CN101847204A - Abnormal face detecting method for embedded digital video recorder (DVR) system - Google Patents
Abnormal face detecting method for embedded digital video recorder (DVR) system Download PDFInfo
- Publication number
- CN101847204A CN101847204A CN 201010138005 CN201010138005A CN101847204A CN 101847204 A CN101847204 A CN 101847204A CN 201010138005 CN201010138005 CN 201010138005 CN 201010138005 A CN201010138005 A CN 201010138005A CN 101847204 A CN101847204 A CN 101847204A
- Authority
- CN
- China
- Prior art keywords
- people
- face
- image
- detect
- dvr
- 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
Images
Landscapes
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention belongs to the technical field of video frequency image analysis and particularly discloses an abnormal face detecting method for embedded DVR system. In the invention, face detection and background identification technology is applied in a digital signal processor (DSP) processing step of the embedded DVR so as to determine that whether people exist in a current image and whether facial features are clear and not sheltered, and if abnormal facial features are detected, an alarm signal is sent to remind related personnel to pay more attention to the current conditions. When the method of the invention is adopted, the technology of detecting whether normal faces exit in a scene is achieved simply by a mode of image processing, and alarm effect that lawbreakers are prevented from committing a crime at an ATM machine of self-service banking by shielding faces is achieved when the method is used in combination with the embedded DVR.
Description
Technical field
The invention belongs to the video image analysis technical field, be meant a kind of abnormal face detecting method that is used for embedded digital video recorder (DVR) system especially.
Background technology
At present, DVR (DVR or digital video recorder) has been substituted analog vtr and has been applied to various fields, and DVR is made of parts such as VI (video acquisition), DSP (digital signal processing), VS (video storage) and VO (video output) usually.Traditional DVR only finishes the work of video compress storage in the DSP part.
Intelligent, networking, integrated are the recent development directions in DVR field.It is exactly the detection and Identification of people's face that a very big application is arranged in the intellectuality.Human face detection tech was risen 20th century 60, the seventies, had experienced the development of decades, had successively used multiple technologies, from initial template matches technology multiple mode till now, all was greatly improved on precision and speed.On precision and speed, reach the requirement of practicality at present based on the adaboost algorithm of feature, had various people's face testing products to come out in succession.
At present, mainly concentrate in gate control system and the doorway security system.Present human face detection tech unusual maturation, but some specific application modes have appearred in some specific industries, in the ATM unmanned plane such as bank, in unserviced ATM room, detection has just proposed new problem for people's face, not only want to detect normal people's face, also want to detect unusual people's face, such as situation masked or that deliberately block facial characteristics.In these cases, hazardous act such as take place that malice is withdrawn the money possibly if can detect this situation, can significantly reduce the generation of hazard event, reduces social harm, and great social effect is arranged.But present human face detection tech can't address this problem.
At first, the problem that solve is when to begin to detect, because only the someone operates, promptly has man-hour could begin to detect before the camera really, and no one's words just can not detect; Secondly need carry out the detection of a period of time just can reach a conclusion, can not only lean on the image of a frame to solve, because present people's face detection limit is still than higher, generally can only carry out the detection of front face, but actual conditions are the facial images that can not guarantee to obtain a front, so just can draw the solution conclusion after must detecting a period of time; Once more, how to solve same people detection repeatedly or a plurality of people detection problem once, because the time length difference that different people operates before ATM, a people walks back another one people and comes over, should regard twice detection as, so just can reach the purpose of not omission.
For these reasons, present people's face detects can't satisfy above-mentioned this special application, can only carry out application simple, that restrictive condition is relatively stricter.
Summary of the invention
The object of the present invention is to provide a kind of abnormal face detecting method that is used for embedded digital video recorder (DVR) system.
Technical scheme of the present invention is:
A kind of abnormal face detecting method that is used for embedded digital video recorder (DVR) system, in the DSP of embedded digital video recorder (DVR) processing links, increasing people's face detects and the background recognition technology, to be able to discern whether the people is arranged in the present image, this human face's feature whether block by clear nothing, if finding has the abnormal face feature, promptly send alerting signal, remind the related personnel that present case is mentioned hig diligence.
A kind of abnormal face detecting method that is used for embedded digital video recorder (DVR) system, its method step comprises:
Whether the people is arranged in step 1, the detection prospect:
Take background modeling, extract the method for prospect and judge before the camera whether the people is arranged, owing to only be used for judging that people and nobody are arranged, simple relatively, obtain background with moving average method, concrete steps are as follows:
A), at first, import first two field picture, requiring can not have foreground object in first two field picture, just background object;
B), with first current frame edge image image K0 as a setting;
C), continue input picture, arrive image Kd with what current images K1 and background image K0 did that the frame difference operates;
D), Kd done the corrosion operation of expanding obtain foreground image Kd ', the ultimate value by calculating K d ' zone obtains the area S of the rectangle that Kd ' constitutes;
E), with S and preset threshold St in advance relatively (St can obtain by the reference material size that the user sets), when S＞St, assert in the current frame image has the people; Otherwise have no talent in the identification current frame image, and deposit present image K1 in buffer memory;
F), continue the operation of C-E, 4 two field pictures are arranged in buffer memory, this 4 two field picture and background image are done average calculating operation obtain new background image, the while, K0 emptied;
G), repeating step C-F can constantly upgrade background, and whether the someone occurs to detect present frame, occurs if detect the someone, provides signal, starts people's face and detects link;
Step 2, detect the unusual people's face that whether occurred in the scene:
Adopt and to detect people's face in a period of time, if having people's face to occur then think normal condition; If continue to monitor face less than the normal person, think that then unusual people's face provides warning, concrete steps are as follows:
A), have the people to occur in the sight that detects through step 1, start people's face detection module this moment;
B), continue whether to have in the detected image people's face to occur, be T detection time, through time of people's normal running before ATM of statistics about 3-5 minute, so T got 1 minute; With frame per second is 25 calculating, can carry out 1500 detections in one minute, blocks if not malice, has enough detected the normal people's face of a frame;
C) if detect people's face in the time at T, then think normal condition, return normal condition this moment, and withdraw from people's face and detect, and waits for next time starting; If do not detect people's face, then enter into next step D;
D), return abnormal conditions this moment, and withdraws from people's face and detect, wait for next time starting;
Step 3, judge whether the people of current operation leaves:
No matter step 2 is returned any situation, should wait for that all current operator restarts detection next time after finishing, and concrete steps are as follows:
A), the input present image, arrive image Kd with what current images K1 and background image K0 did that the frame difference operates;
B), Kd done the corrosion operation of expanding obtain foreground image Kd ', the ultimate value by calculating K d ' zone obtains the area S of the rectangle that Kd ' constitutes;
C), with S and preset threshold St in advance relatively (St can obtain by the reference material size that the user sets), when S＞St, assert in the current frame image has the people, enters step D; Otherwise have no talent in the identification current frame image, enter step e;
D) if the people is arranged in the present image, illustrate that current operator does not leave, return steps A this moment and continue to detect;
E) if having no talent in the present image, illustrate that current operator leaves, can return step 1 this moment, and beginning has detected next time.
The invention has the advantages that: pass through human face detection tech in the past, be merely able to detect normal people's face, can't do the detection of abnormal face, and adopt method of the present invention, only can accomplish mode by Flame Image Process, reach the technology that whether occurs abnormal face in the scene that detects, use, can reach and prevent that lawless person's malice from blocking facial caution effect of committing a crime in self-help bank's ATM in conjunction with embedded digital video recorder (DVR).
Description of drawings
Fig. 1 is the principle of work block diagram of DVR.
Fig. 2 is background testing process figure before the step 1 of the present invention.
Fig. 3 is step 2 people's face testing process figure of the present invention.
Fig. 4 leaves process flow diagram for step 3 detecting operation person of the present invention.
Embodiment
Below just in conjunction with Figure of description, the invention will be further described.
The present invention is intended to reach the detection abnormal face by normal person's face search technique, uses in conjunction with embedded digital video recorder (DVR), reaches the purpose that improves the safety precaution level.
As shown in Figure 1, embedded digital video recorder (DVR) is made of four parts usually, is respectively VO (video acquisition unit), DSP (video processing unit), VS (video storage unit) and VO (video output unit).Wherein DSP partly finishes video compress, functions such as video enhancing and video analysis.The abnormal face test section just belongs to a kind of of video analysis among the present invention.
A kind of abnormal face detecting method that is used for embedded digital video recorder (DVR) system of the present invention, its method step comprises:
In step 1, the detection prospect whether the people is arranged, preceding as shown in Figure 2 background testing process figure:
Take background modeling, extract the method for prospect and judge before the camera whether the people is arranged, owing to only be used for judging that people and nobody are arranged, simple relatively, obtain background with moving average method, concrete steps are as follows:
A), at first, import first two field picture, requiring can not have foreground object in first two field picture, just background object;
B), with first current frame edge image image K0 as a setting;
C), continue input picture, arrive image Kd with what current images K1 and background image K0 did that the frame difference operates;
D), Kd done the corrosion operation of expanding obtain foreground image Kd ', the ultimate value by calculating K d ' zone obtains the area S of the rectangle that Kd ' constitutes;
E), with S and preset threshold St in advance relatively (St can obtain by the reference material size that the user sets), when S＞St, assert in the current frame image has the people; Otherwise have no talent in the identification current frame image, and deposit present image K1 in buffer memory;
F), continue the operation of C-E, 4 two field pictures are arranged in buffer memory, this 4 two field picture and background image are done average calculating operation obtain new background image, the while, K0 emptied;
G), repeating step C-F can constantly upgrade background, and whether the someone occurs to detect present frame, occurs if detect the someone, provides signal, starts people's face and detects link;
Step 2, detect unusual people's face, people's face testing process figure as shown in Figure 3 whether to have occurred in the scene:
Adopt and to detect people's face in a period of time, if having people's face to occur then think normal condition; If continue to monitor face less than the normal person, think that then unusual people's face provides warning, concrete steps are as follows:
A), have the people to occur in the sight that detects through step 1, start people's face detection module this moment;
B), continue whether to have in the detected image people's face to occur, be T detection time, through time of people's normal running before ATM of statistics about 3-5 minute, so T got 1 minute; With frame per second is 25 calculating, can carry out 1500 detections in one minute, blocks if not malice, has enough detected the normal people's face of a frame;
C) if detect people's face in the time at T, then think normal condition, return normal condition this moment, and withdraw from into face and detect, and waits for next time starting; If do not detect people's face, then enter into next step D;
D), return abnormal conditions this moment, and withdraws from people's face and detect, wait for next time starting;
Step 3, judge whether the people of current operation leaves, and the detecting operation person leaves process flow diagram as shown in Figure 4:
No matter step 2 is returned any situation, should wait for that all current operator restarts detection next time after finishing, and concrete steps are as follows:
A), the input present image, arrive image Kd with what current images K1 and background image K0 did that the frame difference operates;
B), Kd done the corrosion operation of expanding obtain foreground image Kd ', the ultimate value by calculating K d ' zone obtains the area S of the rectangle that Kd ' constitutes;
C), with S and preset threshold St in advance relatively (St can obtain by the reference material size that the user sets), when S＞St, assert in the current frame image has the people, enters step D; Otherwise have no talent in the identification current frame image, enter step e;
D) if the people is arranged in the present image, illustrate that current operator does not leave, return steps A this moment and continue to detect;
E) if having no talent in the present image, illustrate that current operator leaves, can return step 1 this moment, and beginning has detected next time.
Show by the experiment that combines with embedded digital video recorder (DVR), the present invention can reach the purpose that detects abnormal face in the scene fully, effectively improve the security of present unmanned ATM, avoided risk for bank and played great meaning, had great practical significance and social effect.
Claims (5)
1. abnormal face detecting method that is used for embedded digital video recorder (DVR) system, in the DSP of embedded digital video recorder (DVR) processing links, increasing people's face detects and the background recognition technology, to be able to discern whether the people is arranged in the present image, this human face's feature whether block by clear nothing, if finding has the abnormal face feature, promptly send alerting signal, remind the related personnel that present case is mentioned hig diligence.
2. abnormal face detecting method that is used for embedded digital video recorder (DVR) system, its method step comprises:
In step 1, the detection prospect whether the people is arranged: take background modeling, extract the method for prospect and judge before the camera whether the people is arranged,, adopt moving average method to obtain background owing to only be used for judging that people and nobody are arranged;
Step 2, detect the unusual people's face that whether occurred in the scene: adopt and detect people's face in a period of time, if having people's face to occur then think normal condition; If continue to monitor face, think that then unusual people's face provides warning less than the normal person;
Step 3, judge whether the people of current operation leaves: no matter step 2 is returned any situation, should wait for all that current operator restarts next time after finishing to detect.
3. a kind of abnormal face detecting method that is used for embedded digital video recorder (DVR) system according to claim 2 is characterized in that: described step 1, and it comprises:
A), at first, import first two field picture, requiring can not have foreground object in first two field picture, just background object;
B), with first current frame edge image image K0 as a setting;
C), continue input picture, arrive image Kd with what current images K1 and background image K0 did that the frame difference operates;
D), Kd done the corrosion operation of expanding obtain foreground image Kd ', the ultimate value by calculating K d ' zone obtains the area S of the rectangle that Kd ' constitutes;
E), with S and preset threshold St in advance relatively (St can obtain by the reference material size that the user sets), when S＞St, assert in the current frame image has the people; Otherwise have no talent in the identification current frame image, and deposit present image K1 in buffer memory;
F), continue the operation of C-E, 4 two field pictures are arranged in buffer memory, this 4 two field picture and background image are done average calculating operation obtain new background image, the while, K0 emptied;
G), repeating step C-F can constantly upgrade background, and whether the someone occurs to detect present frame, occurs if detect the someone, provides signal, starts people's face and detects link.
4. a kind of abnormal face detecting method that is used for embedded digital video recorder (DVR) system according to claim 2 is characterized in that: described step 2, and it comprises:
A), have the people to occur in the sight that detects through step 1, start people's face detection module this moment;
B), continue whether to have in the detected image people's face to occur, be T detection time, through time of people's normal running before ATM of statistics about 3-5 minute, so T got 1 minute; With frame per second is 25 calculating, can carry out 1500 detections in one minute, blocks if not malice, has enough detected the normal people's face of a frame;
C) if detect people's face in the time at T, then think normal condition, return normal condition this moment, and withdraw from people's face and detect, and waits for next time starting; If do not detect people's face, then enter into next step D;
D), return abnormal conditions this moment, and withdraws from people's face and detect, wait for next time starting.
5. a kind of abnormal face detecting method that is used for embedded digital video recorder (DVR) system according to claim 2 is characterized in that: described step 3, and it comprises:
A), the input present image, arrive image Kd with what current images K1 and background image K0 did that the frame difference operates;
B), Kd done the corrosion operation of expanding obtain foreground image Kd ', the ultimate value by calculating K d ' zone obtains the area S of the rectangle that Kd ' constitutes;
C), relatively, St can obtain by the reference material size that the user sets, and when S＞St, assert in the current frame image has the people, enters step D with S and preset threshold St in advance; Otherwise have no talent in the identification current frame image, enter step e;
D) if the people is arranged in the present image, illustrate that current operator does not leave, return steps A this moment and continue to detect;
E) if having no talent in the present image, illustrate that current operator leaves, can return step 1 this moment, and beginning has detected next time.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201010138005 CN101847204A (en) | 2010-03-30 | 2010-03-30 | Abnormal face detecting method for embedded digital video recorder (DVR) system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201010138005 CN101847204A (en) | 2010-03-30 | 2010-03-30 | Abnormal face detecting method for embedded digital video recorder (DVR) system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101847204A true CN101847204A (en) | 2010-09-29 |
Family
ID=42771820
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 201010138005 Pending CN101847204A (en) | 2010-03-30 | 2010-03-30 | Abnormal face detecting method for embedded digital video recorder (DVR) system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101847204A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103440475A (en) * | 2013-08-14 | 2013-12-11 | 北京博思廷科技有限公司 | Automatic teller machine user face visibility judging system and method |
| CN103971100A (en) * | 2014-05-21 | 2014-08-06 | 国家电网公司 | Video-based camouflage and peeping behavior detection method for automated teller machine |
| CN107396029A (en) * | 2017-08-30 | 2017-11-24 | 武汉斗鱼网络科技有限公司 | Method for early warning and device of the duration without face |
| CN108564732A (en) * | 2018-03-22 | 2018-09-21 | 深圳市中科智诚科技有限公司 | A kind of ATM machine that function is blocked with keyboard based on recognition of face |
| CN108734055A (en) * | 2017-04-17 | 2018-11-02 | 杭州海康威视数字技术股份有限公司 | A kind of exception personnel detection method, apparatus and system |
| CN109460729A (en) * | 2018-11-01 | 2019-03-12 | 浩云科技股份有限公司 | A kind of embedded plurality of human faces detection method and system |
| CN109829369A (en) * | 2018-12-25 | 2019-05-31 | 深圳市天彦通信股份有限公司 | Target determines method and relevant apparatus |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1414514A (en) * | 2002-09-18 | 2003-04-30 | 贺贵明 | Video frequency monitoring, identification intelligont device and technical method |
| JP2004062704A (en) * | 2002-07-31 | 2004-02-26 | Hitachi Ltd | Automatic teller machine |
| CN1567344A (en) * | 2003-06-23 | 2005-01-19 | 威视科技股份有限公司 | Face monitoring system and method |
| CN101221663A (en) * | 2008-01-18 | 2008-07-16 | 电子科技大学中山学院 | Intelligent monitoring and alarming method based on movement object detection |
| US20100052904A1 (en) * | 2008-09-04 | 2010-03-04 | Sanyo Electric Co., Ltd. | Abnormality detecting device |
-
2010
- 2010-03-30 CN CN 201010138005 patent/CN101847204A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004062704A (en) * | 2002-07-31 | 2004-02-26 | Hitachi Ltd | Automatic teller machine |
| CN1414514A (en) * | 2002-09-18 | 2003-04-30 | 贺贵明 | Video frequency monitoring, identification intelligont device and technical method |
| CN1567344A (en) * | 2003-06-23 | 2005-01-19 | 威视科技股份有限公司 | Face monitoring system and method |
| CN101221663A (en) * | 2008-01-18 | 2008-07-16 | 电子科技大学中山学院 | Intelligent monitoring and alarming method based on movement object detection |
| US20100052904A1 (en) * | 2008-09-04 | 2010-03-04 | Sanyo Electric Co., Ltd. | Abnormality detecting device |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103440475A (en) * | 2013-08-14 | 2013-12-11 | 北京博思廷科技有限公司 | Automatic teller machine user face visibility judging system and method |
| CN103971100A (en) * | 2014-05-21 | 2014-08-06 | 国家电网公司 | Video-based camouflage and peeping behavior detection method for automated teller machine |
| CN108734055A (en) * | 2017-04-17 | 2018-11-02 | 杭州海康威视数字技术股份有限公司 | A kind of exception personnel detection method, apparatus and system |
| CN108734055B (en) * | 2017-04-17 | 2021-03-26 | 杭州海康威视数字技术股份有限公司 | Abnormal person detection method, device and system |
| CN107396029A (en) * | 2017-08-30 | 2017-11-24 | 武汉斗鱼网络科技有限公司 | Method for early warning and device of the duration without face |
| CN107396029B (en) * | 2017-08-30 | 2020-04-10 | 武汉斗鱼网络科技有限公司 | Duration non-face early warning method and device |
| CN108564732A (en) * | 2018-03-22 | 2018-09-21 | 深圳市中科智诚科技有限公司 | A kind of ATM machine that function is blocked with keyboard based on recognition of face |
| CN109460729A (en) * | 2018-11-01 | 2019-03-12 | 浩云科技股份有限公司 | A kind of embedded plurality of human faces detection method and system |
| CN109829369A (en) * | 2018-12-25 | 2019-05-31 | 深圳市天彦通信股份有限公司 | Target determines method and relevant apparatus |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101847204A (en) | Abnormal face detecting method for embedded digital video recorder (DVR) system | |
| CN103714631B (en) | ATM cash dispenser intelligent monitor system based on recognition of face | |
| US8331674B2 (en) | Rule-based combination of a hierarchy of classifiers for occlusion detection | |
| US20080074496A1 (en) | Video analytics for banking business process monitoring | |
| CN108711208A (en) | A kind of big data access control system and its management method | |
| CN104239851A (en) | Intelligent cell inspection system based on behavior analysis and control method thereof | |
| CN103425997A (en) | Environmental privacy protection processing method and system based on face recognition | |
| CN106951846A (en) | A kind of face 3D models typing and recognition methods and device | |
| CN109598706A (en) | A kind of camera lens occlusion detection method and system | |
| CN110992546A (en) | Face recognition gate and anti-trailing method thereof | |
| CN101303785B (en) | System and method for monitoring safety of ATM | |
| Ray et al. | An intelligent vision system for monitoring security and surveillance of ATM | |
| CN202856897U (en) | Intelligent video monitoring system based on face recognition | |
| CN103400457A (en) | Method for monitoring stealthy ATM (automatic teller machine) bank card substituting behavior | |
| CN107317992A (en) | video communication control device and control method | |
| JP5758165B2 (en) | Article detection device and stationary person detection device | |
| CN106572324A (en) | Energy-efficient smart monitoring device | |
| CN103294981A (en) | Intelligent face recognition system | |
| US10783365B2 (en) | Image processing device and image processing system | |
| Xiya et al. | An abandoned object detection system based on dual background and motion analysis | |
| CN112102543A (en) | Security check system and method | |
| CN116246402A (en) | Monitoring method and device | |
| CN206696918U (en) | Intelligent entrance guard video interlink equipment | |
| CN111277789A (en) | Video-based community security method and device, electronic equipment and storage medium | |
| CN115631562A (en) | Channel gate visitor management system, method and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20100929 |