WO2021223665A1

WO2021223665A1 - Method and device for reducing invalid warnings

Info

Publication number: WO2021223665A1
Application number: PCT/CN2021/091209
Authority: WO
Inventors: 陶冶; 汪红星
Original assignee: 华为技术有限公司
Priority date: 2020-05-08
Filing date: 2021-04-29
Publication date: 2021-11-11
Also published as: CN113628404A

Abstract

A method and device for reducing invalid warnings. The method is applied to a security monitoring system. The method comprises: when a camera or a server detects that a behavior of a first target is a warning triggering behavior, obtaining first feature information of the first target; and then matching the first feature information with feature information in an invalid warning database. Because the invalid warning database stores feature information of targets that triggered invalid warnings before, and/or pre-stored feature information of targets that are prone to triggering invalid warnings, when the matching degree between the first feature information and the feature information in the invalid warning database is high (e.g., greater than or equal to a first threshold), this indicates that the warning triggered by the first target is an invalid warning, and the warning information triggered by the first target can be filtered, i.e., no warning information is outputted. The probability of generating an invalid warning is reduced, and the warning accuracy is improved, thereby improving user experience.

Description

Method and device for reducing invalid alarm

This application claims the priority of the Chinese patent application with the application number 202010383960.6 and the invention title "A method and device for reducing invalid alarms" filed on May 8, 2020, the entire content of which is incorporated into this application by reference.

Technical field

This application relates to the field of image processing, and in particular to a method and device for reducing invalid alarms.

Background technique

For the sake of safety, some special areas, such as railways, zoo fences or precious items on display, need to take security measures. Traditional security measures are mainly manual patrols. Traditional security measures have disadvantages such as high labor costs, low security efficiency, and poor environmental adaptability. Nowadays, more intelligent abnormal intrusion detection technologies are introduced for these special areas, such as triggering an alarm when abnormal intrusion behavior is detected.

The current anomaly intrusion detection is generally based on image processing technology. In the abnormal intrusion detection technology based on image processing technology, at least one camera is first deployed around the security area, the camera can shoot in real time, and then the image captured by the at least one camera is recognized. When it is determined that there is a moving target in the security area, and it is determined that the behavior of the target is a preset behavior that triggers an alarm, an alarm is triggered, and the alarm information of the target is reported to the staff for processing.

For example, the preset behavior that triggers an alarm is human wandering behavior, and because of the diversity of the environment in the security area, the plants around the security area will also shake back and forth under the blow of the wind, which may be misidentified Wandering as a person, triggering an alarm. But in fact, the plant does not belong to the security target of the security area, so the staff will receive invalid warning information. At the same time, plants generally stay in the picture taken by the camera for a long time, which may trigger multiple alarms, and correspondingly, the staff will receive multiple invalid alarm messages. It can be seen that in the current abnormal intrusion detection technology, the probability of generating invalid alarms is high, resulting in low accuracy of alarms.

Summary of the invention

The embodiments of the present application provide a method and device for reducing invalid alarms, which can reduce the probability of generating invalid alarms, improve the accuracy of alarms, and improve user experience.

In the first aspect, the embodiments of the present application provide a method for reducing invalid alarms. The method can be implemented by a camera, or a chip or a chip system in the camera, or a server, or a chip or a chip system in the server. Including: first obtain the first characteristic information of the first target, and then match the first characteristic information with the characteristic information in the invalid alarm database, when the degree of matching between the first characteristic information and the characteristic information in the invalid alarm database is less than At the first threshold, no alarm information is output.

Wherein, the alarm database includes characteristic information of at least one target, and the behavior of each target in the at least one target is a behavior that triggers an invalid alarm. The above-mentioned behavior of the first target is the behavior of triggering an alarm. For example, if the above method is implemented by a camera, the behavior of the first target is a behavior that triggers an alarm of the camera. For another example, if the foregoing method is implemented by a server, the behavior of the first target may be a behavior that triggers a camera alarm, or a behavior that triggers a server alarm.

In the above technical solution, when the camera or the server detects that the behavior of the first target is a behavior that triggers a warning (such as crossing the line, area intrusion, or wandering, etc.), the first characteristic information of the first target is acquired, and then the first characteristic The information is matched with the characteristic information in the invalid alarm database. Since the invalid alarm database stores the characteristic information of the target that previously triggered the invalid alarm, and/or the characteristic information of the target that easily triggers the invalid alarm, when the first characteristic information matches the characteristic information in the invalid alarm database When the level is high (such as greater than or equal to the first threshold), it indicates that the alarm triggered by the first target is an invalid alarm. The alarm information triggered by the first target can be automatically filtered, that is, no alarm information is output, thereby saving manual detection. Link, reduce the probability of generating invalid alarms, improve the accuracy of alarms, thereby improving user experience.

In other words, the device (such as the alarm server or cloud service platform) that implements the method of this solution will receive frequent alarms from the camera, but after further confirmation by the alarm platform, if the matching degree between the feature information and the feature information in the invalid alarm database is better High, which means that the alarm is an error alarm (false alarm) that does not need attention. Therefore, the alarm platform will terminate the alarm and will not actually output the alarm information to the user, reducing the interference of the user from the false alarm.

In a possible design of the first aspect, when the degree of matching between the first feature information and the feature information in the invalid alarm database is less than the first threshold, the first information from the input interface is received, and the first information is used To indicate whether the alarm triggered by the first target is an invalid alarm; when the first information is used to indicate that the alarm triggered by the first target is an invalid alarm, no alarm information is output, and the first characteristic information is added to the invalid alarm database, Otherwise, output alarm information.

Taking the foregoing method implemented by the server as an example, when the degree of matching between the first feature information and the feature information in the invalid alarm database is low (for example, less than the first threshold), the server may include the first image of the first target Display so that the staff can judge whether the alarm triggered by the first target is an invalid alarm according to the first image, and input the judgment result (recorded as the first information) (such as mouse input, keyboard input or voice input) into the input Interface. When the first information is used to indicate that the alarm triggered by the first target is a valid alarm, the server outputs alarm information; when the first information is used to indicate that the alarm triggered by the first target is an invalid alarm, the server filters the alarm information triggered by the first target , That is, no alarm information is output, and the first characteristic information is added to the invalid alarm database. The matching degree between the feature information in the current invalid alarm database and the first feature information is low, but the staff determines that the alarm triggered by the first target is an invalid alarm, and the server adds the first feature information to the invalid alarm database. The diversity of samples in the invalid alarm database can be increased. When the first target triggers an alarm again, the alarm information triggered by the first target can be automatically filtered, without the need for staff to check again, which reduces the overhead of manual maintenance and improves user experience.

In a possible design of the first aspect, adding the first feature information to the invalid alarm database includes: acquiring a first image, the first image includes a first target; determining that the first target is in the first image Then, the area is reduced and/or enlarged according to different sizes to obtain at least one sub-image; the feature information of the first target is extracted from each sub-image in the at least one sub-image to obtain at least one second Characteristic information, the at least one second characteristic information includes the first characteristic information; then the at least one second characteristic information is added to the invalid alarm database.

In the above design, since the feature information of the first target included in images of different sizes may be different, the area including the first target may be reduced according to different reduced sizes, enlarged according to different enlarged sizes, or according to Different reduction sizes and different enlargement sizes are respectively reduced and enlarged to obtain at least one sub-image, and then the feature information obtained based on each sub-image in the at least one sub-image is added to the invalid alarm database, which can enrich the first target To increase the diversity of feature information, the accuracy of automatic judgment of invalid alarms can be improved.

In a possible design of the first aspect, acquiring the first feature information of the first target includes: receiving multimedia information from the camera; when it is detected that there is a moving first target in the shooting area of the camera, and the first target The behavior of is to obtain the first image from the multimedia information when an alarm is triggered, and perform feature extraction on the first target in the first image to obtain the first feature information.

In the above design, the server receives multimedia information from the camera. The multimedia information may be video data or multiple images. The server detects the multimedia information, and when it detects that there is a moving first target in the area captured by the camera, and the behavior of the first target is a behavior that triggers an alarm (such as area intrusion, cross-border or wandering, etc.), it will not include the The first feature information of the first target is extracted from the first image of the first target. That is, the behavior that triggers the alarm is recognized by the server without the need for the camera to recognize, which can reduce the load of the camera.

In a possible design of the first aspect, acquiring the first characteristic information of the first target includes: receiving the first characteristic information from the camera.

In the above design, the server receives the first feature information from the camera. The first feature information is the first target that the camera detects that there is motion in its shooting area, and the behavior of the first target is a behavior that triggers an alarm. It is sent after extracting the first characteristic information of the first target from the first image including the first target. That is, the behavior that triggers the alarm is recognized by the camera without sending multimedia information to the server in real time to identify the behavior that triggers the alarm, which can reduce the load on the server and save network resources.

In a possible design of the first aspect, when the type of the alarm triggered by the first target is the first alarm type, matching the first characteristic information with the characteristic information in the invalid alarm database includes: according to the first alarm Type, determine the third characteristic information in the invalid alarm database, the invalid alarm triggered by the target corresponding to the third characteristic information is an invalid alarm of the first alarm type; the first characteristic information is matched with the third characteristic information.

In the above design, since the behaviors of targets corresponding to different alarm types have large differences, the difference between the characteristic information corresponding to different alarm types is also large. Therefore, when determining whether to filter the alarm information of the first target, there is no need to match the first feature information with all the feature information included in the invalid alarm database. The characteristic information can be matched with the third characteristic information corresponding to the first alarm type, thereby reducing invalid matching, reducing calculation amount, and improving matching efficiency.

In the second aspect, embodiments of the present application provide a device for reducing invalid alarms. The device may be a camera, or a chip or chip system in a camera, or a server, or a chip or chip system in a server. The device includes processing Unit and transceiver unit. in,

A processing unit, configured to obtain first characteristic information of a first target, the behavior of the first target is a behavior that triggers a camera alarm; matching the first characteristic information with the characteristic information in the invalid alarm database, the invalid The alarm database includes characteristic information of at least one target, and the behavior of each target in the at least one target is a behavior that triggers an invalid alarm;

The transceiver unit is configured to not output alarm information when the degree of matching between the first characteristic information and the characteristic information in the invalid alarm database is greater than or equal to a first threshold.

In a possible design of the second aspect, the transceiver unit is further configured to:

When the degree of matching between the first feature information and the feature information in the invalid alarm database is less than a first threshold, receive first information from an input interface, where the first information is used to indicate the first target Whether the triggered alarm is an invalid alarm;

When the first information is used to indicate that the alarm triggered by the first target is an invalid alarm, no alarm information is output, and the processing unit is further configured to add the first characteristic information to the invalid alarm database, Otherwise, the transceiver unit is used to output alarm information.

In a possible design of the second aspect, the processing unit is specifically configured to:

Acquiring a first image, the first image including the first target;

Determine the area of the first target in the first image;

Reduce and/or enlarge the area according to different sizes to obtain at least one sub-image;

Extracting characteristic information of the first target from each of the at least one sub-image to obtain at least one second characteristic information, where the at least one second characteristic information includes the first characteristic information;

The at least one second characteristic information is added to the invalid alarm database.

Receiving multimedia information from the camera;

The processing unit is specifically used for:

When it is detected that there is a moving first target in the shooting area of the camera, and the behavior of the first target is a behavior that triggers an alarm, a first image is obtained from the multimedia information, and the Feature extraction is performed on a first target in an image to obtain the first feature information.

Receiving the first characteristic information from the camera.

In a possible design of the second aspect, when the type of the alarm triggered by the first target is the first alarm type, the processing unit is specifically configured to:

Determine the third characteristic information in the invalid alarm database according to the first alarm type, and the invalid alarm triggered by the target corresponding to the third characteristic information is the invalid alarm of the first alarm type;

Matching the first characteristic information with the third characteristic information.

In a third aspect, an embodiment of the present application provides a device for reducing invalid alarms. The device for reducing invalid alarms includes at least one processor; the at least one processor is configured to run a computer program or instruction to enable the device to execute The method described in the first aspect above.

In a possible design of the third aspect, the at least one processor executes the following steps when running a computer program or instruction:

Acquiring first characteristic information of a first target, where the behavior of the first target is a behavior that triggers a camera alarm;

The first characteristic information is matched with the characteristic information in an invalid alarm database, the invalid alarm database includes characteristic information of at least one target, and the behavior of each target in the at least one target is a behavior that triggers an invalid alarm ；

When the degree of matching between the first characteristic information and the characteristic information in the invalid alarm database is greater than or equal to a first threshold, no alarm information is output.

In a possible design of the third aspect, the device for reducing invalid alarms further includes a memory for storing computer programs or instructions.

In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, the computer-readable storage medium stores a computer program, and the computer program includes program instructions that, when executed by a computer, cause the The computer executes the method described in the first aspect or any one of the possible designs of the first aspect.

In a fifth aspect, an embodiment of the present application provides a computer program product, the computer program product stores a computer program, the computer program includes program instructions, and when executed by a computer, the program instructions cause the computer to execute the first Aspect or the method of any one of the possible designs of the first aspect.

In a sixth aspect, the present application provides a chip system including a processor for implementing the method described in the first aspect or any one of the possible designs of the first aspect. The chip system can be composed of chips, and can also include chips and other discrete devices.

In a seventh aspect, the present application provides an alarm signal processing device, which includes an interface and a processor. Wherein, the interface is used to receive the first feature information sent by the camera or the first image corresponding to the first feature information, and the processor communicates with the interface, and is used to implement the first aspect or any one of the possible designs of the first aspect. The method described.

For the beneficial effects of the aforementioned second aspect to the seventh aspect and possible implementation manners, reference may be made to the description of the beneficial effects of the first aspect and possible implementation manners.

Description of the drawings

FIG. 1 is a schematic diagram of the architecture of a security monitoring system to which an embodiment of the application is applicable;

FIG. 2 is a schematic flowchart of a method for reducing invalid alarms according to an embodiment of the application;

FIG. 3 is a schematic diagram of a display interface provided by an embodiment of this application;

FIG. 4 is a schematic flowchart of another method for reducing invalid alarms according to an embodiment of the application;

Figure 5 is a schematic structural diagram of a device for reducing invalid alarms provided by an embodiment of the application;

Fig. 6 is a schematic structural diagram of an alarm signal processing device provided by an embodiment of the application.

Detailed ways

In order to make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the embodiments of the present application will be further described in detail below with reference to the accompanying drawings.

To facilitate the understanding of the technical personnel, the technical terms involved in the embodiments of the present application will be described first.

(1) Security area can refer to areas that require security protection, such as railways in unmanned areas, fences in parks, or display cabinets where anti-theft objects are located. If an abnormal intrusion situation in a security area needs to be detected, one or more collection devices, such as cameras, can be set around the security area. The collection range of the one or more collection devices can cover the security area, for example, the security area can be part of the collection area covered by the one or more collection devices, or it can also be covered by the one or more collection devices All collection areas.

(2) Security targets can refer to objects that need to be guarded (or need to be monitored), such as people, animals, or vehicles. If the security target enters the security area, an alarm can be triggered. The corresponding security targets can be different in different security scenarios. For example, the security purpose of the railway in the unmanned land is mainly to prevent the surrounding wild animals from wandering on the railway causing accidents such as sudden train braking or derailment. Therefore, the security target in the security scene may be wild animals. For another example, the security purpose of a fence in a zoo is mainly to prevent animals inside the fence from jumping over the fence and causing injury to tourists, and tourists entering the fence illegally, so the security targets in the security scene can be people and animals.

(3) In the embodiments of this application, "multiple" refers to two or more than two. In view of this, "multiple" can also be understood as "at least two" in the embodiments of this application. "At least one" can be understood as one or more, for example, one, two or more. For example, including at least one refers to including one, two or more, and does not limit which ones are included. For example, including at least one of A, B, and C, then the included can be A, B, C, A and B, A and C, B and C, or A and B and C. "And/or" describes the association relationship of the associated objects, indicating that there can be three types of relationships, for example, A and/or B, which can mean: A alone exists, A and B exist at the same time, and B exists alone. In addition, the character "/", unless otherwise specified, generally indicates that the associated objects before and after are in an "or" relationship.

Unless otherwise stated, ordinal numbers such as “first” and “second” mentioned in the embodiments of the present application are used to distinguish multiple objects, and are not used to limit the order, timing, priority, or importance of multiple objects.

In the following, the technical features involved in the embodiments of the present application will be introduced.

Traditional security measures generally monitor the security area by patrolling multiple workers, and manual maintenance costs a lot. Some security scenarios with harsher environmental conditions are not friendly to human health or life. Therefore, in these security scenarios, traditional security measures are no longer applicable, such as security scenarios with high ambient temperatures or railway security in unmanned areas. With the rapid development of image processing technology, especially the development of image recognition technology, the abnormal intrusion detection technology based on image processing technology has gradually replaced the transmission security measures due to its advantages such as automation and low requirements for environmental conditions.

In the current abnormal intrusion detection technology, at least one camera needs to be deployed around the security area, so that the area that needs to be secured falls into the shooting area of the at least one camera in whole or in part. Then, image recognition is performed on the picture captured by the at least one camera to detect whether there is an abnormal intrusion in the security area in real time. When it is recognized that there is a moving target in the security area, and the behavior of the target is a preset behavior that triggers an alarm, an alarm is triggered, and the alarm information of the target is reported to the staff for processing. The warning information may include an image of the target. After receiving the alarm information, the staff can judge whether the alarm triggered by the target is an invalid alarm according to the image. If the staff judges that the alarm triggered by the target is an invalid alarm, then delete the alarm information; if the staff judges that the alarm triggered by the target is a valid alarm, then respond to the alarm information.

Due to the diversity of the environment in the security area, invalid alarms may be triggered. For example, the preset behavior that triggers the alarm is the wandering of a person, and the plants around the security area are swayed by the wind, and are mistakenly identified as the wandering of a person, which triggers the alarm, but in fact the plants do not belong to the security area. Security targets, so the staff will receive invalid warnings. At the same time, plants generally stay in the picture taken by the camera for a long time, and the alarm will be triggered repeatedly in windy weather. Accordingly, the staff will receive multiple invalid alarms. The staff have to screen each alarm and confirm that it is a false alarm. (Alarms caused by plant shaking) is still correct (alarms caused by security targets), which increases the cost of manual maintenance, and the user experience is poor. In addition, when multiple cameras deployed around the security area have the target in their shooting screens, each camera will alert the staff individually, so the staff will also receive information from each of the multiple cameras. Invalid alarms, the staff needs to confirm and delete each invalid alarm one by one, which increases the cost of manual maintenance, and the user experience is poor.

In view of this, the embodiment of the present application provides a method for reducing invalid alarms. In the method, the characteristic information of the target that triggers the alarm is matched with the characteristic information in the invalid alarm database. When the matching degree between the two is greater than or equal to the first When a threshold is reached, the alarm information triggered by the target is filtered, thereby reducing the overhead of manual maintenance and improving user experience.

The method provided by the embodiment of the present application will be described below in conjunction with the accompanying drawings.

Please refer to FIG. 1. FIG. 1 is a schematic structural diagram of a possible security monitoring system to which an embodiment of this application is adapted. As shown in FIG. 1, the security monitoring system 100 may include multiple image acquisition devices 110 (including two image acquisition devices in FIG. 1) and a server 120. The image collection device 110 is deployed around the security area, and is used to collect images of the security area in real time and send the collected images to the server 120. The server 120 is configured to receive the picture taken by the image acquisition device, and recognize the taken picture to determine whether to trigger an alarm. Wherein, multiple image acquisition devices 110 may be deployed around the same security area, or may be deployed in different security areas, which is not limited in the embodiment of the present application.

The image acquisition device 110 includes, but is not limited to, a camera, a video camera, or a smart mobile terminal with a camera function (such as a mobile phone, a tablet computer, a portable computer), etc. The specific implementation form of the image acquisition device 110 is not limited herein. In the following, the image acquisition device 110 is a camera 110 as an example for introduction. The server 120 includes, but is not limited to, an independent server, a server cluster, or a cloud server, etc. The specific implementation form of the server 120 is not limited herein.

Optionally, the security monitoring system 100 may further include a display 130, which is used to display an image including a target that triggers an alarm to the staff. The staff can also perform input operations on the input interface (or called the display interface) of the display 130, such as deleting invalid alarm information.

In a possible implementation manner, the server 120 includes a processor 121 that can recognize the image to determine whether there is an abnormal intrusion in the security area. Optionally, the processor 121 may also extract feature information of the target that triggers the alarm from the image. For example, the processor 121 may be implemented in one or more of the following forms: a general-purpose processor, a central processing unit (CPU), an image signal processor (ISP), a microprocessor, and a digital signal processor. Processor (digital signal processor, DSP), or field-programmable gate array (field-programmable gate array, FPGA), etc.

Exemplarily, the server 120 may further include a memory 122. The memory 122 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, random access memory (RAM), or other types that can store information and instructions The dynamic storage device can also be electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, EEPROM), CD-ROM (Compact Disc Read-Only Memory, CD-ROM) or other optical disc storage, optical disc storage (Including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or can be used to carry or store desired program codes in the form of instructions or data structures and can be used by the Any other medium accessed by the device, but not limited to this. The memory 122 may exist independently and is connected to the processor 121 through a bus. The memory 122 may also be integrated with the processor 121.

Wherein, the memory 122 may be used to store the application program code for executing the solution of the present application, and the processor 121 controls the execution, that is, the processor 121 is used to execute the application program code stored in the memory 122 The method for reducing invalid alarms in the embodiments of the present application is implemented. In some embodiments, the processor 121 may also have the ability to store code inside, and the stored code can be executed without using the memory 122.

In another possible implementation manner, the camera 110 may recognize the captured image to determine whether to send alarm information to the server 120. In this embodiment, the camera 110 may include the aforementioned processor 120, and optionally, the camera 110 may further include the aforementioned memory 122. For the specific implementation of the processor 121 and the memory 122, please refer to the foregoing description, which will not be repeated here.

The embodiments of this application can be applied to scenes that require security, for example, indoor video surveillance scenes, outdoor video surveillance scenes, night video surveillance scenes or security surveillance scenes, or can also be applied to common image shooting scenes, etc., which vary here. An example. The following describes a method for reducing invalid alarms provided by an embodiment of the present application in conjunction with the security monitoring system 100 shown in FIG. 1. Please refer to Figure 2, which shows a flowchart of this method.

S201: The server 120 receives multimedia information from the camera 110.

The camera 110 captures images in its shooting area, obtains multimedia information, and sends the multimedia information to the server 120. The server 120 receives multimedia information. The multimedia information may be video data. For example, the camera 110 compresses a picture taken within a set time period into video data in the H.264/H.265 format, and then sends the video data to the server 120. Alternatively, the multimedia information may also be image data. For example, the camera 110 extracts multiple images from the images captured by the camera 110 within a set time period, and then compresses the multiple images into image data in JPEG format, and then The image data is sent to the server 120. In addition, part or all of the shooting area of the camera 110 is a security area. For example, in an indoor surveillance scene, all areas in the room are security areas, and the camera 110 needs to be deployed in the room, so that the shooting area of the camera 110 is all or part of the area in the room, that is, the shooting area of the camera 110 All areas in are security areas. For another example, in a park surveillance scene, the wall of the park is a security area, and the camera 110 needs to be deployed around the wall, so that part of the picture in the shooting area of the camera 110 is the picture of the wall, that is, in the shooting area of the camera 110 Part of the area is a security area.

S202: The server 120 detects the media information, and if it detects that there is a moving first target in the shooting area of the camera 110 and the behavior of the first target is a behavior that triggers an alarm, perform S203; otherwise, perform S201.

For example, after the server 120 receives the multimedia information from the camera 110, it needs to detect the multimedia information to determine whether there is an abnormal intrusion in the shooting area of the camera 110. Specifically, the server 120 may first determine whether there is a moving target in the shooting area according to the multimedia information (here, the detected moving target is called the first target), and then determine whether the behavior of the first target is triggering an alarm. behavior. For example, if the multimedia information is video data, the server 120 may decompress the video data, extract at least one image from the decompressed video data, and then detect the at least one image through a moving target detection method to determine the shooting area Whether there is the first goal of the movement. For another example, if the multimedia information is image data, the server 120 may decompress the image data to obtain at least one image, and then detect the at least one image by a moving target detection method to determine whether there is a moving first target in the shooting area. Among them, the moving target detection method is a target detection method, and machine learning algorithms can be used.

If the server 120 determines that there is no moving target in the shooting area, the server 120 executes the step of S201, that is, receives multimedia information within the next set time period. If the server 120 determines that there is a moving first target in the shooting area, the server 120 needs to determine whether the behavior of the first target is a behavior that triggers an alarm. Optionally, the server 120 may determine whether the behavior of the first target is a behavior that triggers an alarm through an alarm recognition model. For example, the server 120 may extract the first characteristic information of the first target from the image including the first target, and input the first characteristic information into the alarm recognition model to obtain the recognition result. For another example, the server 120 may reduce the area of the image where the first target is located according to different reduced sizes, enlarge according to different enlarged sizes, or reduce and enlarge according to different reduced sizes and different enlarged sizes, respectively, to obtain at least one For sub-images, the feature information of the first target is extracted from each sub-image in the at least one sub-image to obtain at least one second feature information, and then the at least one second feature information is input into the alarm recognition model to obtain a recognition result. If the recognition result is greater than or equal to the preset second threshold, the server 120 determines that the behavior of the first target is a behavior that triggers an alarm; otherwise, the behavior of the first target is not a behavior that triggers an alarm. Among them, the behavior that triggers the alarm can be a behavior that is preset according to different security scenarios. For example, the wandering of wild animals on the railway in the unmanned area is the behavior that triggers the alarm, and the behavior of people over the wall in the park security is the behavior that triggers the alarm. Wait. The alarm recognition model is obtained by the server 120 performing model training on the feature information corresponding to the preset alarm-triggering behavior. The corresponding model training process can refer to the prior art, which is not limited in the embodiment of the application.

Among them, the method of extracting the feature information of the target from the image can include but is not limited to the feature extraction method based on the histogram of oriented gradient (HOG), and the feature based on the local binary pattern (LBP) Extraction methods, feature extraction methods based on deep neural networks (DNN) or feature extraction methods based on convolutional neural networks (convolutional neural networks, CNN), etc.

It should be understood that there may be multiple moving targets in the shooting area. When there are multiple moving targets in the shooting area, the server 120 may determine, one by one, whether the behavior of each target in the multiple running targets is a behavior that triggers an alarm through an alarm recognition model. If the behavior of one of the multiple running targets is a behavior that triggers an alarm, the server 120 executes S203. In the following description, the first moving target in the shooting area, that is, the first target, is taken as an example for description.

If the server 120 determines that the behavior of the first target is not the behavior that triggers the alarm, the server 120 executes the step of S201, that is, receives the multimedia information within the next set time period. If the server 120 determines that the behavior of the first target is a behavior that triggers an alarm, the server 120 executes S203, that is, acquires a first image including the first target.

It should be noted that the abnormal intrusion detection in step S202 can be implemented by using the above-mentioned method of first detecting a moving target and then recognizing a target behavior, or can be implemented by using existing technology, which is not limited in the embodiment of the present application.

S203: The server 120 obtains a first image from the multimedia information, where the first image includes the first target.

The multimedia information may include multiple images, each of the multiple images includes the first target, and the server 120 may select an image from the multiple images as the first image. For example, the server 120 may select the image with the most feature information of the first target among the plurality of images as the first image, or select the image with the earliest shooting time among the plurality of images as the first image, or select the image from among the plurality of images. The image with the highest resolution is used as the first image, etc., which is not limited in the embodiment of the present application.

S204: The server 120 performs feature extraction on the first target in the first image to obtain first feature information.

In a possible implementation, if the server 120 has extracted the characteristic information of the first target in step S202, the server 120 may not need to perform step S204, that is, the characteristic information extracted by the server 120 in step S202 may be As the first feature information, the amount of calculation can be reduced. Otherwise, the server 120 extracts the first feature information of the first target from the first image. It should be understood that, even if the server 120 has extracted the characteristic information of the first target in step S202, the server 120 may also perform step S204. For example, in step S202, the server 120 may extract the feature information of the first target according to the HOG-based feature extraction method, and in step S204, the server 120 may extract the feature information of the first target according to the CNN-based feature extraction method.

In another possible implementation manner, the server 120 may determine the area of the first target in the first image, and reduce, or enlarge, or reduce and enlarge the area according to different sizes to obtain at least one sub-image, and then At least one piece of second feature information is obtained from the feature information of the first target extracted from each of the at least one sub-image, and the at least one piece of second feature information may include the first feature information. Since the feature information of the first target included in images of different sizes may be different, the area including the first target may be reduced in different reduced sizes, enlarged in accordance with different enlarged sizes, or in accordance with different reduced sizes and differences. The zoomed-in size of is reduced and enlarged to obtain at least one sub-image, and then the feature information of the first target is extracted from each sub-image in the at least one sub-image, so that the feature information of the first target can be enriched and the diversity of feature information is increased .

S205: The server 120 matches the first characteristic information with the characteristic information in the invalid alarm database. If the degree of matching between the first characteristic information and the characteristic information in the invalid alarm database is greater than or equal to the first threshold, execute S206; if the degree of matching between the first characteristic information and the characteristic information in the invalid alarm database is less than For the first threshold, S207 is executed.

Wherein, the invalid alarm database stores characteristic information of at least one target, and the behavior of each target in the at least one target is a behavior that triggers an invalid alarm. For example, the characteristic information in the invalid alarm database may include the characteristic information of the target that easily triggers the invalid alarm obtained in advance according to the security scene, or the characteristic information in the invalid alarm database may include the characteristic information of the target that triggers the invalid alarm in history, or The feature information in the invalid alarm database includes feature information of targets that are easy to trigger invalid alarms obtained in advance according to security scenarios, and feature information of targets that trigger invalid alarms in history. Optionally, the invalid alarm database may also include at least one of information such as an alarm type, a security target, a security area, or an identity (ID) of the camera. Among them, the alarm type may include, but is not limited to, human cross-border behavior, human wandering behavior, human over-wall behavior, animal cross-border behavior, or vehicle cross-border behavior, etc. Optionally, the feature information of a target in the invalid alarm database can be obtained by performing feature extraction on sub-images of different sizes of the region in the image where the target is located, so as to increase invalidity. For the diversity of feature information in the alarm database, the specific extraction process can be referred to the foregoing description, which will not be repeated here.

The server 120 may match the first feature information with the feature information in the invalid alarm database, or match at least one second feature information with the feature information in the invalid alarm database, to determine whether the behavior of the first target is invalid triggering The behavior of the alert. Taking the matching of the first feature information with the feature information in the invalid alarm database as an example, the server 120 may match the first feature information with all the feature information in the invalid alarm database one by one, although the accuracy of the matching result can be guaranteed. However, when there are many feature information in the invalid alarm database, the calculation amount of the server 120 will be greatly increased. Optionally, the server 120 may match the alarm type, security area, or security scenario as consideration factors to eliminate partial invalid matches, reduce the amount of calculation, and improve matching efficiency. For example, the type of the alarm triggered by the first target is the first alarm type (such as a person's behavior of overturning the wall), and the server 120 may obtain the third characteristic information from the invalid alarm database according to the first alarm type, and the third characteristic finely corresponds to The invalid alarm triggered by the target of is the invalid alarm of the first alarm type, and then the first characteristic information is matched with the third characteristic information.

When the server 120 determines that the degree of matching between the first characteristic information and a characteristic information in the invalid alarm data is greater than or equal to the first threshold, the server 120 may determine that the alarm triggered by the first target is an invalid alarm, that is, execute step S206; Or when the server 120 determines that the average matching degree between the first characteristic information and the set number (such as 3) of the characteristic information in the invalid alarm data is greater than or equal to the first threshold, the server 120 may determine that the first target triggers If the alarm is an invalid alarm, step S206 is executed. When the server 120 determines that the degree of matching between the first feature information and any feature information in the invalid alarm data is less than the first threshold, the server 120 may send the first image to the display 130 for display, that is, step S207 is performed.

S206: The server 120 determines that the alarm triggered by the first target is an invalid alarm, and does not output alarm information.

Since the invalid alarm database stores the characteristic information of the target that previously triggered the invalid alarm, and/or the characteristic information of the target that easily triggers the invalid alarm, when the first characteristic information matches the characteristic information in the invalid alarm database When the degree is high (for example, greater than or equal to the first threshold), it indicates that the alarm triggered by the first target is an invalid alarm. Therefore, the alarm information triggered by the first target can be automatically filtered, that is, no alarm information is output, thereby saving manual detection links, reducing manual maintenance costs, and improving the work efficiency and user experience of the alarm process.

S207: The server 120 sends the first image to the display 130 for display.

When the degree of matching between the first feature information and the feature information in the invalid alarm database is low (for example, less than the first threshold), the server 120 sends the first image to the display 130 for display, and the display 130 uses the first image to display the first image. It is displayed on the display interface (or called the input interface), so that the staff can judge whether the alarm triggered by the first target is an invalid alarm according to the first image. In order to improve the efficiency of the alarm process, while the server 120 sends the first image to the display 130 for display, it can also remind the staff to judge the first image displayed on the current display interface through voice, warning lights or animation.

For example, FIG. 3 is a schematic diagram of a display interface. As shown in FIG. 3, the display interface 300 may include a first image 310, an invalid warning option 320, a valid warning option 330, and a second information area 340. Wherein, the first image 310 includes the first target 311. Optionally, a focus frame may be used to clarify the area of the first target 311 in the first image 310. The second information area 340 is used to display second information, which is used to assist the staff in determining whether the alarm triggered by the first target in the first image is an invalid alarm. For example, the second information may include but is not limited to The ID of the camera, the address information of the security zone, the alarm type or time, etc. It should be understood that the position of each option in the display interface shown in FIG. 3 is only an example, which is not limited in the embodiment of the present application. For example, the second information area 340 in FIG. 3 is located on the left of the first image 310, but in fact the second information area 340 may be located above, below, or to the left of the first image 310. When the staff member determines that the alarm triggered by the first target is a valid alarm, the staff member can select the valid alarm option 330 through input methods such as keyboard input, mouse input, or voice input to execute step 210. When the staff member determines that the alarm triggered by the first target is an invalid alarm, the staff member can select the invalid alarm option 320 through input methods such as keyboard input, mouse input, or voice input to execute the content executed in step S209.

S208: The server 120 receives first information from the touch screen display 130, where the first information is used to indicate whether the alarm triggered by the first target is an invalid alarm.

Of course, in addition to the display 130, the first information may also come from other input devices such as a keyboard and a mouse.

For example, the first information may be information triggered by the invalid alarm option 320 or information triggered by the valid alarm option 330. When the first information is information triggered by the invalid alarm option 320, the first information is used to indicate that the alarm triggered by the first target is an invalid alarm. When the first information is the information triggered by the valid alarm option 330, the first information is used to indicate that the alarm triggered by the first target is a valid alarm.

S209: When the first information indicates that the alarm triggered by the first target is an invalid alarm, the server 120 does not output the alarm information, and adds the first characteristic information to the invalid alarm database.

When the first information indicates that the alarm triggered by the first target is an invalid alarm, the server 120 filters the alarm information of the first target, that is, does not output the alarm information, and the server 120 may also add the first characteristic information to the invalid alarm database. In this way, repeated alarms triggered by the first target can be filtered, and manual detection links can be eliminated, manual maintenance costs can be reduced, and user performance can be improved.

Optionally, if in step S202 or step S204, the server 120 has acquired at least one second characteristic information of the first target, the server 120 may add the at least one second characteristic information to the invalid alarm database in step S209 , To enrich the characteristic information of the first target and increase the diversity of the invalid alarm database, thereby improving the accuracy of judging invalid alarms. If in step S202 or step S204, the server 120 does not obtain at least one second feature information of the first target, then in step S209, the server 120 may reduce the area of the first target in the first image according to different sizes, Or zoom in, or reduce and zoom in to obtain at least one sub-image, obtain feature information of the first target from each of the at least one sub-image to obtain at least one second feature information, and then add the at least one second feature information In the invalid alarm database, the feature information of the first target is enriched, and the diversity of the invalid alarm database is increased, so that the accuracy of judging invalid alarms can be improved.

S210: When the first information indicates that the alarm triggered by the first target is a valid alarm, the server 120 outputs alarm information.

When the first information indicates that the alarm triggered by the first target is an invalid alarm, the server 120 outputs the alarm information. The form, or content, or form and content of the alarm information output in different security scenarios may be different, and the ways of responding to the alarm information may also be different. For example, when it is determined that an animal is wandering on a railway in a no-man’s land, the alarm information may include, but is not limited to, the type of alarm, the location information that triggered the alarm, the alarm level, the security target or the number of security targets, etc., and the server 120 outputs the alarm information At the same time, it can also send warning information to the train at a set distance from the warning position to remind the train to decelerate in advance or whistle. Among them, the alarm level is used to indicate the urgency corresponding to the alarm triggered by the target. Generally, the higher the alarm level, the higher the urgency. For another example, when it is determined that someone has crossed the border and entered the security area, the alarm information may include, but is not limited to, the type of alarm, the location information or the alarm level that triggered the alarm, etc. While the server 120 outputs the alarm information, it can also use the speaker near the alarm location. Give voice warnings, or notify relevant security personnel to give manual warnings.

In the foregoing embodiment of the present application, the server 120 receives multimedia information from the camera 110, and the multimedia information may be video data or multiple images. The server 120 detects the multimedia information, and when it detects that there is a moving first target in the shooting area of the camera 110, and the behavior of the first target is a behavior that triggers an alarm (such as area intrusion, cross-border or wandering, etc.), The first feature information of the first target is extracted from the first image including the first target, that is, the behavior of triggering an alarm is recognized by the server 120 without the participation of the camera 110, which can reduce the amount of calculation of the camera 110. Further, the server 120 matches the first characteristic information with the characteristic information in the invalid alarm database. When the matching degree between the two is high (for example, greater than or equal to the first threshold), it indicates that the alarm triggered by the first target is an invalid alarm, and the server 120 can automatically filter the alarm information triggered by the first target, that is, No alarm information is output, thereby saving manual detection links, reducing manual maintenance costs, and improving the efficiency of the alarm process and user experience. When the matching degree between the two is low (for example, less than the first threshold), the server 120 may display the first image including the first target, so that the staff can judge whether the alarm triggered by the first target is based on the first image. It is an invalid alarm. When the staff judges that the alarm triggered by the first target is an invalid alarm based on the first image, the server 120 does not output the alarm information, and at the same time adds the first characteristic information to the invalid alarm database to increase the diversity of samples in the invalid alarm database . When the first target triggers an alarm again, the server 120 can automatically filter the alarm information triggered by the first target without requiring staff to check again, which can reduce the probability of generating an invalid alarm, improve the accuracy of the alarm, and improve user experience.

In the foregoing embodiments of the present application, in order to implement the functions in the methods provided in the foregoing embodiments of the present application, the server 120 may include a hardware structure and/or a software module. Form to achieve the above functions. Whether a certain function among the above-mentioned functions is executed by a hardware structure, a software module, or a hardware structure plus a software module depends on the specific application and design constraint conditions of the technical solution.

In the following, in conjunction with the security monitoring system 100 shown in FIG. 1, another method for reducing invalid alarms provided by an embodiment of the present application is introduced. Please refer to FIG. 4, which is a schematic flowchart of another method for reducing invalid alarms according to an embodiment of the application.

S401: The server 120 receives an alarm request from the camera 110, where the alarm request includes a first image and first feature information.

Wherein, the first image includes a first target, the behavior of the first target is a behavior that triggers an alarm, and the first characteristic information is characteristic information of the first target.

The camera 110 photographs the picture in its shooting area, and at the same time detects the picture in its shooting area to determine whether there is a moving target in its shooting area (here, the detected moving target is called the first target), And the behavior of the first target is a behavior that triggers an alarm. Specifically, the camera 110 may extract at least one image from a shooting frame of a set duration, and detect the at least one image by a moving target detection method to determine whether there is a moving first target in the shooting area. Wherein, part or all of the shooting area of the camera 110 is a security area. For example, in an indoor surveillance scene, all areas in the room are security areas, and the camera 110 needs to be deployed in the room, so that the shooting area of the camera 110 is all or part of the area in the room, that is, the shooting area of the camera 110 All areas in are security areas. For another example, in a park surveillance scene, the wall of the park is a security area, and the camera 110 needs to be deployed around the wall, so that part of the picture in the shooting area of the camera 110 is the picture of the wall, that is, in the shooting area of the camera 110 Part of the area is a security area. Among them, the moving target detection method is a target detection method, and machine learning algorithms can be used.

When the camera 110 detects that there is a moving first target in its shooting area, the camera 110 then determines whether the behavior of the first target is a behavior that triggers an alarm. Optionally, the camera 110 may determine whether the behavior of the first target is a behavior that triggers an alarm through an alarm recognition model. For example, the camera 110 may extract first characteristic information of the first target from an image including the first target, and input the first characteristic information into an alarm recognition model to obtain a recognition result. For another example, the camera 100 may reduce the area of the image where the first target is located according to different reduced sizes, enlarge according to different enlarged sizes, or reduce and enlarge according to different reduced sizes and different enlarged sizes, respectively, to obtain at least one For sub-images, the feature information of the first target is extracted from each sub-image of the at least one sub-image to obtain at least one second feature information, and then the at least one second feature information is input into the alarm recognition model to obtain a recognition result. If the recognition result is greater than or equal to the preset second threshold, the camera 110 determines that the behavior of the first target is a behavior that triggers an alarm; otherwise, the behavior of the first target is not a behavior that triggers an alarm. Among them, the behavior that triggers the alarm can be a behavior that is preset according to different security scenarios. For example, the wandering of wild animals on the railway in the unmanned area is the behavior that triggers the alarm, and the behavior of people over the wall in the park security is the behavior that triggers the alarm. Wait. The alarm recognition model is obtained by the camera 110 performing model training on the feature information corresponding to the preset alarm-triggering behavior. The corresponding model training process can refer to the prior art, which is not limited in the embodiment of the application.

When the camera 110 detects that there is a moving first target in its shooting area, and the behavior of the first target triggers an alert behavior, the camera 110 can record the first feature information of the first target or at least one second target of the first target. The characteristic information is sent to the server 120. The server 120 receives the first characteristic information or at least one second characteristic information. Optionally, the camera 110 may also send the first image including the first target to the server 120. For example, the camera 110 may select the image with the most feature information of the first target among the multiple images as the first image and send it to the server 120, or select the image with the earliest shooting time among the multiple images and send it to the server 120 as the first image. , Or select the image with the highest resolution among the multiple images as the first image and send it to the server 120, etc., which is not limited in the embodiment of the present application.

S402: The server 120 matches the first characteristic information with the characteristic information in the invalid alarm database. If the degree of matching between the first characteristic information and the characteristic information in the invalid alarm database is greater than or equal to the first threshold, execute S403; if the degree of matching between the first characteristic information and the characteristic information in the invalid alarm database is less than For the first threshold, S404 is executed.

Wherein, the invalid alarm database stores characteristic information of at least one target, and the behavior of each target in the at least one target is a behavior that triggers an invalid alarm. For example, the characteristic information in the invalid alarm database may include the characteristic information of the target that easily triggers the invalid alarm obtained in advance according to the security scene, or the characteristic information in the invalid alarm database may include the characteristic information of the target that triggers the invalid alarm in history, or The feature information in the invalid alarm database includes feature information of targets that are easy to trigger invalid alarms obtained in advance according to security scenarios, and feature information of targets that trigger invalid alarms in history. Optionally, the invalid alarm database may also include at least one of information such as an alarm type, a security target, a security area, or the identity of a camera. Among them, the alarm type may include, but is not limited to, human cross-border behavior, human wandering behavior, human over-wall behavior, animal cross-border behavior, or vehicle cross-border behavior, etc. Optionally, the feature information of a target in the invalid alarm database can be obtained by performing feature extraction on sub-images of different sizes of the region in the image where the target is located, so as to increase invalidity. For the diversity of feature information in the alarm database, the specific extraction process can be referred to the foregoing description, which will not be repeated here.

The server 120 may match the first feature information with the feature information in the invalid alarm database, or match at least one second feature information with the feature information in the invalid alarm database, to determine whether the behavior of the first target is invalid triggering The behavior of the alert. Taking the matching of the first feature information with the feature information in the invalid alarm database as an example, the server 120 may match the first feature information with all the feature information in the invalid alarm database one by one, although the accuracy of the matching result can be guaranteed. However, when there are many feature information in the invalid alarm database, the calculation amount of the server 120 will be greatly increased. Optionally, the server 120 may match the alarm type, security area, or security scenario as consideration factors to eliminate partial invalid matches, reduce the amount of calculation, and improve matching efficiency. For example, the type of the alarm triggered by the first target is the first alarm type (such as the behavior of a person overturning the wall), and the server 120 may obtain the third characteristic information from the invalid alarm database according to the first alarm type, and the third characteristic finely corresponds to The invalid alarm triggered by the target of is the invalid alarm of the first alarm type, and then the first characteristic information is matched with the third characteristic information.

When the server 120 determines that the degree of matching between the first characteristic information and a characteristic information in the invalid alarm data is greater than or equal to the first threshold, the server 120 may determine that the alarm triggered by the first target is an invalid alarm, that is, execute step S403; Or when the server 120 determines that the average matching degree between the first characteristic information and the set number (such as 3) of the characteristic information in the invalid alarm data is greater than or equal to the first threshold, the server 120 may determine that the first target triggers If the alarm is an invalid alarm, step S403 is executed. When the server 120 determines that the degree of matching between the first feature information and any feature information in the invalid alarm data is less than the first threshold, the server 120 may send the first image to the display 130 for display, that is, step S404 is performed.

S403: The server 120 determines that the alarm triggered by the first target is an invalid alarm, and does not output alarm information. This step can also be understood as terminating the alarm request sent by the camera.

Since the invalid alarm database stores the characteristic information of the target that previously triggered the invalid alarm, and/or the characteristic information of the target that easily triggers the invalid alarm, when the first characteristic information matches the characteristic information in the invalid alarm database When the degree is high (for example, greater than or equal to the first threshold), it indicates that the alarm triggered by the first target is an invalid alarm. Therefore, the alarm request sent by the camera can be automatically filtered, and the alarm information is not output, thereby saving manual detection links, reducing manual maintenance costs, and improving the efficiency of the alarm process and user experience.

S404: The server 120 sends the first image to the display 130 for display.

When the degree of matching between the first feature information and the feature information in the invalid alarm database is low (for example, less than the first threshold), the server 120 sends the first image to the display 130 for display, and the display 130 uses the first image to display the first image. It is displayed on the display interface (or called the input interface), so that the staff can judge whether the alarm triggered by the first target is an invalid alarm according to the first image. In order to improve the efficiency of the alarm process, while the server 120 sends the first image to the display 130 for display, it can also remind the staff to judge the first image displayed on the current display interface through voice, warning lights or animation. The display interface can be seen in Figure 3, which will not be repeated here.

S405: The server 120 receives the first information from the touch screen display 130, where the first information is used to indicate whether the alarm triggered by the first target is an invalid alarm.

S406: When the first information indicates that the alarm triggered by the first target is an invalid alarm, the server 120 does not output the alarm information, and adds the first characteristic information to the invalid alarm database.

When the first information indicates that the alarm triggered by the first target is an invalid alarm, the server 120 filters the alarm request from the camera and does not output the alarm information. At the same time, the server 120 can also add the first characteristic information to the invalid alarm database, so that it can filter The repetitive alarm triggered by the first target eliminates the need for manual detection, which can reduce the overhead of manual maintenance and improve user performance.

Optionally, if in step S401, the server 120 has acquired at least one second characteristic information of the first target, then in step S406, the server 120 may add the at least one second characteristic information to the invalid alarm database to enrich The feature information of the first target increases the diversity of the invalid alarm database, thereby improving the accuracy of judging invalid alarms. If in step S401, the server 120 does not obtain at least one second feature information of the first target, then in step S406, the server 120 may reduce, or enlarge, or enlarge, or enlarge, or reduce, the area of the first target in the first image according to different sizes. Or reduce and enlarge to obtain at least one sub-image, obtain characteristic information of the first target from each of the at least one sub-image to obtain at least one second characteristic information, and then add the at least one second characteristic information to the invalidity alarm In the database, the feature information of the first target is enriched, and the diversity of the invalid alarm database is increased, so that the accuracy of judging invalid alarms can be improved.

S407: When the first information indicates that the alarm triggered by the first target is a valid alarm, the server 120 outputs alarm information.

In the foregoing embodiment of the present application, the server 120 receives the first characteristic information from the camera 110, the first characteristic information is that the camera 110 detects that there is a first target moving in its shooting area, and the behavior of the first target is to trigger an alarm Is sent after extracting the first characteristic information of the first target from the first image that includes the first target. That is, the behavior that triggers the alarm is recognized by the camera 110 without sending multimedia information to the server 120 in real time to identify the behavior that triggers the alarm, which can reduce the load of the server 120 and save network resources.

Based on the same technical concept, the embodiment of the present application also provides a device for reducing invalid alarms.

FIG. 5 shows a schematic structural diagram of a device 500 for reducing invalid alarms. The device 500 for reducing invalid alarms can be used to implement the function of the server 120 in the embodiment shown in FIG. 2 or 4, that is, the function of the server 120 in the embodiment shown in FIG. 1. The device 500 for reducing invalid alarms may be a hardware structure, a software module, or a hardware structure plus a software module. The device 500 for reducing invalid alarms can be implemented by a chip or a chip system. In the embodiments of the present application, the chip system may be composed of chips, or may include chips and other discrete devices.

The apparatus 500 for reducing invalid alarms may include a processing unit 501 and a transceiver unit 502.

Wherein, the processing unit 501 can execute steps S202 to S206, step S209, and step S210 in the embodiment shown in FIG. 2, and/or other processes for supporting the technology described herein. In a possible implementation, the processing unit 501 can be used to communicate with the receiving unit 502, or the processing unit 501 can communicate with other modules of the security system 100, which can be circuits, devices, interfaces, buses, software modules, transceivers, etc. Device or any other device that can realize communication. The receiving unit 502 may perform step S201, step S207, and step S208 in the embodiment shown in FIG. 2, and/or other processes for supporting the technology described herein. In a possible implementation, the receiving unit 502 can be used to communicate with the processing unit 501, or can communicate with other modules of the security system 100, which can be circuits, devices, interfaces, buses, software modules, transceivers, or other modules. Any device that can realize communication.

Alternatively, the processing unit 501 may execute step S402, step S403, step S406, and step S407 in the embodiment shown in FIG. 4, and/or other processes for supporting the technology described herein. In a possible implementation, the processing unit 501 can be used to communicate with the receiving unit 502, or the processing unit 501 can communicate with other modules of the security system 100, which can be circuits, devices, interfaces, buses, software modules, transceivers, etc. Device or any other device that can realize communication. The receiving unit 502 may perform step S401, step S404, and step S407 in the embodiment shown in FIG. 4, and/or other processes for supporting the technology described herein. In a possible implementation, the receiving unit 502 can be used to communicate with the processing unit 501, or can communicate with other modules of the security system 100, which can be circuits, devices, interfaces, buses, software modules, transceivers, or other modules. Any device that can realize communication.

Among them, all relevant content of the steps involved in the above method embodiments can be cited in the functional description of the corresponding functional module, which will not be repeated here.

The division of modules in the embodiment shown in FIG. 5 is illustrative, and it is only a logical function division. In actual implementation, there may be other division methods. In addition, the functional modules in the various embodiments of the present application may be integrated In a processor, it may exist alone physically, or two or more modules may be integrated into one module. The above-mentioned integrated modules can be implemented in the form of hardware or software functional modules.

FIG. 6 shows an alarm signal processing device 600 provided in an embodiment of the application, where the alarm signal processing device 600 may be used to implement the function of the server 120 in the embodiment shown in FIG. 2 and/or FIG. 4. The processor 601 in the alarm signal processing device 600 may be a chip or a chip system. In the embodiments of the present application, the chip system may be composed of chips, or may include chips and other discrete devices. The alarm signal processing device 600 is, for example, an independent server, a calculator, a server cluster, or a cloud server.

The alarm signal processing device 600 includes at least one processor 601 for implementing or used for the alarm signal processing device 600 to implement the functions of the server 120 in the embodiment shown in FIG. 2 and/or FIG. 4. Exemplarily, the processor 601 may obtain the first characteristic information of the first target, and the behavior of the first target is the behavior that triggers the camera alarm; the first characteristic information is matched with the characteristic information in the invalid alarm database, and the invalid alarm database is The characteristic information of at least one target is included, and the behavior of each target in the at least one target is a behavior that triggers an invalid alarm; if the degree of matching between the first characteristic information and the characteristic information in the invalid alarm database is greater than or equal to the first threshold, No warning information is output. For details, please refer to the detailed description in the method example, which will not be repeated here.

The alarm signal processing device 600 may also include at least one memory 602 for storing program instructions and/or data. The memory 602 and the processor 601 are coupled. The coupling in the embodiments of the present application is an indirect coupling or communication connection between devices, units or modules, and may be in electrical, mechanical or other forms, and is used for information exchange between devices, units or modules. The processor 601 may operate in cooperation with the memory 602. The processor 601 may execute program instructions stored in the memory 602. At least one of the at least one memory may be included in the processor. The memory 602 may be integrated in the processor 601.

The alarm signal processing device 600 may further include an interface 603 for communicating with the processor 601 or for communicating with other devices through a transmission medium, so that the alarm signal processing device 600 can communicate with other devices. Exemplarily, the other device may be a computing module. The processor 601 may use the interface 603 to send and receive data, for example, to receive multimedia information or an alarm request sent by a camera, where the alarm request includes the first characteristic information and a first image corresponding to the first characteristic information.

The specific connection medium between the aforementioned interface 603, the processor 601, and the memory 602 is not limited in the embodiment of the present application. In the embodiment of the present application, in FIG. 6, the memory 602, the processor 601, and the interface 603 are connected by a bus 604. The bus is represented by a thick line in FIG. It is not limited. The bus can be divided into an address bus, a data bus, a control bus, and so on. For ease of representation, only one thick line is used in FIG. 6, but it does not mean that there is only one bus or one type of bus.

In the embodiment of the present application, the processor 601 may be a general-purpose processor, a digital signal processor, an application specific integrated circuit, a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component, which can implement Or execute the methods, steps, and logical block diagrams disclosed in the embodiments of the present application. The general-purpose processor may be a microprocessor or any conventional processor or the like. The steps of the method disclosed in combination with the embodiments of the present application may be directly embodied as being executed and completed by a hardware processor, or executed and completed by a combination of hardware and software modules in the processor.

In the embodiment of the present application, the memory 602 may be a non-volatile memory, such as a hard disk drive (HDD) or a solid-state drive (SSD), etc., or a volatile memory (volatile memory). For example, random access memory (RAM). The memory is any other medium that can be used to carry or store desired program codes in the form of instructions or data structures and that can be accessed by a computer, but is not limited to this. The memory 602 in the embodiment of the present application may also be a circuit or any other device capable of realizing a storage function for storing program instructions and/or data.

An embodiment of the present application also provides a computer-readable storage medium, including instructions, which when run on a computer, cause the computer to execute the method executed by the server 120 in the embodiment shown in FIG. 2 and/or FIG. 4.

The embodiment of the present application provides a chip system. The chip system includes a processor and may also include a memory for implementing the function of the server 120 in the foregoing method. The chip system can be composed of chips, and can also include chips and other discrete devices.

The embodiment of the present application provides a device for reducing invalid alarms, and the device includes the server 120 in the embodiment shown in FIG. 2 and/or FIG. 4.

The methods provided in the embodiments of the present application may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented by software, it can be implemented in the form of a computer program product in whole or in part. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on the computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, network equipment, user equipment, or other programmable devices. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from a website, computer, server, or data center. Transmission to another website, computer, server or data center via wired (such as coaxial cable, optical fiber, digital subscriber line (DSL for short) or wireless (such as infrared, wireless, microwave, etc.). A computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or data center integrated with one or more available media. The available medium may be a magnetic medium (for example, a floppy disk, hard disk, Magnetic tape), optical media (for example, digital video disc (DVD for short)), or semiconductor media (for example, SSD).

Claims

A method for reducing invalid alarms, which is characterized in that it includes:

Acquiring first characteristic information of a first target, where the behavior of the first target is a behavior that triggers a camera alarm;

The first characteristic information is matched with the characteristic information in an invalid alarm database, the invalid alarm database includes characteristic information of at least one target, and the behavior of each target in the at least one target is a behavior that triggers an invalid alarm ；

When the degree of matching between the first characteristic information and the characteristic information in the invalid alarm database is greater than or equal to a first threshold, no alarm information is output.
The method of claim 1, wherein the method further comprises:

When the degree of matching between the first feature information and the feature information in the invalid alarm database is less than a first threshold, receive first information from an input interface, where the first information is used to indicate the first target Whether the triggered alarm is an invalid alarm;

When the first information is used to indicate that the alarm triggered by the first target is an invalid alarm, no alarm information is output, and the first characteristic information is added to the invalid alarm database; otherwise, the alarm information is output.
The method of claim 2, wherein adding the first characteristic information to the invalid alarm database comprises:

Acquiring a first image, the first image including the first target;

Determine the area of the first target in the first image;

Reduce and/or enlarge the area according to different sizes to obtain at least one sub-image;

Extracting characteristic information of the first target from each of the at least one sub-image to obtain at least one second characteristic information, where the at least one second characteristic information includes the first characteristic information;

The at least one second characteristic information is added to the invalid alarm database.
The method according to any one of claims 1 to 3, wherein the acquiring first characteristic information of the first target comprises:

Receiving multimedia information from the camera;

When it is detected that there is a moving first target in the shooting area of the camera, and the behavior of the first target is a behavior that triggers an alarm, a first image is obtained from the multimedia information, and the Feature extraction is performed on a first target in an image to obtain the first feature information.
The method according to any one of claims 1 to 3, wherein the acquiring first characteristic information of the first target comprises:

Receiving the first characteristic information from the camera.
The method according to any one of claims 1 to 5, wherein when the type of the alarm triggered by the first target is the first alarm type, the first characteristic information is combined with the information in the invalid alarm database. Feature information is matched, including:

Determine the third characteristic information in the invalid alarm database according to the first alarm type, and the invalid alarm triggered by the target corresponding to the third characteristic information is the invalid alarm of the first alarm type;

Matching the first characteristic information with the third characteristic information.
A device for reducing invalid alarms, which is characterized in that it comprises:

A processing unit, configured to obtain first characteristic information of a first target, the behavior of the first target is a behavior that triggers a camera alarm; matching the first characteristic information with the characteristic information in the invalid alarm database, the invalid The alarm database includes characteristic information of at least one target, and the behavior of each target in the at least one target is a behavior that triggers an invalid alarm;

The transceiver unit is configured to not output alarm information when the degree of matching between the first characteristic information and the characteristic information in the invalid alarm database is greater than or equal to a first threshold.
The device according to claim 7, wherein the transceiver unit is further configured to:

When the degree of matching between the first feature information and the feature information in the invalid alarm database is less than a first threshold, receive first information from an input interface, where the first information is used to indicate the first target Whether the triggered alarm is an invalid alarm;

When the first information is used to indicate that the alarm triggered by the first target is an invalid alarm, no alarm information is output, and the processing unit is further configured to add the first characteristic information to the invalid alarm database, Otherwise, the transceiver unit is used to output alarm information.
The device according to claim 8, wherein the processing unit is specifically configured to:

Acquiring a first image, the first image including the first target;

Determine the area of the first target in the first image;

Reduce and/or enlarge the area according to different sizes to obtain at least one sub-image;

Extracting characteristic information of the first target from each of the at least one sub-image to obtain at least one second characteristic information, where the at least one second characteristic information includes the first characteristic information;

The at least one second characteristic information is added to the invalid alarm database.
The device according to any one of claims 7-9, wherein the transceiver unit is further configured to:

Receiving multimedia information from the camera;

The processing unit is specifically used for:

When it is detected that there is a moving first target in the shooting area of the camera, and the behavior of the first target is a behavior that triggers an alarm, a first image is obtained from the multimedia information, and the Feature extraction is performed on a first target in an image to obtain the first feature information.
The device according to any one of claims 7-9, wherein the transceiver unit is further configured to:

Receiving the first characteristic information from the camera.
The device according to any one of claims 7 to 11, wherein when the type of the alarm triggered by the first target is the first alarm type, the processing unit is specifically configured to:

Determine the third characteristic information in the invalid alarm database according to the first alarm type, and the invalid alarm triggered by the target corresponding to the third characteristic information is the invalid alarm of the first alarm type;

Matching the first characteristic information with the third characteristic information.
A computer-readable storage medium, characterized in that, a computer program or instruction is stored in the computer-readable storage medium, and when the computer reads and executes the computer program or instruction, the computer executes as claimed in claims 1 to 6 The method described in any one of.
An alarm signal processing equipment, which is characterized in that it comprises:

An interface for receiving the first feature information sent by the camera or the first image corresponding to the first feature information;

The processor communicates with the interface and is configured to execute the method according to any one of claims 1 to 6.