CN105718886A - Moving personnel safety abnormity tumbling detection method - Google Patents
Moving personnel safety abnormity tumbling detection method Download PDFInfo
- Publication number
- CN105718886A CN105718886A CN201610038406.8A CN201610038406A CN105718886A CN 105718886 A CN105718886 A CN 105718886A CN 201610038406 A CN201610038406 A CN 201610038406A CN 105718886 A CN105718886 A CN 105718886A
- Authority
- CN
- China
- Prior art keywords
- pixel
- foreground
- point
- carried out
- detection method
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Abstract
The invention discloses a moving personnel safety abnormity tumbling detection method. The detection method is characterized in that S1, background modeling can be carried out, and current behavior motion characteristic image information of a moving body can be acquired, and a time sequence model of each pixel can be acquired by calculating the current image information by adopting the CodeBook algorithm; S2, prospect detection can be carried out; S3, prospect detection pre-judgment can be carried out; S4, pixel classification can be carried out, and the pixels belonging to the same target can be classified by adopting the connected area marking algorithm; S5, tumbling deduction can be carried out by adopting an aspect ratio. By adopting the technical scheme, the moving condition of the elderly in the room can be identified and controlled; when the elderly tumbles or cannot move for a long time, the nursing personnel can be notified, and the alarming signals can be sent, and therefore the nursing and the caring of the elderly can be provided timely, and the life danger caused by the delayed alarming can be prevented.
Description
Technical field
The invention belongs to the technical field of image procossing, particularly relate to the fall detection method that a kind of activity personal security is abnormal.
Background technology
Present market is provided with intelligent camera first-class monitoring device all, but the security exception situation of Intelligent Recognition indoor activity personnel cannot give early warning simultaneously.
Summary of the invention
The technical problem to be solved in the present invention is to provide and a kind of judges indoor Falls Among Old People the method providing warning based on image procossing.
For solving the problems referred to above, the present invention adopts the following technical scheme that:
The fall detection method that a kind of activity personal security is abnormal comprises the following steps:
Step S1, background modeling
Obtain the current behavior motion characteristic image information of moving human body, present image information is carried out CodeBook algorithm and obtains the time series models of each pixel, setting up a CodeBook structure to each pixel of present image information, each CodeBook comprises multiple CodeWord (CW);
Step S2, prospect are detected
If the value of a certain pixel is not in CodeWords, being labeled as foreground point by this pixel, each pixel judges all foreground pixel points obtaining this frame of video in image;
The detecting of step S3, prospect is pre-to be judged
Pre-determination methods comprises the following steps:
First, picture element matrix is separated into a grid, the neglecting greatly concrete situation and determine of grid,
Then, take the CoodBook of the angle point of each grid, it judged,
Secondly, when there being prospect to occur in screen, foreground object will cover at least one or more grid, therefore the CoodBook group taken out is arranged threshold value, when the number of foreground pixel has exceeded this threshold value, then judge that this frame comprises foreground object, thus carrying out judging to whole pixels and processing;
Step S4, pixel are sorted out
Adopt UNICOM's zone marker algorithm that the pixel belonging to same target is sorted out.
Step S5, deduction of falling
Aspect ratio is adopted to carry out deduction of falling, when aspect ratio ratio is lower than a threshold value, and when continuing for some time, then it is assumed that namely this target person is in the state of falling, and will send alarm.
As preferably, step S4 particularly as follows:
Point on a, progressive scanning picture, checks whether each point is foreground point, if it is not, continue to scan on;
B, check that whether this foreground point identical with the numbering of point around and labelling label label.
C, again scanogram, utilize Union-find Sets algorithm to merge the region of different label of UNICOM, after second step pass, after the different target in same frame will be classified out, can carry out next step and carry out, according to aspect ratio, detecting of falling
The present invention active situation by camera head monitor identification indoor old man, when old man falls in indoor or in long-time irremovable situation, notifies caregiver, it is provided that alarm signal, make old man obtain timely nurse and look after, it is to avoid occur because of can not alarm in time and life problem occurs.
Accompanying drawing explanation
Fig. 1 is the flow chart of fall detection method of the present invention;
Fig. 2 is the enforcement technique effect figure of the inventive method.
Detailed description of the invention
As it is shown in figure 1, the embodiment of the present invention provides the fall detection method that a kind of activity personal security is abnormal, comprise the following steps:
Step S1, background modeling
By the moving human body of monitoring in monitoring camera, obtain the current behavior motion characteristic image information of moving human body;
Present image information being carried out CodeBook algorithm and obtains the time series models of each pixel, detailed process is:
Setting up CodeBook (CB) structure to each pixel of present image information, each CodeBook has again multiple CodeWord (CW) to form.
The form of CB and CW is as follows:
CB={CW1, CW2 ... CWn, t}
CW={lHigh, lLow, max, min, t_last, stale}
Wherein, n is the number of the CW comprised in a CB, when n is too little, deteriorates to simple background, can complex background be modeled when n is bigger;T is the CB number of times updated.CW is 6 tuples, and wherein IHigh and ILow is as the lower bound in study when updating, and max and min records maximum and the minima of current pixel.Time t_last and outmoded time stale (recording how long this CW is not accessed) that last time updates are used for deleting the CodeWord being rarely employed.
Step S2, prospect are detected
If the value of a certain pixel is not in CodeWords, being labeled as foreground point by this pixel, each pixel judges all foreground pixel points obtaining this frame of video in image.
Conventional foreground judge process needs each pixel is modeled, and in real work in most of the cases, video is the people or the object that are absent from movement.Therefore add pre-determination methods for program and will reduce the dry running time of program, thus improving operational efficiency.
The detecting of step S3, prospect is pre-to be judged
Owing to the pixel in screen is the arrangement of two-dimensional matrix form, therefore the pre-determination methods of the present invention comprises the following steps:
First, picture element matrix is separated into a grid, the neglecting greatly concrete situation and determine of grid.
Then, take the CoodBook of the angle point of each grid, it is judged.
Secondly, when there being prospect to occur in screen, foreground object will cover at least one or more grid, therefore the CoodBook group taken out is arranged threshold value, when the number of foreground pixel has exceeded this threshold value, then judge that this frame comprises foreground object, carry out judging to whole pixels thus transferring to and process.
Step S4, pixel are sorted out
Owing to work before simply marks foreground pixel point individually, the pixel belonging to same target is not sorted out.UNICOM's zone marker (ConnectedComponentLabeling) is a technology conventional inside image procossing, and it is used to detect the region of UNICOM in bianry image, serves as the effect of target area detection in many tracing detection algorithms.
Common CCL (ConnectedComponentLabeling) includes the method for Two-Pass and the method for One-Pass, and pass is exactly the pass of scanning here, and Two-Pass scans the algorithm of twice.Concrete step is as follows:
Point on a, progressive scanning picture, checks whether each point is foreground point, if it is not, continue to scan on;
B, check the point on the left side of this foreground point and the point of top, and label is set to corresponding point.
C, again scanogram, utilize Union-find Sets algorithm to merge the region of different label of UNICOM, after second step pass, after the different target in same frame will be classified out, can carry out next step and carry out, according to aspect ratio, detecting of falling.
Step S5, deduction of falling
The present invention adopts aspect ratio to carry out deduction of falling, when namely the profile of foreground target is more than a certain threshold values, then it is assumed that this target person is in standing state, this length-width ratio is lower than a threshold value, and when continuing for some time, then it is assumed that namely this target person is in the state of falling, and will send alarm.
As in figure 2 it is shown, when action figure moves by prospect task and background separation, foreground people task is shown as white, and unrelated background parts shows black.Aspect ratio according to action figure judges that current task is upright or falls turned upside down, and when erectility, personage's display box is shown in green.When aspect ratio detection for fall state time, and the time exceeded a certain threshold values, then it is assumed that action figure falls, and task display box is shown in red and sends alarm to system, prompting caregiver.
The present invention relates to the image processing method of camera head monitor, for the indoor nurse place such as nursing house, family, hospital, school, nursery school, by image procossing fall detection algorithm, identify the alarm in some security exception situations of indoor activity personnel, award caregiver in time currently by the safety prompt function of caregiver and warning.
The present invention improves early warning precision, saves the resource consumption of system, improves program operation speed, it is achieved that the situation of falling of quick and precisely identification activity personnel, it is possible to quickly response, precise positioning, and the front and back process that the personnel that can play back fall.Understand front and back process of falling to caregiver and retain foundation.
Above example is only the exemplary embodiment of the present invention, is not used in the restriction present invention, and protection scope of the present invention is defined by the claims.The present invention in the essence of the present invention and protection domain, can be made various amendment or equivalent replacement by those skilled in the art, and this amendment or equivalent replacement also should be regarded as being within the scope of the present invention.
Claims (2)
1. the fall detection method that an activity personal security is abnormal, it is characterised in that comprise the following steps:
Step S1, background modeling
Obtain the current behavior motion characteristic image information of moving human body, present image information is carried out CodeBook algorithm and obtains the time series models of each pixel, setting up a CodeBook structure to each pixel of present image information, each CodeBook comprises multiple CodeWord (CW);
Step S2, prospect are detected
If the value of a certain pixel is not in CodeWords, being labeled as foreground point by this pixel, each pixel judges all foreground pixel points obtaining this frame of video in image;
The detecting of step S3, prospect is pre-to be judged
Pre-determination methods comprises the following steps:
First, picture element matrix is separated into a grid, the neglecting greatly concrete situation and determine of grid,
Then, take the CoodBook of the angle point of each grid, it judged,
Secondly, when there being prospect to occur in screen, foreground object will cover at least one or more grid, therefore the CoodBook group taken out is arranged threshold value, when the number of foreground pixel has exceeded this threshold value, then judge that this frame comprises foreground object, thus carrying out judging to whole pixels and processing;
Step S4, pixel are sorted out
Adopt UNICOM's zone marker algorithm that the pixel belonging to same target is sorted out.
Step S5, deduction of falling
Aspect ratio is adopted to carry out deduction of falling, when aspect ratio ratio is lower than a threshold value, and when continuing for some time, then it is assumed that namely this target person is in the state of falling, and will send alarm.
2. the fall detection method that activity personal security as claimed in claim 1 is abnormal, it is characterised in that step S4 particularly as follows:
Point on a, progressive scanning picture, checks whether each point is foreground point, if it is not, continue to scan on;
B, check that whether this foreground point identical with the numbering of point around and labelling label label.
C, again scanogram, utilize Union-find Sets algorithm to merge the region of different label of UNICOM, after second step pass, after the different target in same frame will be classified out, can carry out next step and carry out, according to aspect ratio, detecting of falling.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610038406.8A CN105718886A (en) | 2016-01-20 | 2016-01-20 | Moving personnel safety abnormity tumbling detection method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610038406.8A CN105718886A (en) | 2016-01-20 | 2016-01-20 | Moving personnel safety abnormity tumbling detection method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN105718886A true CN105718886A (en) | 2016-06-29 |
Family
ID=56147761
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610038406.8A Pending CN105718886A (en) | 2016-01-20 | 2016-01-20 | Moving personnel safety abnormity tumbling detection method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105718886A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106373336A (en) * | 2016-08-30 | 2017-02-01 | 苏州品诺维新医疗科技有限公司 | Fall detection method and device |
CN107679518A (en) * | 2017-10-27 | 2018-02-09 | 深圳极视角科技有限公司 | A kind of detecting system |
CN111127838A (en) * | 2018-10-30 | 2020-05-08 | 物流及供应链多元技术研发中心有限公司 | System and method for detecting inactive objects |
CN114999108A (en) * | 2022-08-03 | 2022-09-02 | 杭州乐湾科技有限公司 | Old people falling detection method based on image processing |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20080106599A1 (en) * | 2005-11-23 | 2008-05-08 | Object Video, Inc. | Object density estimation in video |
CN102722721A (en) * | 2012-05-25 | 2012-10-10 | 山东大学 | Human falling detection method based on machine vision |
-
2016
- 2016-01-20 CN CN201610038406.8A patent/CN105718886A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20080106599A1 (en) * | 2005-11-23 | 2008-05-08 | Object Video, Inc. | Object density estimation in video |
CN102722721A (en) * | 2012-05-25 | 2012-10-10 | 山东大学 | Human falling detection method based on machine vision |
Non-Patent Citations (4)
Title |
---|
V VISHWAKARMA等: "Automatic Detection of Human Fall in Video", 《PATTERN RECOGNITION & MACHINE INTELLIGENCE SECOND INTERNATIONAL CONFERENCE》 * |
冷斌: "室外监控环境下运动目标检测算法的研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
迟健男等: "《视觉测量技术》", 30 June 2011, 机械工业出版社 * |
雷帮军等: "《视频目标跟踪系统分步详解》", 31 December 2015, 国防工业出版社 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106373336A (en) * | 2016-08-30 | 2017-02-01 | 苏州品诺维新医疗科技有限公司 | Fall detection method and device |
CN107679518A (en) * | 2017-10-27 | 2018-02-09 | 深圳极视角科技有限公司 | A kind of detecting system |
CN111127838A (en) * | 2018-10-30 | 2020-05-08 | 物流及供应链多元技术研发中心有限公司 | System and method for detecting inactive objects |
CN114999108A (en) * | 2022-08-03 | 2022-09-02 | 杭州乐湾科技有限公司 | Old people falling detection method based on image processing |
CN114999108B (en) * | 2022-08-03 | 2022-11-29 | 杭州乐湾科技有限公司 | Old people falling detection method based on image processing |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Gong et al. | A real-time fire detection method from video with multifeature fusion | |
CN112784662A (en) | Video-based fall risk evaluation system | |
KR101607934B1 (en) | The method for monitoring communicable disease and system using the method, recording medium for performing the method | |
CN105718886A (en) | Moving personnel safety abnormity tumbling detection method | |
CN103729858B (en) | A kind of video monitoring system is left over the detection method of article | |
CN105631418A (en) | People counting method and device | |
CN109190475A (en) | A kind of recognition of face network and pedestrian identify network cooperating training method again | |
WO2018120634A1 (en) | Method and apparatus for identifying disease and insect damage | |
CN113992886B (en) | Motion detection method for motion sensor | |
CN109410497A (en) | A kind of monitoring of bridge opening space safety and alarm system based on deep learning | |
CN108460370A (en) | A kind of fixed poultry life-information warning device | |
CN103903020A9 (en) | A kind of fire image recognition methods and device based on CodeBook | |
KR20210133827A (en) | Apparatus for learning deep learning model and method thereof | |
CN110688969A (en) | Video frame human behavior identification method | |
JP2018142173A (en) | Image monitoring device | |
CN107729811B (en) | Night flame detection method based on scene modeling | |
CN106844394A (en) | Based on the video retrieval method that the upper and lower clothing color of pedestrian differentiates | |
CN112633179A (en) | Farmer market aisle object occupying channel detection method based on video analysis | |
JP6275022B2 (en) | Image monitoring device | |
CN111191499A (en) | Fall detection method and device based on minimum center line | |
CN109740527A (en) | Image processing method in a kind of video frame | |
JP2014157453A (en) | Image processing apparatus, image processing method, and image processing program | |
CN106096499A (en) | A kind of video image culminant star moon pattern detection method and system | |
Huang et al. | A method for fast fall detection based on intelligent space | |
JP6155106B2 (en) | Image sensor |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20160629 |
|
RJ01 | Rejection of invention patent application after publication |