CN111353426A - Abnormal behavior detection method and device - Google Patents
Abnormal behavior detection method and device Download PDFInfo
- Publication number
- CN111353426A CN111353426A CN202010126848.4A CN202010126848A CN111353426A CN 111353426 A CN111353426 A CN 111353426A CN 202010126848 A CN202010126848 A CN 202010126848A CN 111353426 A CN111353426 A CN 111353426A
- Authority
- CN
- China
- Prior art keywords
- current
- pedestrian
- abnormal
- behavior
- image
- 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
- 206010000117 Abnormal behaviour Diseases 0.000 title claims abstract description 116
- 238000001514 detection method Methods 0.000 title claims description 5
- 238000012544 monitoring process Methods 0.000 claims abstract description 154
- 230000006399 behavior Effects 0.000 claims abstract description 91
- 230000002159 abnormal effect Effects 0.000 claims abstract description 58
- 238000000034 method Methods 0.000 claims abstract description 31
- 230000005856 abnormality Effects 0.000 claims description 6
- 230000000694 effects Effects 0.000 claims description 5
- 230000036544 posture Effects 0.000 description 24
- 230000009191 jumping Effects 0.000 description 6
- 239000002184 metal Substances 0.000 description 6
- 230000008859 change Effects 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000012806 monitoring device Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Human Computer Interaction (AREA)
- Alarm Systems (AREA)
Abstract
The invention provides a method and a device for detecting abnormal behaviors, wherein the method comprises the following steps: predetermining at least one monitoring area, and storing at least one abnormal behavior characteristic corresponding to each monitoring area; aiming at each current monitoring area in at least one monitoring area, acquiring a monitoring video of at least one pedestrian moving in the current monitoring area; for each current pedestrian in the at least one pedestrian, extracting at least one current behavior feature of the current pedestrian from the monitoring video; determining whether the current pedestrian behavior is abnormal or not according to at least one abnormal behavior characteristic and at least one current behavior characteristic corresponding to the current monitoring area; when the behavior of the current pedestrian is abnormal, acquiring the identity information of the current pedestrian; and storing the identity information and the abnormal monitoring video, wherein the abnormal monitoring video comprises the abnormal current behavior characteristics of the current pedestrian. The scheme can reduce the workload of monitoring video workers.
Description
Technical Field
The invention relates to the technical field of computers, in particular to a method and a device for detecting abnormal behaviors.
Background
In order to meet public safety and intelligent scene application, the demand for monitoring systems is increasing day by day, and cameras are widely applied to various scenes, such as traffic roads, squares, stations, shops and the like.
The traditional video monitoring usually adopts a manual monitoring mode. However, when there are many monitoring pictures, the monitoring is performed continuously for 24 hours by human workers, which results in a large workload of the monitoring video worker.
Disclosure of Invention
The embodiment of the invention provides a method and a device for detecting abnormal behaviors, which can reduce the workload of monitoring video workers.
In a first aspect, the present invention provides a method for detecting abnormal behavior, which determines at least one monitoring area in advance, and stores at least one abnormal behavior feature corresponding to each monitoring area, and further includes:
aiming at each current monitoring area in the at least one monitoring area, acquiring a monitoring video of the activity of at least one pedestrian in the current monitoring area;
for each current pedestrian of at least one pedestrian, extracting at least one current behavior feature of the current pedestrian from the monitoring video;
determining whether the current pedestrian behavior is abnormal or not according to at least one abnormal behavior characteristic corresponding to the current monitoring area and the at least one current behavior characteristic;
when the behavior of the current pedestrian is abnormal, acquiring the identity information of the current pedestrian;
and storing the identity information and an abnormal monitoring video, wherein the abnormal monitoring video comprises the current behavior characteristics of the current pedestrian abnormality.
Preferably, the first and second electrodes are formed of a metal,
the extracting at least one current behavior feature of the current pedestrian from the monitoring video comprises:
when the surveillance video is a real-time surveillance video,
extracting at least two frame images from the monitoring video, wherein the frame number between two adjacent frames in the at least two frame images is preset according to the time sequence;
and respectively extracting at least one current behavior characteristic of the current pedestrian from the at least two frames of images.
Preferably, the first and second electrodes are formed of a metal,
the determining whether the behavior of the current pedestrian is abnormal according to the at least one abnormal behavior feature corresponding to the current monitoring area and the at least one current behavior feature includes:
extracting the human body posture of the current pedestrian from the image aiming at each frame of the image;
and determining whether a target abnormal behavior characteristic matched with the human body posture exists in the at least one abnormal behavior characteristic corresponding to the current monitoring area, if so, executing the acquisition of the identity information of the current pedestrian.
Preferably, the first and second electrodes are formed of a metal,
before the obtaining of the identity information of the current pedestrian, further comprising:
storing at least one type of identification information representing the identity of at least one internal person;
the acquiring of the identity information of the current pedestrian includes:
d1: determining an unprocessed set comprising the at least two frame images;
d2: determining a current image from the unprocessed set;
d3: determining whether at least one current identification information representing the identity of the current pedestrian can be extracted from the current image, if so, executing D4, otherwise, executing D6;
d4: determining whether identification information matching the current identification information exists in the stored identification information, if so, performing D5, otherwise, performing D8;
d5: taking the current identification information as the identity information of the current pedestrian;
d6: deleting the current image, and executing D7;
d7: determining whether the current image is the last image in the unprocessed set, if so, executing D8, otherwise, returning to D2;
d8: and marking the marking information representing that the current pedestrian is the non-interior personnel, and taking the marking information as the identity information of the current pedestrian.
In a second aspect, the present invention provides an apparatus for detecting abnormal behavior, comprising:
the system comprises a determining module, a judging module and a judging module, wherein the determining module is used for determining at least one monitoring area in advance and storing at least one abnormal behavior characteristic corresponding to each monitoring area;
the first acquisition module is used for acquiring a monitoring video of the activity of at least one pedestrian in each current monitoring area in the at least one monitoring area;
the processing module is used for extracting at least one current behavior characteristic of at least one current pedestrian from the monitoring video acquired by the first acquisition module aiming at each current pedestrian in the at least one pedestrian; determining whether the current pedestrian behavior is abnormal or not according to at least one abnormal behavior characteristic corresponding to the current monitoring area and the at least one current behavior characteristic;
the second acquisition module is used for acquiring the identity information of the current pedestrian when the processing module determines that the behavior of the current pedestrian is abnormal;
and the storage module is used for storing the identity information and the abnormal monitoring video acquired by the second acquisition module, wherein the abnormal monitoring video comprises the current behavior characteristics of the current pedestrian abnormality.
Preferably, the first and second electrodes are formed of a metal,
the processing module is used for extracting at least two frames of images from the monitoring video when the monitoring video is a real-time monitoring video, wherein the number of frames between two adjacent frames in the at least two frames of images according to a time sequence is preset; and respectively extracting at least one current behavior characteristic of the current pedestrian from the at least two frames of images.
Preferably, the first and second electrodes are formed of a metal,
the processing module is further configured to extract, for each frame of the image, a human body posture of the current pedestrian from the image; and determining whether a target abnormal behavior characteristic matched with the human body posture exists in the at least one abnormal behavior characteristic corresponding to the current monitoring area, and if so, triggering a second acquisition module.
Preferably, the first and second electrodes are formed of a metal,
the storage module is further used for storing at least one identification information representing the identity of at least one internal person before the identity information of the current pedestrian is acquired;
the second obtaining module is configured to perform:
d1: determining an unprocessed set comprising the at least two frame images;
d2: determining a current image from the unprocessed set;
d3: determining whether at least one current identification information representing the identity of the current pedestrian can be extracted from the current image, if so, executing D4, otherwise, executing D6;
d4: determining whether identification information matching the current identification information exists in the stored identification information, if so, performing D5, otherwise, performing D8;
d5: taking the current identification information as the identity information of the current pedestrian;
d6: deleting the current image, and executing D7;
d7: determining whether the current image is the last image in the unprocessed set, if so, executing D8, otherwise, returning to D2;
d8: and marking the marking information representing that the current pedestrian is the non-interior personnel, and taking the marking information as the identity information of the current pedestrian.
In a third aspect, the present invention provides an apparatus for detecting abnormal behavior, comprising: at least one memory and at least one processor;
the at least one memory to store a machine readable program;
the at least one processor is configured to invoke the machine-readable program to perform the method of any of the first aspects.
In a fourth aspect, the present invention provides a computer readable medium having stored thereon computer instructions which, when executed by a processor, cause the processor to perform the method of any of the first aspects.
The embodiment of the invention provides a method and a device for detecting abnormal behaviors, wherein pedestrians may be prohibited from making some abnormal actions in different monitoring areas, so that in order to monitor the behaviors of the pedestrians in the monitoring areas, the current behavior characteristics of each pedestrian can be extracted from a monitoring video of the current monitoring area, then whether the behaviors of the current pedestrians are abnormal or not is judged based on at least one abnormal behavior characteristic corresponding to the current monitoring area, when the behaviors of the current pedestrians are determined to be abnormal, the identity information of the current pedestrians with the abnormal behaviors is obtained, and the abnormal monitoring video including the abnormal behaviors of the current pedestrians and the identity information of the current pedestrians are stored, so that the pedestrians with the abnormal behaviors can be inquired and checked later. By means of the method, abnormal behaviors are detected, and a worker who does not need to monitor videos can monitor the monitoring area continuously for 24 hours, so that the workload of the worker who monitors the videos can be reduced.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
Fig. 1 is a flowchart of a method for detecting abnormal behavior according to an embodiment of the present invention;
FIG. 2 is a flow chart of another abnormal behavior detection method provided by an embodiment of the present invention;
fig. 3 is a schematic diagram of an apparatus for detecting abnormal behavior according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer and more complete, the technical solutions in the embodiments of the present invention will be described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention, and based on the embodiments of the present invention, all other embodiments obtained by a person of ordinary skill in the art without creative efforts belong to the scope of the present invention.
As shown in fig. 1, an embodiment of the present invention provides a method for detecting abnormal behavior, which may include the following steps:
step 101: predetermining at least one monitoring area, and storing at least one abnormal behavior characteristic corresponding to each monitoring area;
step 102: aiming at each current monitoring area in at least one monitoring area, acquiring a monitoring video of at least one pedestrian moving in the current monitoring area;
step 103: for each current pedestrian in the at least one pedestrian, extracting at least one current behavior feature of the current pedestrian from the monitoring video;
step 104: determining whether the current pedestrian behavior is abnormal or not according to at least one abnormal behavior characteristic and at least one current behavior characteristic corresponding to the current monitoring area;
step 105: when the behavior of the current pedestrian is abnormal, acquiring the identity information of the current pedestrian;
step 106: and storing the identity information and the abnormal monitoring video, wherein the abnormal monitoring video comprises the abnormal current behavior characteristics of the current pedestrian.
In the embodiment of the invention, pedestrians may be prohibited from making some abnormal actions in different monitoring areas, therefore, in order to supervise the behaviors of the pedestrians in the monitoring areas, the current behavior feature of each pedestrian may be extracted from the monitoring video of the current monitoring area, then whether the behavior of the current pedestrian is abnormal is judged based on at least one abnormal behavior feature corresponding to the current monitoring area, and when the behavior of the current pedestrian is determined to be abnormal, the identity information of the current pedestrian with the abnormal behavior is acquired, and the abnormal monitoring video including the abnormal behavior of the current pedestrian and the identity information of the current pedestrian are stored, so that the pedestrian with the abnormal behavior is inquired and checked at a later stage. By means of the method, abnormal behaviors are detected, and a worker who does not need to monitor videos can monitor the monitoring area continuously for 24 hours, so that the workload of the worker who monitors the videos can be reduced.
In order to extract the current behavior feature to determine whether there is an abnormality in the behavior of the current pedestrian, in an embodiment of the present invention, at least one current behavior feature of the current pedestrian is extracted from the monitoring video in step 103 in the foregoing embodiment, which may be specifically implemented by:
when the surveillance video is a real-time surveillance video,
extracting at least two frames of images from the monitoring video, wherein the frames of the at least two frames of images which are adjacent to each other in time sequence are preset at intervals;
and respectively extracting at least one current behavior characteristic of the current pedestrian from the at least two frames of images.
In the embodiment of the invention, because the actions of the pedestrians in the monitoring video are continuous, in order to identify whether the current pedestrian behavior is abnormal or not, multiple frames of images can be extracted from the monitoring video, so that the behavior characteristics of the current pedestrian at the moment can be extracted from a static image. Since the human body behaves less in a short time (e.g., 100 ms). For example, in a first frame image and a second frame image extracted from the monitoring video, the feet of the pedestrian in the first frame image directly contact with the ground and the legs of the pedestrian keep an upright state, the left leg of the pedestrian in the second frame image directly contacts with the ground and still keeps the upright state, the angle change between the right leg and the left leg is very small, and the change of the current behavior characteristics of the pedestrian in the first frame image and the second frame image is small at this time. Therefore, in order to prevent repeated extraction of a large number of behavior features with small changes in a short time and increase the data processing amount, a plurality of frames of images are extracted with a preset number of frames spaced between two adjacent frames in time sequence, for example, a surveillance video is composed of 5 frames, the 1 st frame image and the 5 th frame image can be extracted when the images are extracted, and the 3 frame image is spaced between the 2 frame images.
In order to determine whether the current behavior of the pedestrian is abnormal, in an embodiment of the present invention, in step 104 in the foregoing embodiment, whether the current behavior of the pedestrian is abnormal is determined according to at least one abnormal behavior feature and at least one current behavior feature corresponding to the current monitoring area, which may specifically be implemented as follows:
extracting the human body posture of the current pedestrian from the image aiming at each frame of image;
and determining whether at least one abnormal behavior characteristic corresponding to the current monitoring area has a target abnormal behavior characteristic matched with the human body posture, if so, executing to acquire the identity information of the current pedestrian.
In the embodiment of the present invention, a human body posture of a current pedestrian may be extracted for each frame of image as a current behavior feature of the current pedestrian, for example, if the extracted human body posture of the current pedestrian is a jumping human body posture, it may be determined whether there is an abnormal behavior of the current pedestrian by determining whether there is a target abnormal behavior feature matching with the jumping human body posture in at least one abnormal behavior feature corresponding to the current monitoring region. If the current pedestrian has abnormal behaviors, the identity information of the current pedestrian can be acquired so as to search and check the abnormal behaviors at a later period.
In order to obtain the identity information of the current pedestrian, in an embodiment of the present invention, before obtaining the identity information of the current pedestrian, the method further includes:
storing at least one type of identification information representing the identity of at least one internal person;
acquiring identity information of a current pedestrian, comprising:
d1: determining an unprocessed set comprising at least two frames of images;
d2: determining a current image from the unprocessed set;
d3: determining whether at least one current identification information representing the identity of the current pedestrian can be extracted from the current image, if so, executing D4, otherwise, executing D6;
d4: determining whether identification information matching the current identification information exists in the stored identification information, if so, performing D5, otherwise, performing D8;
d5: taking the current identification information as the identity information of the current pedestrian;
d6: delete the current image, perform D7;
d7: determining whether the current image is the last image in the unprocessed set, if so, executing D8, otherwise, returning to D2;
d8: the mark represents the mark information of the current pedestrian as the non-internal person, and the mark information is used as the identity information of the current pedestrian.
In the embodiment of the invention, the identity information of the current pedestrian can be acquired for the detected abnormal behavior so as to search and check the abnormal behavior at a later period, and the identity information of the current pedestrian can be acquired by determining the identifier of the current pedestrian, which can represent the identity information, for example, the face information, the worker plate information and the mark information of a company on clothes of the current pedestrian can be determined so as to acquire the identity information of the current pedestrian. The method comprises the steps of taking an image of a current pedestrian including at least two frames of images as an unprocessed set, and determining whether at least one current identification information representing the identity of the current pedestrian can be extracted from a first current image in the unprocessed set. If the identification information can be extracted, the identification information is identified to determine the identity information of the pedestrian, and the current identification information is used as the identity information of the current pedestrian. And if the first current image can not extract the identification information, deleting the identification information, and identifying the identification information of the second current image until the identification information representing the identity of the current pedestrian can not be extracted from the last current image of the unprocessed set, and marking the identity information of the current pedestrian as non-insider personnel.
The invention realizes a method for detecting the abnormal behaviors of the pedestrians based on the video image of the monitoring area, and the method can also automatically extract the current behavior characteristics of the current pedestrians from the video image through a CNN model, and determine the abnormal behaviors of the current pedestrians according to at least one kind of abnormal behavior characteristics stored in advance and the behavior characteristics of the current pedestrians. When the current behavior characteristic is confirmed to be abnormal behavior, the identity recognition module can be called to acquire the identity information of the current pedestrian, and the identity information of the current pedestrian and the abnormal behavior monitoring video are stored in the database, so that violation information of personnel can be inquired conveniently, and the operation of intelligently recognizing the abnormal behavior of the pedestrian is realized.
The monitoring system is widely applied, but the actual monitoring task still needs more manpower to be completed, and the existing video monitoring system usually only records video images and can only be used for obtaining evidence afterwards, and the real-time performance and the initiative of monitoring are not fully exerted. The invention can analyze, track and judge the monitored object in real time, and prompt when an abnormal event occurs, thereby facilitating the unit to master the violation condition of the internal personnel.
The invention has the following outstanding advantages in the aspect of detecting abnormal behaviors: (1) the cost is low, new equipment does not need to be purchased, and the method can be realized only by using the existing office computer; (2) the abnormal behaviors and the identity information of the pedestrians are automatically identified, the abnormal behavior records are stored, and the manual workload is reduced; (3) the method has universality and can be applied to all monitoring areas needing to detect abnormal behaviors.
As shown in fig. 2, in order to more clearly illustrate the technical solution and advantages of the present invention, the following describes in detail a method for detecting abnormal behavior provided by an embodiment of the present invention, and specifically includes the following steps:
step 201: at least one monitoring area is predetermined, and at least one abnormal behavior characteristic corresponding to each monitoring area is stored.
Specifically, at least one abnormal behavior characteristic corresponding to each monitoring area is stored as a reference object so as to determine whether the current pedestrian behavior is abnormal. A personnel behavior characteristic class library exists in the system, and abnormal behavior characteristics can be customized. In a personnel behavior characteristic class library of the system, the abnormal behavior characteristics of the personnel can be customized and the default abnormal behavior characteristics can be modified.
For example, assuming that the monitored area a prohibits the running of the pedestrian and the monitored area B prohibits the jumping of the pedestrian, the abnormal behavior feature of the monitored area a may be stored as a plurality of running human postures as the abnormal behavior feature, and the abnormal behavior feature of the monitored area B may be stored as a plurality of jumping human postures as the abnormal behavior feature.
Step 202: at least one kind of identification information of at least one kind of internal person is stored.
Specifically, for the detected abnormal behavior of the pedestrian, the identity information of the pedestrian needs to be identified, so as to search and check the identity information of the pedestrian corresponding to the abnormal behavior at a later period, and therefore, at least one identification information representing the identity of at least one inside person needs to be stored, so as to determine whether the pedestrian with the abnormal behavior is the inside person.
For example, face recognition information, card recognition information, and company logo information of the insiders may be stored as identification information representing the identities of the insiders.
Step 203: and acquiring a monitoring video of the activity of at least one pedestrian in the current monitoring area aiming at each current monitoring area in the at least one monitoring area.
Step 204: for each current pedestrian in at least one pedestrian, when the monitoring video is a real-time monitoring video, at least two frames of images are extracted from the monitoring video, wherein the number of frames between two adjacent frames in the at least two frames of images is preset according to the time sequence.
Step 205: and respectively extracting at least one current behavior characteristic of the current pedestrian from the at least two frames of images.
Step 206: and extracting the human body posture of the current pedestrian from the image according to each frame of image.
Step 207: and determining whether a target abnormal behavior characteristic matched with the human body posture exists in at least one abnormal behavior characteristic corresponding to the current monitoring area, if so, executing the step 208.
For example, software installed with a CNN model and capable of intelligently identifying abnormal behavior characteristics of a pedestrian may be connected to a monitoring device, at least two frames of images are extracted from a monitoring video in real time through the CNN model, a human posture of a current pedestrian is extracted from each frame of image, and whether the current pedestrian has abnormal behavior or not is determined according to the abnormal behavior characteristics in the pedestrian behavior characteristic class library and current behavior characteristics corresponding to the human posture of the current pedestrian. For example, the first frame image and the fifth frame image of the current pedestrian a in the monitoring area a can be extracted through the CNN model, the body posture of the running pedestrian a in the first frame image and the fifth frame image is recognized through software for intelligently recognizing abnormal behavior characteristics of people, and whether the current behavior characteristics are abnormal is determined by determining whether the current behavior characteristics corresponding to the body posture of the running pedestrian a are matched with the prestored abnormal behavior characteristics.
Step 208: an unprocessed set comprising at least two frames of images is determined.
Step 209: a current image is determined from the unprocessed set.
Step 210: it is determined whether at least one current identification information characterizing the identity of the current pedestrian can be extracted from the current image, if so, step 211 is performed, otherwise, step 213 is performed.
Step 211: it is determined whether there is identification information matching the current identification information among the stored identification information, and if so, step 212 is performed, otherwise, step 213 is performed.
Step 212: and taking the current identification information as the identity information of the current pedestrian.
Step 213: the current picture is deleted and step 214 is performed.
Step 214: it is determined whether the current image is the last image in the unprocessed set, and if so, step 215 is performed, otherwise, step 209 is returned.
Step 215: the mark represents the mark information of the current pedestrian as the non-internal person, and the mark information is used as the identity information of the current pedestrian.
Specifically, identification information can be identified for the current pedestrian with abnormal behavior to acquire identity information of the current pedestrian, so that violation information of the pedestrian can be inquired later.
For example, assuming that a first frame, a third frame, and a fifth frame of image corresponding to a current pedestrian are extracted, identification information (for example, face information, worker plate information, and a company logo) in the first frame of image is first identified, if the stored identification information includes identification information matching the current identification information, the current identification information is used as identity information of the current pedestrian, if the first frame cannot identify the identification information, the first frame of image is deleted, the third frame of image is still not identified, the third frame of image is deleted, the fifth frame of image is identified again, the fifth frame of image still cannot identify the identification information, and the fifth frame is a last image, and the current pedestrian is marked as a non-insider. Step 216: and storing the identity information and the abnormal monitoring video, wherein the abnormal monitoring video comprises the abnormal current behavior characteristics of the current pedestrian.
Specifically, the identity information and the abnormal behavior of the pedestrian with the abnormal behavior are recorded into the database, and the system end can inquire and check the abnormal behavior information of the pedestrian through an inquiry interface.
As shown in fig. 3, an embodiment of the present invention provides an apparatus for detecting an abnormal behavior, including:
a determining module 301, configured to determine at least one monitoring area in advance, and store at least one abnormal behavior feature corresponding to each monitoring area;
a first obtaining module 302, configured to obtain, for each current monitoring area in the at least one monitoring area, a monitoring video of at least one pedestrian moving in the current monitoring area;
a processing module 303, configured to, for each current pedestrian of at least one pedestrian, extract at least one current behavior feature of the current pedestrian from the monitoring video acquired by the first acquiring module 302; determining whether the current pedestrian behavior is abnormal or not according to at least one abnormal behavior characteristic corresponding to the current monitoring area and the at least one current behavior characteristic;
a second obtaining module 304, configured to obtain identity information of the current pedestrian when the processing module 303 determines that the behavior of the current pedestrian is abnormal;
a storage module 305, configured to store the identity information and the abnormal monitoring video acquired by the second acquiring module 304, where the abnormal monitoring video includes the current behavior characteristic of the current pedestrian abnormality.
In the embodiment of the invention, pedestrians may be prohibited from making some abnormal actions in different monitoring areas, therefore, in order to supervise the behaviors of the pedestrians in the monitoring areas, the processing module may extract the current behavior feature of each pedestrian from the monitoring video of the current monitoring area acquired by the first acquiring module, then determine whether the behavior of the current pedestrian is abnormal based on at least one abnormal behavior feature corresponding to the current monitoring area determined by the determining module, and when it is determined that the behavior of the current pedestrian is abnormal, acquire the identity information of the current pedestrian with abnormal behavior by the second acquiring module, and store the abnormal monitoring video including the abnormal behavior of the current pedestrian and the identity information of the current pedestrian by the storage module, so as to query and check the pedestrian with abnormal behavior at a later stage. By means of the method, abnormal behaviors are detected, and a worker who does not need to monitor videos can monitor the monitoring area continuously for 24 hours, so that the workload of the worker who monitors the videos can be reduced.
In an embodiment of the present invention, the processing module 303 is configured to extract at least two frames of images from the monitoring video when the monitoring video is a real-time monitoring video, where a preset number of frames are spaced between two adjacent frames in a time sequence in the at least two frames of images; and respectively extracting at least one current behavior characteristic of the current pedestrian from the at least two frames of images.
In an embodiment of the present invention, the processing module 303 is configured to extract, for each frame of the image, a human body posture of the current pedestrian from the image; and determining whether a target abnormal behavior characteristic matched with the human body posture exists in the at least one abnormal behavior characteristic corresponding to the current monitoring area, and if so, triggering a second acquisition module.
In an embodiment of the present invention, the storage module 305 is further configured to store at least one identification information representing an identity of at least one inside person before the obtaining of the identity information of the current pedestrian;
the second obtaining module 304 is configured to perform:
d1: determining an unprocessed set comprising the at least two frame images;
d2: determining a current image from the unprocessed set;
d3: determining whether at least one current identification information representing the identity of the current pedestrian can be extracted from the current image, if so, executing D4, otherwise, executing D6;
d4: determining whether identification information matching the current identification information exists in the stored identification information, if so, performing D5, otherwise, performing D8;
d5: taking the current identification information as the identity information of the current pedestrian;
d6: deleting the current image, and executing D7;
d7: determining whether the current image is the last image in the unprocessed set, if so, executing D8, otherwise, returning to D2;
d8: and marking the marking information representing that the current pedestrian is the non-interior personnel, and taking the marking information as the identity information of the current pedestrian.
It is to be understood that the illustrated structure of the embodiment of the present invention does not constitute a specific limitation to the detection device of abnormal behavior. In other embodiments of the invention, the means for detecting abnormal behavior may include more or fewer components than shown, or some components may be combined, some components may be separated, or a different arrangement of components. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.
Because the information interaction, execution process, and other contents between the units in the device are based on the same concept as the method embodiment of the present invention, specific contents may refer to the description in the method embodiment of the present invention, and are not described herein again.
The embodiment of the present invention further provides a device for detecting an abnormal behavior, including: at least one memory and at least one processor;
the at least one memory to store a machine readable program;
the at least one processor is configured to invoke the machine-readable program to perform a method for detecting abnormal behavior according to any embodiment of the present invention.
An embodiment of the present invention further provides a computer-readable medium, where a computer instruction is stored on the computer-readable medium, and when the computer instruction is executed by a processor, the processor is caused to execute the method for detecting an abnormal behavior in any embodiment of the present invention. Specifically, a system or an apparatus equipped with a storage medium on which software program codes that realize the functions of any of the above-described embodiments are stored may be provided, and a computer (or a CPU or MPU) of the system or the apparatus is caused to read out and execute the program codes stored in the storage medium.
In this case, the program code itself read from the storage medium can realize the functions of any of the above-described embodiments, and thus the program code and the storage medium storing the program code constitute a part of the present invention.
Examples of the storage medium for supplying the program code include a floppy disk, a hard disk, a magneto-optical disk, an optical disk (e.g., CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD + RW), a magnetic tape, a nonvolatile memory card, and a ROM. Alternatively, the program code may be downloaded from a server computer via a communications network.
Further, it should be clear that the functions of any one of the above-described embodiments may be implemented not only by executing the program code read out by the computer, but also by causing an operating system or the like operating on the computer to perform a part or all of the actual operations based on instructions of the program code.
Further, it is to be understood that the program code read out from the storage medium is written to a memory provided in an expansion board inserted into the computer or to a memory provided in an expansion unit connected to the computer, and then causes a CPU or the like mounted on the expansion board or the expansion unit to perform part or all of the actual operations based on instructions of the program code, thereby realizing the functions of any of the above-described embodiments.
The embodiments of the invention have at least the following beneficial effects:
1. in the embodiment of the invention, pedestrians may be prohibited from making some abnormal actions in different monitoring areas, therefore, in order to supervise the behaviors of the pedestrians in the monitoring areas, the current behavior feature of each pedestrian may be extracted from the monitoring video of the current monitoring area, then whether the behavior of the current pedestrian is abnormal is judged based on at least one abnormal behavior feature corresponding to the current monitoring area, and when the behavior of the current pedestrian is determined to be abnormal, the identity information of the current pedestrian with the abnormal behavior is acquired, and the abnormal monitoring video including the abnormal behavior of the current pedestrian and the identity information of the current pedestrian are stored, so that the pedestrian with the abnormal behavior is inquired and checked at a later stage. By means of the method, abnormal behaviors are detected, and a worker who does not need to monitor videos can monitor the monitoring area continuously for 24 hours, so that the workload of the worker who monitors the videos can be reduced.
2. In an embodiment of the present invention, since the actions of the pedestrian in the surveillance video are continuous, in order to facilitate identifying whether the current behavior of the pedestrian is abnormal, multiple frames of images may be extracted from the surveillance video, so as to extract the behavior feature of the current pedestrian at the moment from the static image. Since the human body behaves less in a short time (e.g., 100 ms). For example, in a first frame image and a second frame image extracted from the monitoring video, the feet of the pedestrian in the first frame image directly contact with the ground and the legs of the pedestrian keep an upright state, the left leg of the pedestrian in the second frame image directly contacts with the ground and still keeps the upright state, the angle change between the right leg and the left leg is very small, and the change of the current behavior characteristics of the pedestrian in the first frame image and the second frame image is small at this time. Therefore, in order to prevent repeated extraction of a large number of behavior features with small changes in a short time and increase the data processing amount, a plurality of frames of images are extracted with a preset number of frames spaced between two adjacent frames in time sequence, for example, a surveillance video is composed of 5 frames, the 1 st frame image and the 5 th frame image can be extracted when the images are extracted, and the 3 frame image is spaced between the 2 frame images.
3. In an embodiment of the present invention, a body posture of a current pedestrian may be extracted for each frame of image as a current behavior feature of the current pedestrian, for example, if the extracted body posture of the current pedestrian is a jumping body posture, it may be determined whether there is a target abnormal behavior feature matching the jumping body posture in at least one abnormal behavior feature corresponding to the current monitoring area, so as to determine whether there is an abnormal behavior of the current pedestrian. If the current pedestrian has abnormal behaviors, the identity information of the current pedestrian can be acquired so as to search and check the abnormal behaviors at a later period.
It should be noted that not all steps and modules in the above flows and system structure diagrams are necessary, and some steps or modules may be omitted according to actual needs. The execution order of the steps is not fixed and can be adjusted as required. The system structure described in the above embodiments may be a physical structure or a logical structure, that is, some modules may be implemented by the same physical entity, or some modules may be implemented by a plurality of physical entities, or some components in a plurality of independent devices may be implemented together.
In the above embodiments, the hardware unit may be implemented mechanically or electrically. For example, a hardware element may comprise permanently dedicated circuitry or logic (such as a dedicated processor, FPGA or ASIC) to perform the corresponding operations. The hardware elements may also comprise programmable logic or circuitry, such as a general purpose processor or other programmable processor, that may be temporarily configured by software to perform the corresponding operations. The specific implementation (mechanical, or dedicated permanent, or temporarily set) may be determined based on cost and time considerations.
While the invention has been shown and described in detail in the drawings and in the preferred embodiments, it is not intended to limit the invention to the embodiments disclosed, and it will be apparent to those skilled in the art that various combinations of the code auditing means in the various embodiments described above may be used to obtain further embodiments of the invention, which are also within the scope of the invention.
Claims (10)
1. The abnormal behavior detection method is characterized by further comprising the following steps of:
aiming at each current monitoring area in the at least one monitoring area, acquiring a monitoring video of the activity of at least one pedestrian in the current monitoring area;
for each current pedestrian of at least one pedestrian, extracting at least one current behavior feature of the current pedestrian from the monitoring video;
determining whether the current pedestrian behavior is abnormal or not according to at least one abnormal behavior characteristic corresponding to the current monitoring area and the at least one current behavior characteristic;
when the behavior of the current pedestrian is abnormal, acquiring the identity information of the current pedestrian;
and storing the identity information and an abnormal monitoring video, wherein the abnormal monitoring video comprises the current behavior characteristics of the current pedestrian abnormality.
2. The method of claim 1,
the extracting at least one current behavior feature of the current pedestrian from the monitoring video comprises:
when the surveillance video is a real-time surveillance video,
extracting at least two frame images from the monitoring video, wherein the frame number between two adjacent frames in the at least two frame images is preset according to the time sequence;
and respectively extracting at least one current behavior characteristic of the current pedestrian from the at least two frames of images.
3. The method of claim 2,
the determining whether the behavior of the current pedestrian is abnormal according to the at least one abnormal behavior feature corresponding to the current monitoring area and the at least one current behavior feature includes:
extracting the human body posture of the current pedestrian from the image aiming at each frame of the image;
and determining whether a target abnormal behavior characteristic matched with the human body posture exists in the at least one abnormal behavior characteristic corresponding to the current monitoring area, if so, executing the acquisition of the identity information of the current pedestrian.
4. The method according to claim 2 or 3,
before the obtaining of the identity information of the current pedestrian, further comprising:
storing at least one type of identification information representing the identity of at least one internal person;
the acquiring of the identity information of the current pedestrian includes:
d1: determining an unprocessed set comprising the at least two frame images;
d2: determining a current image from the unprocessed set;
d3: determining whether at least one current identification information representing the identity of the current pedestrian can be extracted from the current image, if so, executing D4, otherwise, executing D6;
d4: determining whether identification information matching the current identification information exists in the stored identification information, if so, performing D5, otherwise, performing D8;
d5: taking the current identification information as the identity information of the current pedestrian;
d6: deleting the current image, and executing D7;
d7: determining whether the current image is the last image in the unprocessed set, if so, executing D8, otherwise, returning to D2;
d8: and marking the marking information representing that the current pedestrian is the non-interior personnel, and taking the marking information as the identity information of the current pedestrian.
5. An apparatus for detecting abnormal behavior, comprising:
the system comprises a determining module, a judging module and a judging module, wherein the determining module is used for determining at least one monitoring area in advance and storing at least one abnormal behavior characteristic corresponding to each monitoring area;
the first acquisition module is used for acquiring a monitoring video of the activity of at least one pedestrian in each current monitoring area in the at least one monitoring area;
the processing module is used for extracting at least one current behavior characteristic of at least one current pedestrian from the monitoring video acquired by the first acquisition module aiming at each current pedestrian in the at least one pedestrian; determining whether the current pedestrian behavior is abnormal or not according to at least one abnormal behavior characteristic corresponding to the current monitoring area and the at least one current behavior characteristic;
the second acquisition module is used for acquiring the identity information of the current pedestrian when the processing module determines that the behavior of the current pedestrian is abnormal;
and the storage module is used for storing the identity information and the abnormal monitoring video acquired by the second acquisition module, wherein the abnormal monitoring video comprises the current behavior characteristics of the current pedestrian abnormality.
6. The apparatus of claim 5,
the processing module is used for extracting at least two frames of images from the monitoring video when the monitoring video is a real-time monitoring video, wherein the number of frames between two adjacent frames in the at least two frames of images according to a time sequence is preset; and respectively extracting at least one current behavior characteristic of the current pedestrian from the at least two frames of images.
7. The apparatus of claim 6,
the processing module is further configured to extract, for each frame of the image, a human body posture of the current pedestrian from the image; and determining whether a target abnormal behavior characteristic matched with the human body posture exists in the at least one abnormal behavior characteristic corresponding to the current monitoring area, and if so, triggering a second acquisition module.
8. The apparatus according to claim 6 or 7,
the storage module is further used for storing at least one identification information representing the identity of at least one internal person before the identity information of the current pedestrian is acquired;
the second obtaining module is configured to perform:
d1: determining an unprocessed set comprising the at least two frame images;
d2: determining a current image from the unprocessed set;
d3: determining whether at least one current identification information representing the identity of the current pedestrian can be extracted from the current image, if so, executing D4, otherwise, executing D6;
d4: determining whether identification information matching the current identification information exists in the stored identification information, if so, performing D5, otherwise, performing D8;
d5: taking the current identification information as the identity information of the current pedestrian;
d6: deleting the current image, and executing D7;
d7: determining whether the current image is the last image in the unprocessed set, if so, executing D8, otherwise, returning to D2;
d8: and marking the marking information representing that the current pedestrian is the non-interior personnel, and taking the marking information as the identity information of the current pedestrian.
9. An apparatus for detecting abnormal behavior, comprising: at least one memory and at least one processor;
the at least one memory to store a machine readable program;
the at least one processor, configured to invoke the machine readable program to perform the method of any of claims 1 to 4.
10. Computer readable medium, characterized in that it has stored thereon computer instructions which, when executed by a processor, cause the processor to carry out the method of any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010126848.4A CN111353426A (en) | 2020-02-28 | 2020-02-28 | Abnormal behavior detection method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010126848.4A CN111353426A (en) | 2020-02-28 | 2020-02-28 | Abnormal behavior detection method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111353426A true CN111353426A (en) | 2020-06-30 |
Family
ID=71194142
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010126848.4A Pending CN111353426A (en) | 2020-02-28 | 2020-02-28 | Abnormal behavior detection method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111353426A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111814775A (en) * | 2020-09-10 | 2020-10-23 | 平安国际智慧城市科技股份有限公司 | Target object abnormal behavior identification method, device, terminal and storage medium |
| CN112541661A (en) * | 2020-12-01 | 2021-03-23 | 中国工商银行股份有限公司 | Method and device for detecting personnel behavior and environmental specification of network points |
| CN112633133A (en) * | 2020-12-18 | 2021-04-09 | 江苏省苏力环境科技有限责任公司 | AI-based intelligent water station operation and maintenance method, system, terminal and storage medium |
| CN113469080A (en) * | 2021-07-08 | 2021-10-01 | 中国科学院自动化研究所 | Individual, group and scene interactive collaborative perception method, system and equipment |
| CN113989934A (en) * | 2021-10-29 | 2022-01-28 | 北京百度网讯科技有限公司 | Behavior detection method, behavior detection device, electronic apparatus, and medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106384237A (en) * | 2016-08-31 | 2017-02-08 | 北京志光伯元科技有限公司 | Member authentication-management method, device and system based on face identification |
| CN108399665A (en) * | 2018-01-03 | 2018-08-14 | 平安科技（深圳）有限公司 | Method for safety monitoring, device based on recognition of face and storage medium |
| CN108664608A (en) * | 2018-05-11 | 2018-10-16 | 中国联合网络通信集团有限公司 | Recognition methods, device and the computer readable storage medium of a suspect |
| CN108765935A (en) * | 2018-06-12 | 2018-11-06 | 佛山市灏金赢科技有限公司 | A kind of pedestrian running red light is broken rules and regulations processing system and its method |
| CN109460702A (en) * | 2018-09-14 | 2019-03-12 | 华南理工大学 | Passenger's abnormal behaviour recognition methods based on human skeleton sequence |
| CN110176117A (en) * | 2019-06-17 | 2019-08-27 | 广东翔翼科技信息有限公司 | A kind of monitoring device and monitoring method of Behavior-based control identification technology |
| CN110609920A (en) * | 2019-08-05 | 2019-12-24 | 华中科技大学 | Method and system for mixed pedestrian search in video surveillance scene |
-
2020
- 2020-02-28 CN CN202010126848.4A patent/CN111353426A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106384237A (en) * | 2016-08-31 | 2017-02-08 | 北京志光伯元科技有限公司 | Member authentication-management method, device and system based on face identification |
| CN108399665A (en) * | 2018-01-03 | 2018-08-14 | 平安科技（深圳）有限公司 | Method for safety monitoring, device based on recognition of face and storage medium |
| CN108664608A (en) * | 2018-05-11 | 2018-10-16 | 中国联合网络通信集团有限公司 | Recognition methods, device and the computer readable storage medium of a suspect |
| CN108765935A (en) * | 2018-06-12 | 2018-11-06 | 佛山市灏金赢科技有限公司 | A kind of pedestrian running red light is broken rules and regulations processing system and its method |
| CN109460702A (en) * | 2018-09-14 | 2019-03-12 | 华南理工大学 | Passenger's abnormal behaviour recognition methods based on human skeleton sequence |
| CN110176117A (en) * | 2019-06-17 | 2019-08-27 | 广东翔翼科技信息有限公司 | A kind of monitoring device and monitoring method of Behavior-based control identification technology |
| CN110609920A (en) * | 2019-08-05 | 2019-12-24 | 华中科技大学 | Method and system for mixed pedestrian search in video surveillance scene |
Non-Patent Citations (1)
| Title |
|---|
| 马义飞: "基于异常行为识别在公共区域的安全防范建模", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111814775A (en) * | 2020-09-10 | 2020-10-23 | 平安国际智慧城市科技股份有限公司 | Target object abnormal behavior identification method, device, terminal and storage medium |
| CN111814775B (en) * | 2020-09-10 | 2020-12-11 | 平安国际智慧城市科技股份有限公司 | Target object abnormal behavior identification method, device, terminal and storage medium |
| CN112541661A (en) * | 2020-12-01 | 2021-03-23 | 中国工商银行股份有限公司 | Method and device for detecting personnel behavior and environmental specification of network points |
| CN112633133A (en) * | 2020-12-18 | 2021-04-09 | 江苏省苏力环境科技有限责任公司 | AI-based intelligent water station operation and maintenance method, system, terminal and storage medium |
| CN113469080A (en) * | 2021-07-08 | 2021-10-01 | 中国科学院自动化研究所 | Individual, group and scene interactive collaborative perception method, system and equipment |
| CN113989934A (en) * | 2021-10-29 | 2022-01-28 | 北京百度网讯科技有限公司 | Behavior detection method, behavior detection device, electronic apparatus, and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111353426A (en) | Abnormal behavior detection method and device | |
| WO2019134346A1 (en) | Face recognition method, application server, and computer-readable storage medium | |
| CN110139075B (en) | Video data processing method, video data processing device, computer equipment and storage medium | |
| CN110738178A (en) | Garden construction safety detection method and device, computer equipment and storage medium | |
| CN110705477A (en) | Behavior analysis method and apparatus, electronic device, and computer storage medium | |
| JP2005501351A (en) | Vision-based method and apparatus for detecting fraud events in a retail environment | |
| CN113269127B (en) | Face recognition and pedestrian re-recognition monitoring method and system for real-time automatic database establishment | |
| CN110781733B (en) | Image duplicate removal method, storage medium, network equipment and intelligent monitoring system | |
| CN113095132B (en) | Neural network based gas field identification method, system, terminal and storage medium | |
| US10271018B2 (en) | Method of detecting critical objects from CCTV video using metadata filtering | |
| CN108446681B (en) | Pedestrian analysis method, device, terminal and storage medium | |
| CN107564225A (en) | An image detection and alarm system | |
| CN110705476A (en) | Data analysis method and device, electronic equipment and computer storage medium | |
| CN110852148A (en) | Visitor destination verification method and system based on target tracking | |
| CN111476685B (en) | Behavior analysis method, device and equipment | |
| CN111368619A (en) | Method, device and equipment for detecting suspicious people | |
| CN111260869B (en) | Method and device for extracting video frames in monitoring video and computer equipment | |
| CN110705472A (en) | Off-duty method and system based on video image recognition | |
| CN111045889A (en) | Closed network equipment state monitoring system, method and device and readable storage medium | |
| CN110717358A (en) | Visitor number counting method and device, electronic equipment and storage medium | |
| CN113343913A (en) | Target determination method, target determination device, storage medium and computer equipment | |
| CN112418063A (en) | Face recognition method and device, electronic equipment and storage medium | |
| CN115410112A (en) | Elevator abnormal stay recognition method and device and storage medium | |
| CN112001280B (en) | Real-time and online optimized face recognition system and method | |
| CN112733737A (en) | Human body weight recognition method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200630 |
|
| RJ01 | Rejection of invention patent application after publication |