小区内/小区间远程单元天线结合
US10055669B2 (en) 2016-08-12 2018-08-21 Qualcomm Incorporated Methods and systems of determining a minimum blob size in video analytics
US10169661B2 (en) 2014-03-28 2019-01-01 International Business Machines Corporation Filtering methods for visual object detection

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5311306A (en) * 1991-07-19 1994-05-10 Kabushiki Kaisha Toshiba Motion detecting circuit for video signal processor
US20070104462A1 (en) * 2005-11-10 2007-05-10 Sony Corporation Image signal processing device, imaging device, and image signal processing method
US20100265344A1 (en) * 2009-04-15 2010-10-21 Qualcomm Incorporated Auto-triggered fast frame rate digital video recording

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5311306A (en) * 1991-07-19 1994-05-10 Kabushiki Kaisha Toshiba Motion detecting circuit for video signal processor
US20070104462A1 (en) * 2005-11-10 2007-05-10 Sony Corporation Image signal processing device, imaging device, and image signal processing method
US20100265344A1 (en) * 2009-04-15 2010-10-21 Qualcomm Incorporated Auto-triggered fast frame rate digital video recording

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105308876A (zh) * 2012-11-29 2016-02-03 康宁光电通信有限责任公司 在多输入、多输出(mimo)分布式天线系统(das)中的混合式小区内/小区间远程单元天线结合
CN105308876B (zh) * 2012-11-29 2018-06-22 康宁光电通信有限责任公司 分布式天线系统中的远程单元天线结合
US10169661B2 (en) 2014-03-28 2019-01-01 International Business Machines Corporation Filtering methods for visual object detection
US10055669B2 (en) 2016-08-12 2018-08-21 Qualcomm Incorporated Methods and systems of determining a minimum blob size in video analytics

Also Published As

Publication number Publication date
MY178641A (en) 2020-10-19

Similar Documents

Publication Publication Date Title
CN109887281B (zh) 一种监控交通事件的方法及系统
CN104966304B (zh) 基于卡尔曼滤波与非参数背景模型的多目标检测跟踪方法
JP5644097B2 (ja) 画像処理装置、画像処理方法及びプログラム
AU2014240669B2 (en) Object monitoring system, object monitoring method, and monitoring target extraction project
KR101910542B1 (ko) 객체 검출을 위한 영상분석 서버장치 및 방법
US10853949B2 (en) Image processing device
KR102002812B1 (ko) 객체 검출을 위한 영상분석 서버장치 및 방법
KR101541272B1 (ko) 사람들의 움직임 불규칙성을 이용한 폭력 행위 검출 장치 및 방법
CN101483001A (zh) 基于视频的入侵检测方法、检测装置、芯片
KR101750094B1 (ko) 영상의 실시간 모니터링을 통한 비정상 집단행동 분류 방법
CN108280953A (zh) 视频侦测报警方法及装置、存储介质、摄像头
KR20090044957A (ko) 도난 및 방치 감시 시스템 및 도난 및 방치 검출 방법
WO2012148258A1 (fr) Système et procédé de détection de mouvements brusques d'objets
US8842178B2 (en) Supervising system for image
KR101214858B1 (ko) 클러스터링 기법을 이용한 이동물체 검출장치 및 방법
US9001207B1 (en) Apparatus and method for motion detection in video
KR100887942B1 (ko) 실시간 이상현상 감지 시스템 및 그 제어방법
KR101395666B1 (ko) 비디오 영상의 변화를 이용한 도난 감시 장치 및 방법
CN109215150A (zh) 人脸点名与计数方法及其系统
CN104392201A (zh) 一种基于全向视觉的人体跌倒识别方法
Lee et al. Multiple object tracking for fall detection in real-time surveillance system
KR101848367B1 (ko) 모션벡터와 dct 계수를 이용한 메타데이터 기반 의심영상 구분식별 방식의 영상 관제 방법
US11341773B2 (en) Detection device and control method of the same
Van Den Hengel et al. Finding camera overlap in large surveillance networks
CN110572618B (zh) 一种非法拍照行为监控方法、装置及系统

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 12776518

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 12776518

Country of ref document: EP

Kind code of ref document: A1