CN112419638B - Method and device for acquiring alarm video - Google Patents
Method and device for acquiring alarm video Download PDFInfo
- Publication number
- CN112419638B CN112419638B CN202011088030.4A CN202011088030A CN112419638B CN 112419638 B CN112419638 B CN 112419638B CN 202011088030 A CN202011088030 A CN 202011088030A CN 112419638 B CN112419638 B CN 112419638B
- Authority
- CN
- China
- Prior art keywords
- alarm
- video
- video stream
- analyzed
- frame
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 47
- 230000006399 behavior Effects 0.000 claims description 46
- 238000013528 artificial neural network Methods 0.000 claims description 34
- 238000004590 computer program Methods 0.000 claims description 6
- 239000000284 extract Substances 0.000 claims description 6
- 238000012544 monitoring process Methods 0.000 description 13
- 238000000605 extraction Methods 0.000 description 5
- 238000004891 communication Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000010276 construction Methods 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 241001465754 Metazoa Species 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000005856 abnormality Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B13/00—Burglar, theft or intruder alarms
- G08B13/18—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength
- G08B13/189—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength using passive radiation detection systems
- G08B13/194—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength using passive radiation detection systems using image scanning and comparing systems
- G08B13/196—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength using passive radiation detection systems using image scanning and comparing systems using television cameras
- G08B13/19602—Image analysis to detect motion of the intruder, e.g. by frame subtraction
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Alarm Systems (AREA)
Abstract
The disclosure relates to a method and a device for acquiring an alarm video. Firstly, performing video analysis on an original video stream, generating alarm starting information when a preset event occurs, and sending the alarm starting information to an alarm terminal so that the alarm terminal records the alarm starting time; when the preset event is determined to be finished, generating alarm release information, and sending the alarm release information to an alarm terminal so that the alarm terminal records the alarm finishing time; receiving an alarm video to be analyzed sent by an alarm terminal; the alarm video to be analyzed is obtained by intercepting an original video stream of the acquisition terminal by the alarm terminal according to the alarm starting time and the alarm ending time; and identifying the alarm video to be analyzed according to the local processing resource to form the marked alarm video. The method and the device do not need to send the video stream to the alarm terminal in real time to cause overlarge load of network resources and processing resources, can fully utilize local processing resources of the video analysis module, provide the alarm video for the alarm terminal and ensure the complete recording of the alarm video.
Description
Technical Field
The disclosure relates to a method for acquiring an alarm video, and also relates to a corresponding device for acquiring the alarm video, belonging to the technical field of video analysis.
Background
In some security, monitoring and other scenes, video analysis is generally required to be performed on video images acquired by an acquisition terminal, so that related scenes are effectively monitored, and an alarm record when an abnormality is monitored is generated.
In order to enable the alarm terminal in the background to know the monitoring condition in time, when the monitoring is abnormal, the alarm record can be pushed to the alarm terminal for displaying. Generally, due to limited network resources, the same alarm terminal may need to receive a large number of alarm records of the collection terminal, so the alarm records generally only include one alarm image, and if the dangerous area does not allow the people to appear, when a person appears, a frame of alarm image with the person is collected. Thus, only one alarm image is presented at the alarm terminal, and the information presented at the alarm terminal is limited. In addition, if a complete recorded and analyzed warning video is provided for the warning terminal, each frame of the real-time video needs to be analyzed and processed, the frame extraction frequency needs to be increased when the warning condition is identified, and the consumption of image processing resources is increased when each frame of the real-time video is analyzed. Therefore, if a real-time warning video is provided for the warning terminal at present, a heavy burden is caused to image processing resources and network resources, and particularly when a plurality of acquisition terminals give a warning at the same time, the instantaneous burden of the analysis service is too heavy, so that the request for the real-time video fails, and then the recording also fails. In severe cases, the service being analyzed may be affected, and the user experience is poor.
Disclosure of Invention
The first technical problem to be solved by the present disclosure is to provide a method for acquiring an alarm video.
Another technical problem to be solved by the present disclosure is to provide an apparatus for acquiring an alert video.
In order to achieve the purpose, the following technical scheme is adopted in the disclosure:
according to a first aspect of the embodiments of the present disclosure, a method for acquiring an alert video is provided, which includes the following steps:
performing video analysis according to an original video stream acquired by an acquisition terminal, and generating alarm starting information when a preset event occurs in the original video stream;
sending the alarm starting information to an alarm terminal so that the alarm terminal records the alarm starting time;
performing video analysis according to an original video stream acquired by an acquisition terminal, and generating alarm release information when a preset event in the original video stream is determined to be finished;
sending the alarm release information to an alarm terminal so that the alarm terminal records the alarm ending time;
receiving an alarm video to be analyzed sent by an alarm terminal; the alarm video to be analyzed is obtained by the alarm terminal by intercepting the original video stream of the acquisition terminal according to the alarm starting time and the alarm ending time;
and identifying the alarm video to be analyzed according to local processing resources to form a marked alarm video.
Preferably, the video analysis is performed according to the original video stream acquired by the acquisition terminal, and includes:
extracting the original video stream frame by frame according to the original video stream acquired by the acquisition terminal to obtain each frame of image;
and inputting each frame of image into a pre-trained neural network to perform preset object type recognition or preset object behavior recognition.
Preferably, when it is determined that a preset event occurs in the original video stream, generating the alarm start information includes:
within a preset frame range, if a preset object type or a preset object behavior is recognized for the first time according to a pre-trained neural network, determining that a preset event occurs in an original video stream;
generating alarm starting information; the alarm starting information comprises a target frame image carrying an alarm starting mark; the target frame image is a frame image corresponding to the type or behavior of a preset object which is recognized for the first time according to a pre-trained neural network.
Preferably, after the alarm start information is generated, the method further comprises the following steps:
and sending the alarm starting information to an alarm terminal so that the alarm terminal displays a target frame image carrying an alarm starting mark.
Preferably, when it is determined that the preset event in the original video stream is ended, generating the alarm release information includes:
in a preset frame range, if the preset object type or the preset object behavior is identified last time according to a pre-trained neural network, determining that the preset event in the original video stream is ended if the corresponding preset object type or the preset object behavior does not exist in continuous multiple frames;
generating alarm release information; the alarm release information includes an alarm release flag.
Preferably, the method for identifying the alarm video to be analyzed according to the local processing resource to form the marked alarm video includes:
putting the alarm video to be analyzed into a local processing queue;
when the local processing resource has the idle resource for processing the alarm video to be analyzed, extracting the alarm video to be analyzed frame by frame, inputting the extracted alarm video to a pre-trained neural network for recognizing the type of a preset object or the behavior of the preset object so as to generate a recognition frame and generate the marked alarm video.
According to a second aspect of the embodiments of the present disclosure, a method for acquiring an alert video is provided, which includes the following steps:
receiving alarm starting information sent by a video analysis module, and recording the alarm starting time; the alarm starting information is generated when the video analysis module performs video analysis according to the original video stream acquired by the acquisition terminal and determines that a preset event occurs in the original video stream;
receiving alarm release information sent by a video analysis module, and recording the alarm ending time; the alarm releasing information is generated when the video analysis module performs video analysis according to the original video stream acquired by the acquisition terminal and determines that the preset event in the original video stream is finished;
and intercepting the alarm video to be analyzed from the original video stream of the acquisition terminal according to the alarm starting time and the alarm ending time, and sending the alarm video to be analyzed to a video analysis module, so that the video analysis module identifies the alarm video to be analyzed according to local processing resources to form a marked alarm video.
Preferably, the alarm start information is determined by:
the video analysis module extracts the original video stream frame by frame according to the original video stream acquired by the acquisition terminal to obtain each frame of image; inputting each frame of image into a pre-trained neural network to perform preset object type recognition or preset object behavior recognition; within a preset frame range, if a preset object type or a preset object behavior is recognized for the first time according to a pre-trained neural network, determining that a preset event occurs in an original video stream, and generating alarm starting information; the alarm starting information comprises a target frame image carrying an alarm starting mark; the target frame image is a frame image corresponding to the type or behavior of a preset object which is recognized for the first time according to a pre-trained neural network.
Preferably, after receiving the alarm start information sent by the video analysis module, the method further includes:
and displaying the target frame image carrying the alarm starting mark.
Preferably, the alarm release information is determined by:
the video analysis module extracts the original video stream frame by frame according to the original video stream acquired by the acquisition terminal to obtain each frame of image; inputting each frame of image into a pre-trained neural network to perform preset object type recognition or preset object behavior recognition; within a preset frame range, if the preset object type or the preset object behavior is identified last time according to a pre-trained neural network, determining that the preset event in the original video stream is finished and generating alarm release information if the corresponding preset object type or the corresponding preset object behavior does not exist in continuous multiple frames; the alarm release information includes an alarm release flag.
Preferably, the method for capturing the alarm video to be analyzed from the original video stream of the acquisition terminal according to the alarm starting time and the alarm ending time includes:
sending a video stream request instruction to the acquisition terminal, so that the acquisition terminal intercepts an acquired alarm video to be analyzed from an original video stream; the video stream request instruction is provided with a video stream starting time and a video stream ending time; the video stream starting time is equal to the alarm starting time or is the time within a preset range before the alarm starting time; the video stream ending time is equal to the alarm ending time or is a time within a preset range after the alarm ending time;
and receiving the alarm video to be analyzed sent by the acquisition terminal.
Preferably, the sending the alarm video to be analyzed to a video analysis module so that the video analysis module identifies the alarm video to be analyzed according to local processing resources to form a marked alarm video, includes:
and sending the alarm video to be analyzed to a video analysis module so that the video analysis module puts the alarm video to be analyzed into a local processing queue, when the local processing resource has an idle resource for processing the alarm video to be analyzed, extracting the alarm video to be analyzed frame by frame, inputting the extracted alarm video to be analyzed into a pre-trained neural network to perform preset object type identification or preset object behavior identification so as to generate an identification frame, and generating the marked alarm video.
According to a third aspect of the embodiments of the present disclosure, there is provided an apparatus for acquiring an alert video, including a processor and a memory, where the processor reads a computer program or an instruction in the memory to perform the following operations:
performing video analysis according to an original video stream acquired by an acquisition terminal, and generating alarm starting information when a preset event occurs in the original video stream;
sending the alarm starting information to an alarm terminal so that the alarm terminal records the alarm starting time;
performing video analysis according to an original video stream acquired by an acquisition terminal, and generating alarm release information when a preset event in the original video stream is determined to be finished;
sending the alarm release information to an alarm terminal so that the alarm terminal records the alarm ending time;
receiving an alarm video to be analyzed sent by an alarm terminal; the alarm video to be analyzed is obtained by the alarm terminal by intercepting the original video stream of the acquisition terminal according to the alarm starting time and the alarm ending time;
and identifying the alarm video to be analyzed according to local processing resources to form a marked alarm video.
According to a fourth aspect of the embodiments of the present disclosure, there is provided an apparatus for acquiring an alert video, including a processor and a memory, where the processor reads a computer program or instructions in the memory and is configured to:
receiving alarm starting information sent by a video analysis module, and recording the alarm starting time; the alarm starting information is generated when the video analysis module performs video analysis according to the original video stream acquired by the acquisition terminal and determines that a preset event occurs in the original video stream;
receiving alarm release information sent by a video analysis module, and recording the alarm ending time; the alarm release information is generated when the video analysis module performs video analysis according to the original video stream acquired by the acquisition terminal and determines that the preset event in the original video stream is finished;
and intercepting the alarm video to be analyzed from the original video stream of the acquisition terminal according to the alarm starting time and the alarm ending time, and sending the alarm video to be analyzed to a video analysis module, so that the video analysis module identifies the alarm video to be analyzed according to local processing resources to form a marked alarm video.
The method and the device for acquiring the alarm video can accurately perform video analysis on the original video stream, determine the conditions of alarm starting and alarm relieving, and do not need to send the related alarm video to the alarm terminal in real time, but the alarm terminal is responsible for recording the alarm starting time and the alarm ending time, and then the alarm terminal intercepts the related alarm video from the original video stream of the acquisition terminal by means of the alarm starting time and the alarm ending time; therefore, the video analysis module can identify the alarm video to be analyzed according to the local processing resources to form the marked alarm video. Therefore, the embodiment of the disclosure does not need to send the video stream to the alarm terminal in real time to cause overlarge load of network resources and processing resources, and can also fully utilize the local processing resources of the video analysis module to provide the alarm video for the alarm terminal and ensure the complete recording of the alarm video.
Drawings
Fig. 1 is a schematic diagram of an implementation provided by an embodiment of the present disclosure;
fig. 2 is a first flowchart of a method for acquiring an alarm video according to an embodiment of the present disclosure;
fig. 3 is a second flowchart of a method for acquiring an alarm video according to an embodiment of the present disclosure;
FIG. 4 is a schematic diagram of another implementation provided by an embodiment of the present disclosure;
fig. 5 is a schematic structural diagram of an apparatus for acquiring an alert video according to an embodiment of the present invention.
Detailed Description
The technical contents of the present disclosure are further described in detail below with reference to the accompanying drawings and specific embodiments.
In order to make the present disclosure better understood by those skilled in the art, some technical terms related to the present disclosure are explained as follows:
and (4) warning terminal: the terminal for displaying the alarm may display an alarm video, an alarm image, and the like, and the alarm terminal may be a computer, a mobile phone, a tablet computer, a PDA (personal Di digital As s i state) handheld terminal, a video phone, and the like.
Collecting a terminal: the terminal for completing the video image acquisition work to provide a video stream to a video analysis service (which can run on a video analysis service module, and the video analysis service module can be arranged on an external server, a computer or the acquisition terminal), and the acquisition terminal can be an image acquisition device, such as a camera.
Frame extraction: the video analysis system is used for converting an original video stream acquired by the acquisition terminal into each frame of image for subsequent video analysis service.
Video analysis service: the method is characterized in that each frame of image is identified through an image identification technology, and whether the condition that the alarm rule is met exists in each frame of image is determined.
And (3) alarm rules: and when the alarm condition is met, the identified alarm image needs to be transmitted to an alarm terminal.
In the process of implementing the embodiments of the present disclosure, the inventor finds that the embodiments of the present disclosure can be implemented in the following manner, as shown in fig. 1, which generally includes the following processes:
A. the acquisition terminal sends the video stream to a video analysis service.
B. The video analysis service obtains a video stream, and the video stream is converted into a picture by the frame extraction module and is transmitted to the analysis module.
C. The analysis module carries out algorithm analysis on the received picture.
D. And after the analysis module identifies the condition needing alarming, the alarm identification is added to the picture, and the alarm information and the alarm picture are pushed to an alarm service. The alarm is divided into two types of generation and release.
E. After receiving the alarm push, the alarm service distributes the alarm information to a terminal (alarm terminal), and the terminal finishes the presentation.
F. After receiving the alarm notification, the alarm service may request the video analysis module for a video stream to which a label of the identification result has been added. And the video analysis module stores the video stream and starts to record.
G. And after receiving the alarm release push, the alarm service distributes the alarm release to the terminal. And simultaneously requesting the video analysis module to end the video stream to finish recording.
However, the above process of steps a-G has the following drawbacks:
requesting real-time video places a significant burden on the analysis service. If multiple devices alarm simultaneously, it may cause the analysis service to be overloaded instantaneously, resulting in failure to request real-time video and then failure to record. In severe cases, the service being analyzed may be affected, and the user experience is poor.
When no alarm is given, the video analysis service is light in load, and waste of hardware resources is generated.
The recording action occurs after the real-time video request action, at this time, an alarm has been generated, and the recorded video cannot completely record the alarm process.
In order to overcome the above drawback, as shown in fig. 2, an embodiment of the present disclosure provides a method for acquiring an alarm video, which takes a video analysis module as an execution main body, and includes the following steps:
And 102, sending the alarm starting information to an alarm terminal so that the alarm terminal records the alarm starting time.
And 103, performing video analysis according to the original video stream acquired by the acquisition terminal, and generating alarm release information when the preset event in the original video stream is determined to be finished.
And step 104, sending the alarm release information to an alarm terminal so that the alarm terminal records the alarm ending time.
105, receiving an alarm video to be analyzed, which is sent by an alarm terminal; and the alarm video to be analyzed is obtained by intercepting the original video stream of the acquisition terminal by the alarm terminal according to the alarm starting time and the alarm ending time.
And 106, identifying the alarm video to be analyzed according to local processing resources to form a marked alarm video.
It should be noted that the method for acquiring an alarm video provided in the embodiment of the present disclosure may be applied to a video information processing system 30 shown in fig. 3, where the system 30 may include an acquisition terminal 31 (which may be an image acquisition device, such as a camera, etc.), an alarm terminal 32 (which may be a computer, a mobile phone, a tablet computer, a PDA handheld terminal, a video telephone, etc.), and a video analysis module 33 (which may be disposed at the acquisition terminal 31, or may be disposed at a separate computer or server, and for example, may be implemented by using a video analysis server loaded with various algorithms and services).
In the above steps 101 and 103, performing video analysis according to the original video stream acquired by the acquisition terminal may be implemented in the following manner:
and extracting the original video stream frame by frame according to the original video stream acquired by the acquisition terminal to obtain each frame of image. The acquisition terminal acquires the original video stream, and can acquire the original video stream in real time according to the preset acquisition frequency to form the original video stream. Therefore, when the video analysis module obtains the original video stream, it also needs to perform frame-by-frame extraction to obtain each frame image within a period of time, so as to facilitate the subsequent analysis processing.
And inputting each frame of image into a pre-trained neural network to perform preset object type recognition or preset object behavior recognition. Specifically, the neural network may be a convolutional neural network capable of performing object type identification, or may be a convolutional neural network capable of performing object behavior identification, but is not limited thereto. The neural network can be trained through sample data in advance, for example, in an object type recognition task, the neural network can recognize objects such as people, vehicles, animals and the like; for example, in the object behavior recognition task, the neural network may recognize a human behavior (whether a specific behavior such as whether a helmet is worn or whether a user is focused on or not), and recognize an object behavior such as a vehicle (e.g., collision or movement). There are many prior arts for specific neural networks to implement object type recognition and object behavior recognition, and details thereof are not repeated here.
In step 101, when it is determined that a preset event occurs in the original video stream, the alarm start information is generated, which may be implemented in the following manner:
within a preset frame range, if a preset object type or a preset object behavior is recognized for the first time according to a pre-trained neural network, a preset event in an original video stream is determined to occur. For example, in the range of n frames (100 frames, 1000 frames, etc., but not limited thereto, any frame range may exist), if it is recognized that someone appears in a certain monitoring area for the first time, it is determined that a preset event appears in the original video stream; for another example, in the range of n frames, if it is recognized for the first time that a safety helmet is not worn in a certain monitoring area, it is determined that a preset event occurs in the original video stream. The above is merely an example of two preset events, but not limited thereto, and it should be understood that any identification cases obtained by performing type identification and behavior identification on an object are included in the protection scope of the present disclosure. The object type identification and behavior identification disclosed by the present disclosure can be applied to regional anti-theft monitoring (such as anti-theft of vehicles, cultural relics, etc.), regional personnel monitoring (people are not allowed to appear in some regions), construction specification monitoring (corresponding safety equipment must be worn in some construction regions), concentration monitoring (such as monitoring whether cheating is done in an examination room, monitoring whether sleeping is done in a classroom, talking, etc.), and the like, but are not limited thereto.
Generating alarm starting information; the alarm starting information comprises a target frame image carrying an alarm starting mark; the target frame image is a frame image corresponding to the type or behavior of the preset object which is recognized for the first time according to a pre-trained neural network. For example, in the range of n frames, for example, in the range of 100 frames, the preset object type or the preset object behavior is not recognized in the first 20 frames, and the 21 st frame image recognizes the preset object type or the preset object behavior, the 21 st frame image is taken as the target frame image. The above is only a specific example, but not limited thereto.
In step 101, after generating the alert start information, the method for acquiring an alert video according to the embodiment of the present disclosure may further include the following steps:
and sending the alarm starting information to an alarm terminal so that the alarm terminal displays a target frame image carrying an alarm starting mark. For example, the target frame image of the alarm start flag is displayed on a computer screen of a security monitoring room, a mobile phone of a security person, or the like, but is not limited thereto. Therefore, the alarm terminal can be firstly alarmed under the condition that the complete video is not provided for the alarm terminal temporarily, so that the alarm terminal can obtain the occurrence of the preset event in time.
In step 103, when it is determined that the preset event in the original video stream is ended, the alarm release information is generated, which may be implemented as follows:
and in a preset frame range, if the preset object type or the preset object behavior is identified last time according to a pre-trained neural network, determining that the preset event in the original video stream is ended if the corresponding preset object type or the preset object behavior does not exist in continuous multiple frames. For example, in the range of n frames, taking 100 frames as an example, the preset object type or the preset object behavior is recognized at the beginning of the 21 st frame, then the preset object type or the preset object behavior is recognized at the 22 nd to 35 th frames respectively, then the corresponding preset object type or the preset object behavior is not recognized at the following frames, and assuming that the threshold of the consecutive frames is 50 frames, it is determined that the preset event in the original video stream is ended after the recognition result of the 85 th frame. This is merely an example, and not intended to be limiting.
Generating alarm release information; the alarm release information includes an alarm release flag.
In step 106, the alarm video to be analyzed is identified according to the local processing resource to form a marked alarm video, which may be implemented as follows:
and putting the alarm video to be analyzed into a local processing queue. The local processing queue is used for determining the processing sequence of the alarm video to be analyzed, so that the subsequent processing of the alarm video to be analyzed is smoother, and the condition that a large number of alarm videos to be analyzed are stacked to occupy resources is avoided.
When the local processing resources have idle resources for processing the alarm video to be analyzed, extracting the alarm video to be analyzed frame by frame, inputting the extracted alarm video to a pre-trained neural network for identifying the type of a preset object or identifying the behavior of the preset object so as to generate an identification frame and generate the labeled alarm video. Specifically, the resources for processing the alarm video to be analyzed may include a memory, a CPU, a disk, a network, and the like, and when the resources of the video analysis module are idle, the alarm video to be analyzed in the local processing queue may be processed. The specific processing procedure may still adopt the processing of the neural network, and is not described herein again.
In addition, as shown in fig. 3, an embodiment of the present disclosure provides a method for acquiring an alert video, where an alert terminal is used as an execution main body, and the method includes the following steps:
Wherein, the alarm starting information can be determined by the following method:
the video analysis module extracts the original video stream frame by frame according to the original video stream acquired by the acquisition terminal to obtain each frame of image; inputting each frame of image into a pre-trained neural network to perform preset object type recognition or preset object behavior recognition; within a preset frame range, if a preset object type or a preset object behavior is recognized for the first time according to a pre-trained neural network, determining that a preset event occurs in an original video stream, and generating alarm starting information; the alarm starting information comprises a target frame image carrying an alarm starting mark; the target frame image is a frame image corresponding to the type or behavior of the preset object which is recognized for the first time according to a pre-trained neural network. Here, the determination of the alarm start information may refer to step 101 described above, and is not described here again.
In step 201, after receiving the alarm start information sent by the video analysis module, the method may further include:
and displaying the target frame image carrying the alarm starting mark. For example, the target frame image of the alarm start flag is displayed on a computer screen of a security monitoring room, a mobile phone of a security personnel, or the like, but is not limited thereto. Therefore, the alarm terminal can be firstly alarmed under the condition that the complete video is not provided for the alarm terminal temporarily, so that the alarm terminal can obtain the occurrence of the preset event in time.
The alarm release information may be determined in the following manner:
the video analysis module extracts the original video stream frame by frame according to the original video stream acquired by the acquisition terminal to obtain each frame of image; inputting each frame of image into a pre-trained neural network to perform preset object type recognition or preset object behavior recognition; within a preset frame range, if the preset object type or the preset object behavior is identified last time according to a pre-trained neural network, determining that the preset event in the original video stream is finished and generating alarm release information if the corresponding preset object type or the corresponding preset object behavior does not exist in continuous multiple frames; the alarm release information includes an alarm release flag. Here, the determination of the alarm release information may refer to step 103 described above, and is not described here in detail.
In step 203, the alarm video to be analyzed, which is obtained by intercepting the original video stream of the acquisition terminal according to the alarm starting time and the alarm ending time, may be implemented in the following manner:
sending a video stream request instruction to the acquisition terminal, so that the acquisition terminal intercepts an acquired alarm video to be analyzed from an original video stream; the video stream request instruction is provided with a video stream starting time and a video stream ending time; the video stream starting time is equal to the alarm starting time or is the time within a preset range before the alarm starting time; the video stream ending time is equal to the alarm ending time or is a time within a preset range after the alarm ending time. Here, in order to obtain the alarm video to be analyzed by interception, the alarm video may be intercepted by the alarm starting time and the alarm ending time, however, since the preset event may have occurred for a while when the alarm starts, in order to enable the alarm video to be analyzed to more comprehensively show the occurrence process of the preset event, the video stream starting time may be a time within a preset range before the alarm starting time, and the video stream ending time may be a time within a preset range after the alarm ending time, so that the intercepted alarm video to be analyzed may further include the occurrence of the preset event, and the missing is avoided.
And receiving the alarm video to be analyzed sent by the acquisition terminal.
In step 203, the alarm video to be analyzed is sent to a video analysis module, so that the video analysis module identifies the alarm video to be analyzed according to local processing resources to form a marked alarm video, which may be implemented as follows:
and sending the alarm video to be analyzed to a video analysis module so that the video analysis module puts the alarm video to be analyzed into a local processing queue, when the local processing resource has a free resource for processing the alarm video to be analyzed, extracting the alarm video to be analyzed frame by frame, inputting the extracted alarm video to be analyzed into a pre-trained neural network for recognizing a preset object type or a preset object behavior so as to generate a recognition frame and generate a marked alarm video. The specific implementation process may refer to step 106, which is not described herein again.
Therefore, through the above steps 101 to 106, the following specific process shown in fig. 4 can be implemented in steps 201 to 203:
A. and the acquisition terminal sends the video stream to a video analysis service.
B. The video analysis service obtains a video stream, and the video stream is converted into a picture by the frame extraction module and is transmitted to the analysis module.
C. The analysis module carries out algorithm analysis on the received picture.
D. And after the analysis module identifies the alarm, the alarm identification is added to the picture, and the alarm information and the alarm picture are pushed to an alarm service. The alarm is divided into two types of generation and release.
E. And after the alarm service is subjected to alarm pushing, the alarm information is distributed to the terminal, and the terminal finishes presenting.
F. After receiving the alarm notice, the alarm service records the alarm starting event and the alarm ending time.
G. The alarm service requests a complete video recording from the video acquisition terminal according to the alarm starting time and the alarm ending time, and can slightly push the alarm generating time forward to ensure the integrity of the alarm process.
H. And after the video recording is requested, pushing the video stream to a video analysis service.
I. After receiving the video stream, the video analysis service can be put into a queue, and then analysis is carried out when the service is idle, frame-by-frame identification is carried out during analysis, an identification frame is added, a marked alarm video is generated, a magnetic disk is stored, and recording is completed.
Wherein, the terminal and the alarm service can be understood as the alarm terminal as an execution subject; the acquisition terminal is the execution main body of the acquisition terminal; the frame extracting module and the analyzing module in the video analyzing service can be understood as the video analyzing module as an executing body. The processes a to I herein are only for the purpose of more clearly and intuitively understanding the present disclosure by those skilled in the art, and are not intended to limit the present disclosure.
In addition, as shown in fig. 5, an apparatus for acquiring an alert video is further provided in the embodiment of the present disclosure, and includes a processor 32 and a memory 31, and may further include a communication component, a sensor component, a power component, a multimedia component, and an input/output interface according to actual needs. The memory, communication components, sensor components, power components, multimedia components, and input/output interfaces are all connected to the processor 32. As mentioned above, the memory 31 may be a Static Random Access Memory (SRAM), an Electrically Erasable Programmable Read Only Memory (EEPROM), an Erasable Programmable Read Only Memory (EPROM), a Programmable Read Only Memory (PROM), a Read Only Memory (ROM), a magnetic memory, a flash memory, etc.; the processor 32 may be a Central Processing Unit (CPU), graphics Processing Unit (GPU), field Programmable Gate Array (FPGA), application specific integrated circuit (AS ic), digital Signal Processing (DSP) chip, or the like. Other communication components, sensor components, power components, multimedia components, etc. may be implemented using common components found in existing smartphones and are not specifically described herein.
In addition, an apparatus for acquiring an alert video according to an embodiment of the present disclosure includes a processor 32 and a memory 31, where the processor 32 reads a computer program or an instruction in the memory 31 to perform the following operations:
and performing video analysis according to the original video stream acquired by the acquisition terminal, and generating alarm starting information when a preset event occurs in the original video stream.
And sending the alarm starting information to an alarm terminal so that the alarm terminal records the alarm starting moment.
And performing video analysis according to the original video stream acquired by the acquisition terminal, and generating alarm release information when the preset event in the original video stream is determined to be finished.
And sending the alarm release information to an alarm terminal so that the alarm terminal records the alarm ending time.
Receiving an alarm video to be analyzed sent by an alarm terminal; and the alarm video to be analyzed is obtained by intercepting the original video stream of the acquisition terminal by the alarm terminal according to the alarm starting time and the alarm ending time.
And identifying the alarm video to be analyzed according to local processing resources to form the marked alarm video.
In addition, an apparatus for acquiring an alert video is further provided in the embodiments of the present disclosure, and includes a processor 32 and a memory 31, where the processor 32 reads a computer program or an instruction in the memory 31, and is configured to perform the following operations:
receiving alarm starting information sent by a video analysis module, and recording the alarm starting time; the alarm starting information is generated when the video analysis module performs video analysis according to the original video stream acquired by the acquisition terminal and determines that a preset event occurs in the original video stream.
Receiving alarm release information sent by a video analysis module, and recording the alarm ending time; the alarm release information is generated when the video analysis module performs video analysis according to the original video stream acquired by the acquisition terminal and determines that the preset event in the original video stream is finished.
And intercepting the alarm video to be analyzed from the original video stream of the acquisition terminal according to the alarm starting time and the alarm ending time, and sending the alarm video to be analyzed to a video analysis module, so that the video analysis module identifies the alarm video to be analyzed according to local processing resources to form a marked alarm video.
In addition, an embodiment of the present disclosure further provides a computer-readable storage medium, where instructions are stored on the computer-readable storage medium, and when the instructions are run on a computer, the computer is enabled to execute the method for acquiring an alert video according to fig. 2 or fig. 3, and details of a specific implementation manner of the method are not repeated here.
In addition, an embodiment of the present disclosure further provides a computer program product including instructions, which when run on a computer, causes the computer to execute the method for acquiring an alarm video according to fig. 2 or fig. 3, and details of the specific implementation manner are not repeated here.
The method and the device for acquiring the alarm video can accurately perform video analysis on the original video stream, determine the conditions of alarm starting and alarm releasing, and the alarm terminal is responsible for recording the alarm starting time and the alarm ending time without transmitting the related alarm video to the alarm terminal in real time, and then the alarm terminal intercepts the related alarm video from the original video stream of the acquisition terminal by means of the alarm starting time and the alarm ending time; therefore, the video analysis module can identify the alarm video to be analyzed according to the local processing resources to form the marked alarm video. Therefore, the method and the device do not need to send the video stream to the alarm terminal in real time to cause overlarge load of network resources and processing resources, can fully utilize local processing resources of the video analysis module, provide the alarm video for the alarm terminal and ensure the complete recording of the alarm video.
The method and the device for acquiring the warning video provided by the present disclosure are described in detail above. It will be apparent to those skilled in the art that various changes may be made without departing from the spirit and scope of the disclosure.
Claims (7)
1. A method for acquiring an alarm video is characterized by comprising the following steps:
receiving alarm starting information sent by a video analysis module, and recording the alarm starting time; the alarm starting information is generated when the video analysis module performs video analysis according to the original video stream acquired by the acquisition terminal and determines that a preset event occurs in the original video stream;
receiving alarm release information sent by a video analysis module, and recording the alarm ending time; the alarm releasing information is generated when the video analysis module performs video analysis according to the original video stream acquired by the acquisition terminal and determines that the preset event in the original video stream is finished;
intercepting an alarm video to be analyzed from an original video stream of the acquisition terminal according to the alarm starting time and the alarm ending time, and sending the alarm video to be analyzed to a video analysis module, so that the video analysis module identifies the alarm video to be analyzed according to local processing resources to form a marked alarm video; wherein,
the video analysis module is arranged at a server, a computer or the acquisition terminal, the acquisition terminal is an image acquisition device,
wherein the alarm start information is determined by: the video analysis module extracts the original video stream frame by frame according to the original video stream acquired by the acquisition terminal to obtain each frame of image; inputting each frame of image into a pre-trained neural network to perform preset object type recognition or preset object behavior recognition; within a preset frame range, if a preset object type or a preset object behavior is recognized for the first time according to a pre-trained neural network, determining that a preset event occurs in an original video stream, and generating alarm starting information;
the alarm release information is determined by the following method: the video analysis module extracts the original video stream frame by frame according to the original video stream acquired by the acquisition terminal to obtain each frame of image; inputting each frame of image into a pre-trained neural network to perform preset object type recognition or preset object behavior recognition; within a preset frame range, if the preset object type or the preset object behavior is identified last time according to a pre-trained neural network, determining that the preset event in the original video stream is finished if the corresponding preset object type or the corresponding preset object behavior does not exist in continuous multiple frames, and generating alarm release information.
2. The method for acquiring an alert video according to claim 1, wherein the alert start information includes a target frame image carrying an alert start flag; the target frame image is a frame image corresponding to the type or behavior of a preset object which is recognized for the first time according to a pre-trained neural network.
3. The method for acquiring an alert video according to claim 2, further comprising, after receiving the alert start information sent by the video analysis module: and displaying the target frame image carrying the alarm starting mark.
4. The method for acquiring an alert video according to claim 1, wherein the alert release information includes an alert release flag.
5. The method for acquiring an alarm video according to claim 1, wherein the step of capturing the alarm video to be analyzed from the original video stream of the collection terminal according to the alarm start time and the alarm end time comprises:
sending a video stream request instruction to the acquisition terminal, so that the acquisition terminal intercepts an acquired alarm video to be analyzed from an original video stream; the video stream request instruction is provided with a video stream starting time and a video stream ending time; the video stream starting time is equal to the alarm starting time or is the time within a preset range before the alarm starting time; the video stream ending time is equal to the alarm ending time or is a time within a preset range after the alarm ending time;
and receiving the alarm video to be analyzed sent by the acquisition terminal.
6. The method for acquiring an alarm video according to claim 1, wherein the sending the alarm video to be analyzed to a video analysis module so that the video analysis module identifies the alarm video to be analyzed according to local processing resources to form a labeled alarm video comprises:
and sending the alarm video to be analyzed to a video analysis module so that the video analysis module puts the alarm video to be analyzed into a local processing queue, when the local processing resource has a free resource for processing the alarm video to be analyzed, extracting the alarm video to be analyzed frame by frame, inputting the extracted alarm video to be analyzed into a pre-trained neural network for recognizing a preset object type or a preset object behavior so as to generate a recognition frame and generate a marked alarm video.
7. An apparatus for acquiring an alert video, comprising a processor and a memory, wherein the processor reads a computer program or instructions in the memory for executing the method for acquiring an alert video according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011088030.4A CN112419638B (en) | 2020-10-13 | 2020-10-13 | Method and device for acquiring alarm video |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011088030.4A CN112419638B (en) | 2020-10-13 | 2020-10-13 | Method and device for acquiring alarm video |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112419638A CN112419638A (en) | 2021-02-26 |
| CN112419638B true CN112419638B (en) | 2023-03-14 |
Family
ID=74854853
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011088030.4A Active CN112419638B (en) | 2020-10-13 | 2020-10-13 | Method and device for acquiring alarm video |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112419638B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114419502A (en) * | 2022-01-12 | 2022-04-29 | 深圳力维智联技术有限公司 | Data analysis method and device and storage medium |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| FR2856222B3 (en) * | 2003-06-16 | 2005-12-09 | Daniel Armand | IMAGE AND / OR SOUND RECORDING DEVICE FOR CAPTURING AND RESTITUING A VIDEO SEQUENCE RETRACING THE CIRCUMSTANCES OF THE TRIGGERING OF AN ALERT DEVICE |
| KR20080075091A (en) * | 2005-09-12 | 2008-08-14 | 쓰리브이알 시큐리티, 인크. | Storage of video analysis data for real-time alerting and forensic analysis |
| US7570159B2 (en) * | 2007-08-24 | 2009-08-04 | Robelight Llc | System and method for providing visual and physiological cues in a security matching system |
| CN104349119A (en) * | 2013-07-24 | 2015-02-11 | 浙江大华技术股份有限公司 | Monitoring video recording method and device |
| CN104581064A (en) * | 2014-12-26 | 2015-04-29 | 深圳市兰丁科技有限公司 | Method and device for integrating camera alarm information |
| CN109257569A (en) * | 2018-10-24 | 2019-01-22 | 广东佳鸿达科技股份有限公司 | Security protection video monitoring analysis method |
| CN109462738A (en) * | 2018-10-25 | 2019-03-12 | 江苏跃鑫科技有限公司 | One kind automatically tracking monitoring method |
| CN110211316A (en) * | 2019-06-05 | 2019-09-06 | 深圳市梦网百科信息技术有限公司 | A kind of monitoring and early warning method and system |
| CN110782639A (en) * | 2019-10-28 | 2020-02-11 | 深圳奇迹智慧网络有限公司 | Abnormal behavior warning method, device, system and storage medium |
| CN111105047A (en) * | 2019-12-12 | 2020-05-05 | 陕西瑞海工程智慧数据科技有限公司 | Operation and maintenance monitoring method and device, electronic equipment and storage medium |
| CN111212321A (en) * | 2020-01-10 | 2020-05-29 | 上海摩象网络科技有限公司 | Video processing method, device, equipment and computer storage medium |
| CN111405222A (en) * | 2019-12-12 | 2020-07-10 | 杭州海康威视系统技术有限公司 | Video alarm method, video alarm system and alarm picture acquisition method |
| CN111601080A (en) * | 2020-05-12 | 2020-08-28 | 杭州武盛广告制作有限公司 | Video management system for community security monitoring video storage |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103945156A (en) * | 2014-04-16 | 2014-07-23 | 深圳英飞拓科技股份有限公司 | Alarm video replaying method and system |
-
2020
- 2020-10-13 CN CN202011088030.4A patent/CN112419638B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| FR2856222B3 (en) * | 2003-06-16 | 2005-12-09 | Daniel Armand | IMAGE AND / OR SOUND RECORDING DEVICE FOR CAPTURING AND RESTITUING A VIDEO SEQUENCE RETRACING THE CIRCUMSTANCES OF THE TRIGGERING OF AN ALERT DEVICE |
| KR20080075091A (en) * | 2005-09-12 | 2008-08-14 | 쓰리브이알 시큐리티, 인크. | Storage of video analysis data for real-time alerting and forensic analysis |
| US7570159B2 (en) * | 2007-08-24 | 2009-08-04 | Robelight Llc | System and method for providing visual and physiological cues in a security matching system |
| CN104349119A (en) * | 2013-07-24 | 2015-02-11 | 浙江大华技术股份有限公司 | Monitoring video recording method and device |
| CN104581064A (en) * | 2014-12-26 | 2015-04-29 | 深圳市兰丁科技有限公司 | Method and device for integrating camera alarm information |
| CN109257569A (en) * | 2018-10-24 | 2019-01-22 | 广东佳鸿达科技股份有限公司 | Security protection video monitoring analysis method |
| CN109462738A (en) * | 2018-10-25 | 2019-03-12 | 江苏跃鑫科技有限公司 | One kind automatically tracking monitoring method |
| CN110211316A (en) * | 2019-06-05 | 2019-09-06 | 深圳市梦网百科信息技术有限公司 | A kind of monitoring and early warning method and system |
| CN110782639A (en) * | 2019-10-28 | 2020-02-11 | 深圳奇迹智慧网络有限公司 | Abnormal behavior warning method, device, system and storage medium |
| CN111105047A (en) * | 2019-12-12 | 2020-05-05 | 陕西瑞海工程智慧数据科技有限公司 | Operation and maintenance monitoring method and device, electronic equipment and storage medium |
| CN111405222A (en) * | 2019-12-12 | 2020-07-10 | 杭州海康威视系统技术有限公司 | Video alarm method, video alarm system and alarm picture acquisition method |
| CN111212321A (en) * | 2020-01-10 | 2020-05-29 | 上海摩象网络科技有限公司 | Video processing method, device, equipment and computer storage medium |
| CN111601080A (en) * | 2020-05-12 | 2020-08-28 | 杭州武盛广告制作有限公司 | Video management system for community security monitoring video storage |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112419638A (en) | 2021-02-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112419639A (en) | Video information acquisition method and device | |
| AU2009243916B2 (en) | A system and method for electronic surveillance | |
| US8854474B2 (en) | System and method for quick object verification | |
| JP2022048147A5 (en) | ||
| KR20110130033A (en) | Active image monitoring system using motion pattern database, and method thereof | |
| CN108200405A (en) | A kind of video monitoring system based on recognition of face | |
| CN108288025A (en) | A kind of car video monitoring method, device and equipment | |
| CN112017323A (en) | Patrol alarm method and device, readable storage medium and terminal equipment | |
| CN110751125A (en) | Wearing detection method and device | |
| CN115273231A (en) | Information processing method, information processing apparatus, storage medium, and electronic device | |
| CN111191507A (en) | Safety early warning analysis method and system for smart community | |
| CN111178241A (en) | Intelligent monitoring system and method based on video analysis | |
| US10783365B2 (en) | Image processing device and image processing system | |
| CN114187561A (en) | Abnormal behavior identification method and device, terminal equipment and storage medium | |
| KR102182660B1 (en) | System and method for detecting violence using spatiotemporal features and computer readable recording media storing program for executing method thereof | |
| CN112419638B (en) | Method and device for acquiring alarm video | |
| JP6437217B2 (en) | Image output device, image management system, image processing method, and program | |
| CN111581436A (en) | Target identification method and device, computer equipment and storage medium | |
| JPWO2018037665A1 (en) | INFORMATION PROCESSING APPARATUS, INFORMATION PROCESSING SYSTEM, CONTROL METHOD, AND PROGRAM | |
| CN116797993B (en) | Monitoring method, system, medium and equipment based on intelligent community scene | |
| CN114764895A (en) | Abnormal behavior detection device and method | |
| CN112883816A (en) | Information pushing method and device | |
| CN109120896B (en) | Security video monitoring guard system | |
| CN111695445A (en) | Face recognition method, device, equipment and computer readable storage medium | |
| CN113038261A (en) | Video generation method, device, equipment, system and storage medium |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |