A repeated alarm identification method and device, medium and electronic equipment

Abstract

This application provides a method, apparatus, medium, and electronic device for identifying duplicate alarms, relating to the field of computer technology. The duplicate alarm identification method acquires multi-dimensional data information and first-moment information of the alarm information to be processed; determines the associated alarm information corresponding to the alarm information to be processed based on the first-moment information and preset time period determination rules; inputs the alarm information to be processed and the associated alarm information as source information into a trained feature extraction network model to obtain multi-dimensional basic features; the multi-dimensional basic features include a first multi-dimensional basic feature of the alarm information to be processed and a second multi-dimensional basic feature of the associated alarm information; based on the first and second multi-dimensional basic features, determines the comprehensive matching similarity information between the alarm information to be processed and the associated alarm information; and determines the duplicate alarm identification result of the alarm information to be processed based on the comprehensive matching similarity information and preset matching conditions.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a method, apparatus, medium and electronic device for identifying repeated alarms. Background Technology

[0002] In police emergency response, with the diversification of reporting channels and the continuous increase in the number of reports, identifying duplicate reports has become a crucial aspect of improving police efficiency. During the response process, due to factors such as the caller's anxiety, information transmission errors, or inaccurate location of the incident, complex scenarios frequently occur, including "one person reporting multiple times" (the same person reporting the same event multiple times), "one location reporting multiple times" (different people reporting the same event at the same location repeatedly), and "multiple people reporting one thing" (multiple independent callers reporting the same related event separately). These repeated reports not only increase the daily workload of police agencies and waste valuable public service resources, but also seriously disrupt their routine emergency response work.

[0003] Existing methods for identifying duplicate alarms typically rely on manual screening, which suffers from drawbacks such as long processing time, susceptibility to omissions, and high false positive rates, resulting in low efficiency. Therefore, developing a new method for identifying duplicate alarms that addresses this inefficiency is of significant practical importance. Summary of the Invention

[0004] This application provides a method, apparatus, medium, and electronic device for identifying duplicate alarms, which can improve the efficiency of duplicate alarm identification.

[0005] To achieve the above objectives, the technical solution of this application embodiment is implemented as follows:

[0006] In a first aspect, embodiments of this application provide a method for identifying repeated alarms, the method comprising:

[0007] Acquire multidimensional data information and first-moment information of the alarm information to be processed; the multidimensional data information is alarm information of a preset baseline data type; the baseline data type includes voiceprint data, location data and text data;

[0008] Based on the first time information and the preset time period determination rule, the associated alarm information corresponding to the alarm information to be processed is determined; the second time information corresponding to the associated alarm information satisfies the time period determination rule with the first time information.

[0009] The alarm information to be processed and the associated alarm information are respectively used as source information and input into the trained feature extraction network model to obtain multi-dimensional basic features; the multi-dimensional basic features include the first multi-dimensional basic features of the alarm information to be processed and the second multi-dimensional basic features of the associated alarm information.

[0010] Based on the first multi-dimensional basic features and the second multi-dimensional basic features, comprehensive matching similarity information between the alarm information to be processed and the associated alarm information is determined; the comprehensive matching similarity information is a feature value that characterizes the information similarity between the alarm information to be processed and the associated alarm information.

[0011] Based on the comprehensive matching similarity information and the preset matching conditions, the duplicate alarm identification result of the alarm information to be processed is determined.

[0012] The duplicate alarm identification method provided in this application, during the process of identifying duplicate alarms, acquires multi-dimensional data information and first-time information of the alarm information to be processed; the multi-dimensional data information is alarm information of a preset benchmark data type; the benchmark information type includes voiceprint data, location data, and text data; based on the first-time information and a preset time period determination rule, determines the associated alarm information corresponding to the alarm information to be processed; the second-time information corresponding to the associated alarm information satisfies the time period determination rule with the first-time information; the alarm information to be processed and the associated alarm information are respectively used as source information and input into a trained feature extraction network model to obtain multi-dimensional basic features; the multi-dimensional basic features include the first multi-dimensional basic features of the alarm information to be processed and the second multi-dimensional basic features of the associated alarm information; based on the first multi-dimensional basic features and the second multi-dimensional basic features, determines the comprehensive matching similarity information between the alarm information to be processed and the associated alarm information; the comprehensive matching similarity information represents the feature value of the information similarity between the alarm information to be processed and the associated alarm information; based on the comprehensive matching similarity information and a preset matching condition, determines the duplicate alarm identification result of the alarm information to be processed. This method improves the accuracy of duplicate alarm identification and reduces the time required for duplicate alarm identification through efficient correlation analysis of multi-dimensional information, thus effectively improving the efficiency of duplicate alarm identification.

[0013] In an optional embodiment, determining the duplicate alarm identification result of the alarm information to be processed based on the comprehensive matching similarity information and preset matching conditions includes:

[0014] If the comprehensive matching similarity information is greater than or equal to the preset first matching degree threshold, then the duplicate alarm identification result of the alarm information to be processed is determined to be a duplicate alarm.

[0015] The method of this embodiment determines the duplicate alarm identification result of the alarm information to be processed as a duplicate alarm by using the preset first matching degree threshold when the comprehensive matching similar information is greater than or equal to the first matching degree threshold. This can accurately identify duplicate alarms in the alarm information to be processed by using the preset first matching degree threshold, thereby improving the accuracy of duplicate alarm identification, reducing the time of duplicate alarm identification, and effectively improving the efficiency of duplicate alarm identification.

[0016] In an optional embodiment, determining the duplicate alarm identification result of the alarm information to be processed based on the comprehensive matching similarity information and preset matching conditions further includes:

[0017] If the comprehensive matching similarity information is less than the preset second matching degree threshold, then the duplicate alarm identification result of the alarm information to be processed is determined to be a new alarm; the second matching degree threshold is less than the first matching degree threshold.

[0018] The method in this embodiment further includes determining the duplicate alarm identification result of the alarm information to be processed as a new alarm if the comprehensive matching similarity information is less than a preset second matching degree threshold; since the second matching degree threshold is less than the first matching degree threshold, the new alarm information to be processed can also be accurately identified based on the preset second matching degree threshold, further improving the accuracy of duplicate alarm identification, reducing the time of duplicate alarm identification, and effectively improving the efficiency of duplicate alarm identification.

[0019] In an optional embodiment, determining the duplicate alarm identification result of the alarm information to be processed based on the comprehensive matching similarity information and preset matching conditions further includes:

[0020] If the comprehensive matching similarity information is greater than the second matching degree threshold and less than the first matching degree threshold, then the duplicate alarm identification result of the alarm information to be processed is determined as an alarm to be verified; the alarm information to be verified needs to be additionally verified to determine whether the alarm information is duplicated.

[0021] The method in this embodiment further includes determining the duplicate alarm identification result of the alarm information to be processed as an alarm to be verified if the comprehensive matching similarity information is greater than the second matching degree threshold and less than the first matching degree threshold. The alarm to be verified needs to be additionally verified to determine whether the alarm information is duplicated. Therefore, the method can also select the alarm to be verified from the alarm information to be processed based on the preset first matching degree threshold and second matching degree threshold, which can prompt the corresponding personnel to perform additional verification to determine whether the alarm information is duplicated, thereby helping to improve the accuracy of duplicate alarm identification, reduce the time of duplicate alarm identification, and effectively improve the efficiency of duplicate alarm identification.

[0022] In one optional embodiment, the feature extraction network model includes a data preprocessing module and a feature extraction module; the step of using the alarm information to be processed and the associated alarm information as source information, respectively, and inputting them into the trained feature extraction network model to obtain multi-dimensional basic features includes:

[0023] Select the alarm information to be processed and the associated alarm information one by one as the source information;

[0024] For each piece of source information obtained, the current source information is input into the data preprocessing module to obtain intermediate information;

[0025] The intermediate information is input into the feature extraction module to obtain the multi-dimensional basic features corresponding to the current source information.

[0026] The method of this embodiment selects the alarm information to be processed and the associated alarm information one by one as source information; for each source information obtained, the current source information is input into the data preprocessing module to obtain intermediate information; the intermediate information is input into the feature extraction module to obtain the multi-dimensional basic features corresponding to the current source information. The data preprocessing module can preprocess the alarm information to be processed before feature extraction, which can achieve data standardization and eliminate data interference, improve the data quality of multi-dimensional data, improve the accuracy of duplicate alarm identification, reduce the time of duplicate alarm identification, and effectively improve the efficiency of duplicate alarm identification.

[0027] In one optional embodiment, the data preprocessing module includes a first submodule, a second submodule, and a third submodule; the step of inputting the current source information into the data preprocessing module to obtain intermediate information includes:

[0028] The first sub-class information in the current source information is input into the first sub-module for sound information standardization to obtain the first sub-information; the sound information standardization operation includes noise reduction; the base data type corresponding to the first sub-class information is the voiceprint data;

[0029] The second sub-class information from the current source information is input into the second sub-module for geographic information standardization to obtain the second sub-information; the geographic information standardization operation includes coordinate unification, outlier removal, and map matching; the base data type corresponding to the second sub-class information is the location data;

[0030] The third subclass information in the current source information is input into the third submodule for text information standardization to obtain the third sub-information; the text information standardization operation includes dialect translation; the base data type corresponding to the third subclass information is the text data;

[0031] The intermediate information is obtained based on the first sub-information, the second sub-information, and the third sub-information.

[0032] The method in this embodiment, by setting up sub-modules for data preprocessing according to each dimension of multidimensional data information, can perform targeted data standardization and noise reduction on alarm information of different dimensions. Dialect translation can be performed through a dialect translation standardization model, which is constructed based on a dialect semantic standardization system driven by a time relation network. This can improve the accuracy of dialect conversion, thereby further improving the accuracy of duplicate alarm identification, reducing the time for duplicate alarm identification, and more effectively improving the efficiency of duplicate alarm identification.

[0033] In an optional embodiment, determining the comprehensive matching similarity information between the alarm information to be processed and the associated alarm information based on the first multi-dimensional basic features and the second multi-dimensional basic features includes:

[0034] Based on the first multi-dimensional basic features and the second multi-dimensional basic features, multi-dimensional similarity information is determined; the multi-dimensional similarity information includes voiceprint similarity, spatial similarity and text similarity;

[0035] Based on the multidimensional data information and the first time information, the target scene type corresponding to the alarm information to be processed is determined;

[0036] Based on the target scene type and the preset mapping relationship between scene type and multidimensional weight information, the target multidimensional weight information corresponding to the alarm information to be processed is determined; the multidimensional weight information includes voiceprint weight, spatial weight and text weight.

[0037] Based on the multidimensional similarity information and the target multidimensional weight information, comprehensive matching similarity information between the alarm information to be processed and the associated alarm information is obtained.

[0038] The method of this embodiment determines multidimensional similarity information based on the first and second multidimensional basic features; the multidimensional similarity information includes voiceprint similarity, spatial similarity, and text similarity; based on the multidimensional data information and the first time information, the target scene type corresponding to the alarm information to be processed is determined; based on the target scene type and the mapping relationship between the preset scene type and multidimensional weight information, the target multidimensional weight information corresponding to the alarm information to be processed is determined; the multidimensional weight information includes voiceprint weight, spatial weight, and text weight; based on the multidimensional similarity information and the target multidimensional weight information, comprehensive matching similarity information between the alarm information to be processed and the associated alarm information is obtained, thereby providing a mechanism for multidimensional dynamic decision-making in the identification of repeated alarms, achieving more accurate multimodal fusion identification, enhancing the robustness of repeated alarm identification, and further improving the efficiency of repeated alarm identification.

[0039] Secondly, embodiments of this application also provide a repeat alarm identification device, comprising:

[0040] The basic information acquisition module is used to acquire multi-dimensional data information and first-moment information of the alarm information to be processed; the multi-dimensional data information is alarm information of a preset baseline data type; the baseline information type includes voiceprint data, location data and text data;

[0041] The associated alarm selection module is used to determine the associated alarm information corresponding to the alarm information to be processed based on the first time information and the preset time period determination rules; the second time information corresponding to the associated alarm information satisfies the time period determination rules with the first time information.

[0042] The multidimensional feature acquisition module is used to take the alarm information to be processed and the associated alarm information as source information, respectively, and input them into the trained feature extraction network model to obtain multidimensional basic features; the multidimensional basic features include the first multidimensional basic features of the alarm information to be processed and the second multidimensional basic features of the associated alarm information.

[0043] The multidimensional matching evaluation module is used to determine the comprehensive matching similarity information between the alarm information to be processed and the associated alarm information based on the first multidimensional basic features and the second multidimensional basic features; the comprehensive matching similarity information is a feature value that characterizes the information similarity between the alarm information to be processed and the associated alarm information.

[0044] The identification result determination module is used to determine the duplicate alarm identification result of the alarm information to be processed based on the comprehensive matching similarity information and the preset matching conditions.

[0045] In an optional embodiment, the identification result determination module is specifically used for:

[0046] If the comprehensive matching similarity information is greater than or equal to the preset first matching degree threshold, then the duplicate alarm identification result of the alarm information to be processed is determined to be a duplicate alarm.

[0047] In an optional embodiment, the identification result determination module is further configured to:

[0048] If the comprehensive matching similarity information is less than the preset second matching degree threshold, then the duplicate alarm identification result of the alarm information to be processed is determined to be a new alarm; the second matching degree threshold is less than the first matching degree threshold.

[0049] In an optional embodiment, the identification result determination module is further configured to:

[0050] If the comprehensive matching similarity information is greater than the second matching degree threshold and less than the first matching degree threshold, then the duplicate alarm identification result of the alarm information to be processed is determined as an alarm to be verified; the alarm information to be verified needs to be additionally verified to determine whether the alarm information is duplicated.

[0051] In one optional embodiment, the feature extraction network model includes a data preprocessing module and a feature extraction module; the multidimensional feature acquisition module is specifically used for:

[0052] Select the alarm information to be processed and the associated alarm information one by one as the source information;

[0053] For each piece of source information obtained, the current source information is input into the data preprocessing module to obtain intermediate information;

[0054] The intermediate information is input into the feature extraction module to obtain the multi-dimensional basic features corresponding to the current source information.

[0055] In one optional embodiment, the data preprocessing module includes a first sub-module, a second sub-module, and a third sub-module; the multidimensional feature acquisition module is specifically used for:

[0056] The first sub-class information in the current source information is input into the first sub-module for sound information standardization to obtain the first sub-information; the sound information standardization operation includes noise reduction; the base data type corresponding to the first sub-class information is the voiceprint data;

[0057] The second sub-class information from the current source information is input into the second sub-module for geographic information standardization to obtain the second sub-information; the geographic information standardization operation includes coordinate unification, outlier removal, and map matching; the base data type corresponding to the second sub-class information is the location data;

[0058] The third subclass information in the current source information is input into the third submodule for text information standardization to obtain the third sub-information; the text information standardization operation includes dialect translation; the base data type corresponding to the third subclass information is the text data;

[0059] The intermediate information is obtained based on the first sub-information, the second sub-information, and the third sub-information.

[0060] In one optional embodiment, the multidimensional matching evaluation module is specifically used for:

[0061] Based on the first multi-dimensional basic features and the second multi-dimensional basic features, multi-dimensional similarity information is determined; the multi-dimensional similarity information includes voiceprint similarity, spatial similarity and text similarity;

[0062] Based on the multidimensional data information and the first time information, the target scene type corresponding to the alarm information to be processed is determined;

[0063] Based on the target scene type and the preset mapping relationship between scene type and multidimensional weight information, the target multidimensional weight information corresponding to the alarm information to be processed is determined; the multidimensional weight information includes voiceprint weight, spatial weight and text weight.

[0064] Based on the multidimensional similarity information and the target multidimensional weight information, comprehensive matching similarity information between the alarm information to be processed and the associated alarm information is obtained.

[0065] Thirdly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the repeat alarm identification method of the first aspect.

[0066] Fourthly, embodiments of this application also provide an electronic device, including a memory and a processor, wherein the memory stores a computer program that can run on the processor, and when the computer program is executed by the processor, the processor enables the processor to implement the repeat alarm identification method of the first aspect.

[0067] The technical effects of any of the implementation methods in the second to fourth aspects can be found in the technical effects of the corresponding implementation methods in the first aspect, and will not be repeated here. Attached Figure Description

[0068] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0069] Figure 1 A flowchart illustrating a method for identifying repeated alarms provided in an embodiment of this application;

[0070] Figure 2 A flowchart illustrating the process of obtaining multi-dimensional basic features in a method for identifying repeat alarms provided in this application embodiment;

[0071] Figure 3 A flowchart illustrating the process of inputting current source information into a data preprocessing module to obtain intermediate information, as provided in an embodiment of this application, for a method of identifying repeated alarms.

[0072] Figure 4 A schematic diagram illustrating the process of determining and comprehensively matching similar information for a method of identifying repeated alarms provided in this application embodiment;

[0073] Figure 5 This is one of the flowcharts illustrating the process of determining the duplicate alarm identification result of the duplicate alarm information to be processed in a duplicate alarm identification method provided in this application embodiment;

[0074] Figure 6 The second schematic diagram of the process for determining the duplicate alarm identification result of the duplicate alarm information to be processed in a duplicate alarm identification method provided in the embodiments of this application;

[0075] Figure 7 This is a schematic diagram of the structure of a repeat alarm identification device provided in an embodiment of this application;

[0076] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0077] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0078] It should be noted that the terms "comprising" and "having" and their variations used in this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to such process, method, product, or device.

[0079] The following are explanations of some of the words that appear in the text:

[0080] (1) Multiple reports by one person: The same person reports the same event multiple times in a short period of time (such as an intoxicated person dialing 110 repeatedly).

[0081] (2) Multiple alarms in one location: Multiple independent alarm events occur in the same geographical location (such as theft in a shopping mall).

[0082] (3) Multiple people report one: Different people report the same event at the same time (such as a traffic accident).

[0083] (4) Dialect translation standardization model: a knowledge base that translates dialect descriptions (such as “fighting”) into standard event codes (“brawl”).

[0084] (5) Density adaptive radius algorithm: an algorithm that dynamically adjusts the spatial matching radius based on real-time pedestrian flow density, calculates the spatial matching degree through dynamic radius, and improves the recognition accuracy of spatial overlap events.

[0085] In police emergency response, with the diversification of reporting channels and the continuous increase in the number of reports, identifying duplicate reports has become a crucial aspect of improving police efficiency. During the response process, due to factors such as the caller's anxiety, information transmission errors, or inaccurate location of the incident, complex scenarios frequently occur, including "one person reporting multiple times" (the same person reporting the same event multiple times), "one location reporting multiple times" (different people reporting the same event at the same location repeatedly), and "multiple people reporting one thing" (multiple independent callers reporting the same related event separately). These repeated reports not only increase the daily workload of police agencies and waste valuable public service resources, but also seriously disrupt their routine emergency response work.

[0086] Existing methods for identifying duplicate alarms typically rely on manual screening, which suffers from drawbacks such as long processing time, susceptibility to omissions, and high false positive rates, resulting in low efficiency. Therefore, developing a new method for identifying duplicate alarms that addresses this inefficiency is of significant practical importance.

[0087] To address existing technical problems, embodiments of this application provide a method, apparatus, medium, and electronic device for identifying duplicate alarms. In the process of identifying duplicate alarms, multi-dimensional data information and first-moment information of the alarm information to be processed are acquired. The multi-dimensional data information is alarm information of a preset baseline data type. The baseline information types include voiceprint data, location data, and text data. Based on the first-moment information and preset time period determination rules, associated alarm information corresponding to the alarm information to be processed is determined. The second-moment information corresponding to the associated alarm information satisfies the time period determination rules with the first-moment information. The alarm information to be processed and the associated alarm information are respectively used as source information and input into a trained feature extraction network model to obtain multi-dimensional basic features. The multi-dimensional basic features include the first multi-dimensional basic features of the alarm information to be processed and the second multi-dimensional basic features of the associated alarm information. Based on the first and second multi-dimensional basic features, comprehensive matching similarity information between the alarm information to be processed and the associated alarm information is determined. The comprehensive matching similarity information is a feature value representing the information similarity between the alarm information to be processed and the associated alarm information. Based on the comprehensive matching similarity information and preset matching conditions, the duplicate alarm identification result of the alarm information to be processed is determined. This method improves the accuracy of duplicate alarm identification and reduces the time required for duplicate alarm identification through efficient correlation analysis of multi-dimensional information, thus effectively improving the efficiency of duplicate alarm identification.

[0088] To make the inventive objectives, technical solutions, and advantages of the embodiments of this application clearer, the present application will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0089] The technical solutions provided in the embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0090] The duplicate alarm identification method provided in this application can be applied to a server or a terminal device. The following embodiments of this application illustrate the application of the duplicate alarm identification method to a terminal device.

[0091] This application provides a method for identifying repeated alarms, such as... Figure 1 As shown, it includes the following steps:

[0092] Step S101: Obtain multidimensional data information and first-moment information of the alarm information to be processed; the multidimensional data information is alarm information of a preset baseline data type; the baseline information type includes voiceprint data, location data and text data.

[0093] In practice, the terminal device acquires multidimensional data information and first-moment information of the alarm information to be processed; the multidimensional data information is alarm information of a preset baseline data type; the baseline information types include voiceprint data, location data and text data.

[0094] In some embodiments of this application, the location data is GIS information.

[0095] Step S102: Based on the first moment information and the preset time period determination rules, determine the associated alarm information corresponding to the alarm information to be processed; the second moment information corresponding to the associated alarm information and the first moment information satisfy the time period determination rules.

[0096] In practice, the terminal device determines the associated alarm information corresponding to the alarm information to be processed based on the first moment information and the preset time period determination rules; the second moment information corresponding to the associated alarm information satisfies the time period determination rules with the first moment information.

[0097] In some embodiments of this application, the time period determination rule may be that the time interval between two pieces of time information does not exceed a preset time threshold.

[0098] For example, in one embodiment, the time period determination rule can be that the second time information corresponding to the associated alarm information and the first time information meet the requirement that the time interval does not exceed 1 hour.

[0099] Step S103: The alarm information to be processed and the associated alarm information are used as source information and input into the trained feature extraction network model to obtain multi-dimensional basic features. The multi-dimensional basic features include the first multi-dimensional basic features of the alarm information to be processed and the second multi-dimensional basic features of the associated alarm information.

[0100] In practice, the terminal device takes the alarm information to be processed and the associated alarm information as source information and inputs them into the trained feature extraction network model to obtain multi-dimensional basic features. The multi-dimensional basic features include the first multi-dimensional basic features of the alarm information to be processed and the second multi-dimensional basic features of the associated alarm information.

[0101] In some optional embodiments, the feature extraction network model includes a data preprocessing module and a feature extraction module; in step S103 above, the process of using the alarm information to be processed and the associated alarm information as source information, respectively, and inputting them into the trained feature extraction network model to obtain multi-dimensional basic features, such as... Figure 2 As shown, this can be achieved through the following steps:

[0102] Step S201: Select the alarm information to be processed and the associated alarm information one by one as the source information.

[0103] In practice, the terminal device selects the alarm information to be processed and the associated alarm information one by one as the source information.

[0104] In step S202, for each piece of source information obtained, the current source information is input into the data preprocessing module to obtain intermediate information.

[0105] In practice, each time the terminal device receives a source information, it inputs the current source information into the data preprocessing module to obtain intermediate information.

[0106] Step S203: Input the intermediate information into the feature extraction module to obtain the multi-dimensional basic features corresponding to the current source information.

[0107] In practice, the terminal device inputs the intermediate information into the feature extraction module to obtain the multi-dimensional basic features corresponding to the current source information.

[0108] The method in this embodiment can preprocess the alarm information to be processed before feature extraction through the data preprocessing module, which can achieve data standardization and eliminate data interference, improve the data quality of multi-dimensional data, improve the accuracy of duplicate alarm identification, reduce the time of duplicate alarm identification, and effectively improve the efficiency of duplicate alarm identification.

[0109] In some optional embodiments, the process of inputting the current source information into the data preprocessing module to obtain intermediate information in step S203 above, such as... Figure 3 As shown, this can be achieved through the following steps:

[0110] Step S301: Input the first subclass information in the current source information into the first submodule for sound information standardization operation to obtain the first sub-information; the sound information standardization operation includes noise reduction; the reference data type corresponding to the first subclass information is voiceprint data.

[0111] Step S302: Input the second subclass information from the current source information into the second submodule for geographic information standardization to obtain the second sub-information; the geographic information standardization operation includes coordinate unification, outlier removal, and map matching; the base data type corresponding to the second subclass information is location data.

[0112] Step S303: Input the third subclass information in the current source information into the third submodule for text information standardization to obtain the third sub-information; the text information standardization operation includes dialect translation; the base data type corresponding to the third subclass information is text data.

[0113] In some embodiments of this application, the dialect translation process is carried out through a dialect translation standardization model, which is constructed based on a dialect semantic standardization system driven by an event relation network. Specifically, an alarm recording corpus is used as training data to annotate event type codes, achieving synonymous event chain mapping (e.g., "fighting" is mapped to "fighting") and dialect translation standardization (e.g., translating "fighting" into the standard event code "fighting"). Through the above method, the dialect alarm text is transformed into a standard semantic expression, thereby effectively improving the accuracy of text matching. Examples of dialect translation are shown in Table 1:

[0114] Table 1

[0115]

[0116] In some embodiments of this application, numerous event relationships can constitute an event relationship network.

[0117] Step S304: Based on the first sub-information, the second sub-information, and the third sub-information, intermediate information is obtained.

[0118] In some embodiments of this application, intermediate information is obtained based on the first sub-information, the second sub-information, and the third sub-information. Specifically, the intermediate information is obtained by synchronizing the first sub-information, the second sub-information, and the third sub-information in time.

[0119] In some embodiments of this application, the data preprocessing module mainly performs comprehensive standardization processing on the input voiceprint data, GIS information and text data to eliminate data interference, unify data format and solve dialect semantic problems.

[0120] In some embodiments of this application, the standardization process mainly includes: preprocessing the voiceprint data such as noise reduction; performing coordinate unification, outlier removal, and map matching on the GIS information; performing word segmentation, stop word removal, part-of-speech tagging, and dialect translation on the text data; at the same time, the system also performs data alignment and time synchronization to ensure the consistency of different modal data in time and space.

[0121] In some embodiments of this application, the feature extraction network model is used to mine multi-dimensional and multi-level modal features from preprocessed data, providing basic feature vectors for subsequent analysis.

[0122] In some embodiments of this application, source information is input into a trained feature extraction network model to obtain multi-dimensional basic features, including: extracting voiceprint features from the voiceprint data of the source information; extracting spatial features from the GIS information of the source information; and extracting text features from the text data of the source information.

[0123] In some embodiments of this application, source information is input into a trained feature extraction network model to obtain multi-dimensional basic features, including: extracting fundamental frequency features and voiceprint vectors from the voiceprint data of the source information and normalizing them; extracting spatial location features, movement trajectory features, and regional hotspot features from the GIS information of the source information; and extracting word embedding vectors and semantic features from the text data.

[0124] In the method of this embodiment, by setting up sub-modules for data preprocessing according to the dimensions of multidimensional data information, targeted data standardization and noise reduction can be performed on alarm information of different dimensions. Dialect translation can be performed through a dialect translation standardization model, which is constructed based on a dialect semantic standardization system driven by a time relation network. This can improve the accuracy of dialect conversion, thereby further improving the accuracy of duplicate alarm identification, reducing the time for duplicate alarm identification, and more effectively improving the efficiency of duplicate alarm identification.

[0125] Step S104: Based on the first multi-dimensional basic features and the second multi-dimensional basic features, determine the comprehensive matching similarity information between the alarm information to be processed and the associated alarm information; the comprehensive matching similarity information is a feature value that represents the information similarity between the alarm information to be processed and the associated alarm information.

[0126] In practice, the terminal device determines the comprehensive matching similarity information between the alarm information to be processed and the associated alarm information based on the first multi-dimensional basic features and the second multi-dimensional basic features; the comprehensive matching similarity information is a feature value that represents the information similarity between the alarm information to be processed and the associated alarm information.

[0127] In some embodiments of this application, the feature value of information similarity is specifically the maximum value of information similarity.

[0128] In some optional embodiments, the process of determining the comprehensive matching similarity information between the alarm information to be processed and the associated alarm information based on the first multi-dimensional basic features and the second multi-dimensional basic features in step S104 above, such as... Figure 4 As shown, this can be achieved through the following steps:

[0129] Step S401: Based on the first multi-dimensional basic features and the second multi-dimensional basic features, determine multi-dimensional similarity information; the multi-dimensional similarity information includes voiceprint similarity, spatial similarity and text similarity.

[0130] In practice, the terminal device determines multidimensional similarity information based on the first and second multidimensional basic features; the multidimensional similarity information includes voiceprint similarity, spatial similarity and text similarity.

[0131] In some embodiments of this application, voiceprint similarity can be determined by cosine similarity calculation.

[0132] The process of determining voiceprint similarity includes: when voiceprint features are input, the data vector is first normalized, and then the voiceprint similarity S_v is calculated using cosine similarity.

[0133] The formula for calculating voiceprint similarity is:

[0134]

[0135] Where V_i represents the sound vector numbered i;

[0136] V_j represents the sound vector numbered j;

[0137] × indicates calculating the cosine similarity.

[0138] In some embodiments of this application, spatial similarity can be determined using a density-adaptive radius algorithm.

[0139] In some embodiments of this application, during the determination of spatial similarity, when spatial features are input, a density-adaptive radius algorithm is used to overcome the fixed radius threshold limitation and achieve dynamic spatial matching. First, the dynamic radius is calculated using the dynamic radius formula. Then, the ratio of the number of spatial points within the radius to the total number of spatial points is calculated to obtain the spatial matching degree G_d. The formula for calculating the spatial matching degree is:

[0140]

[0141] in,

[0142] R is the dynamic radius.

[0143] In some embodiments of this application, the dynamic radius can be determined by the following formula:

[0144]

[0145] R_base is the baseline radius. Typically, the baseline radius for residential areas is 50m, and the baseline radius for commercial areas is 150m.

[0146] ρ represents the real-time pedestrian density;

[0147] K is the adjustment coefficient (default value 20);

[0148] , where is the number of spatial points within radius r;

[0149] , where is the total number of spatial points.

[0150] In some embodiments of this application, text similarity can be determined using a dialect translation standardization model and cosine similarity.

[0151] In some embodiments of this application, the process of determining text similarity includes: when text features are input, firstly, the dialect translation standardization model is standardized, and then the data vector is normalized; then, the text similarity T_s is calculated using cosine similarity.

[0152] The formula for calculating text similarity is:

[0153]

[0154] Where T_i represents the text vector numbered i;

[0155] T_j represents the text vector with index j;

[0156] × indicates calculating the cosine similarity.

[0157] Step S402: Based on the multidimensional data information and the first-moment information, determine the target scene type corresponding to the alarm information to be processed.

[0158] In practice, the terminal device determines the target scenario type corresponding to the alarm information to be processed based on multi-dimensional data information and first-moment information.

[0159] In some embodiments of this application, the scene types include nighttime alarm type, group dispute type, and dialect alarm type; the scene type is used to dynamically adjust the weight ratio of multidimensional similarity information of different modal data.

[0160] In other embodiments of this application, the scene type also includes nighttime group type, nighttime dialect type, and group dialect type.

[0161] Step S403: Determine the target multidimensional weight information corresponding to the alarm information to be processed based on the target scene type and the mapping relationship between the preset scene type and multidimensional weight information; the multidimensional weight information includes voiceprint weight, spatial weight and text weight.

[0162] In some embodiments, the weight ratio of different modal data is dynamically adjusted according to the alarm type scenario (such as nighttime, group, or dialect alarm).

[0163] The dynamic weight allocation rules are shown in Table 2:

[0164] Table 2

[0165]

[0166] In some embodiments of this application, a dynamic weight allocation rule is adopted in the process of determining the comprehensive matching similar information. Specifically, the weights of the three modal features of voiceprint (S_v), space (G_d) and text (T_s) are adaptively adjusted according to different scene types, and the voiceprint weight, space weight and text weight are dynamically determined to improve the accuracy and robustness of fusion recognition.

[0167] In some embodiments, the dynamic weight allocation rule is as follows:

[0168] Nighttime alarm scenario: Due to the high ambient noise at night, voiceprint information is easily distorted. Therefore, the voiceprint weight is appropriately reduced (α=0.2), while the spatial weight is increased (β=0.5) to make full use of GIS information to make up for the lack of voiceprint information. The text weight is kept at a medium level (γ=0.3).

[0169] Group dispute scenario: In this scenario, voiceprint, spatial and text information are relatively rich and their importance is relatively balanced. Therefore, the weight allocation of the three modalities is relatively balanced (α=0.3, β=0.4, γ=0.3) to make full use of the comprehensive information of multimodal features;

[0170] Dialect alarm scenario: Since dialect differences may reduce the accuracy of voiceprint recognition, the voiceprint weight is significantly reduced (α=0.1), while the text weight is greatly increased (γ=0.6) to enhance the semantic understanding of text information. The spatial weight is appropriately reduced (β=0.3) to avoid spatial information from interfering with text information.

[0171] Through the dynamic weight allocation strategy described above, the system can adaptively adjust the contribution of each modality feature according to the characteristics of different scenarios, thereby achieving more accurate and robust multimodal fusion recognition.

[0172] Similarly, in some embodiments of this application, the dynamic weight allocation rules may also be as shown in Table 3:

[0173] Table 3

[0174]

[0175] Through the dynamic weight allocation strategy described above, the system can more precisely and adaptively adjust the contribution of each modal feature according to the characteristics of different scenarios, thereby further improving accuracy, providing better robustness, and achieving better multimodal fusion recognition.

[0176] Step S404: Based on the multidimensional similarity information and the target multidimensional weight information, obtain the comprehensive matching similarity information between the alarm information to be processed and the associated alarm information.

[0177] In practice, the terminal device is based on a multimodal dynamic decision engine and adopts a scene-adaptive weight allocation mechanism. First, it calculates voiceprint similarity, spatial similarity and text similarity respectively. Then, according to the specific alarm scene type (such as nighttime, group or dialect alarm), it dynamically adjusts the weight ratio of each modality data. Finally, it uses a cross-modal association formula to comprehensively calculate the overall matching degree, so as to realize multimodal collaborative multi-dimensional dynamic decision-making.

[0178] Terminal devices use cross-modal association formulas to comprehensively calculate the overall matching degree, obtain comprehensive matching similarity information, and realize intelligent decision-making through multimodal collaboration.

[0179] The cross-modal correlation formula is:

[0180]

[0181] Where Sim_ai represents the comprehensive matching similarity information, indicating the comprehensive matching degree;

[0182] , which is the voiceprint weight;

[0183] , where is the spatial weight;

[0184] , which represents the text weight.

[0185] The method in this embodiment provides a mechanism for identifying repeated alarm incidents through multi-dimensional dynamic decision-making, achieving more accurate multi-modal fusion identification, enhancing the robustness of repeated alarm incident identification, and further improving the efficiency of repeated alarm incident identification.

[0186] Step S105: Based on the comprehensive matching of similar information and the preset matching conditions, determine the duplicate alarm identification result of the alarm information to be processed.

[0187] In one embodiment, the duplicate alarm identification result of the alarm information to be processed is determined based on comprehensive matching similarity information and preset matching conditions, including:

[0188] If the overall matching similarity information is greater than or equal to the preset first matching degree threshold, then the duplicate alarm identification result of the alarm information to be processed is determined to be a duplicate alarm.

[0189] The method in this embodiment determines that the duplicate alarm information to be processed is a duplicate alarm if the comprehensive matching similar information is greater than or equal to a preset first matching degree threshold. By using the preset first matching degree threshold, duplicate alarms in the alarm information to be processed can be accurately identified, thereby improving the accuracy of duplicate alarm identification, reducing the time of duplicate alarm identification, and effectively improving the efficiency of duplicate alarm identification.

[0190] In one embodiment, determining the duplicate alarm identification result of the alarm information to be processed based on comprehensive matching similarity information and preset matching conditions further includes:

[0191] If the overall matching similarity information is less than the preset second matching degree threshold, then the duplicate alarm identification result of the alarm information to be processed is determined as a new alarm; the second matching degree threshold is less than the first matching degree threshold.

[0192] The method of this embodiment further includes determining the duplicate alarm identification result of the alarm information to be processed as a new alarm if the comprehensive matching similarity information is less than a preset second matching degree threshold; since the second matching degree threshold is less than the first matching degree threshold, the new alarm information in the alarm information to be processed can also be accurately identified based on the preset second matching degree threshold, thereby further improving the accuracy of duplicate alarm identification, reducing the time of duplicate alarm identification, and effectively improving the efficiency of duplicate alarm identification.

[0193] In one embodiment, determining the duplicate alarm identification result of the alarm information to be processed based on comprehensive matching similarity information and preset matching conditions further includes:

[0194] If the overall matching similarity information is greater than the second matching degree threshold and less than the first matching degree threshold, then the duplicate alarm identification result of the alarm information to be processed is determined as an alarm to be verified; the alarm information to be verified needs to be additionally verified to determine whether the alarm information is duplicated.

[0195] The method in this embodiment further includes determining the duplicate alarm identification result of the alarm information to be processed as an alarm to be verified if the comprehensive matching similarity information is greater than the second matching degree threshold and less than the first matching degree threshold. The alarm to be verified needs to be additionally verified to determine whether the alarm information is duplicated. Therefore, the method can also select the alarm to be verified from the alarm information to be processed based on the preset first matching degree threshold and second matching degree threshold, which can prompt the corresponding personnel to perform additional verification to determine whether the alarm information is duplicated, thereby helping to improve the accuracy of duplicate alarm identification, reduce the time of duplicate alarm identification, and effectively improve the efficiency of duplicate alarm identification.

[0196] In some optional embodiments, the process of determining the duplicate alarm identification result of the alarm information to be processed based on the comprehensive matching similarity information and preset matching conditions in step S105 above, such as... Figure 5 As shown, this can be achieved through the following steps:

[0197] Step S501: If the comprehensive matching similarity information is greater than or equal to the preset first matching degree threshold, then the duplicate alarm identification result of the alarm information to be processed is determined to be a duplicate alarm.

[0198] Step S502: If the comprehensive matching similarity information is less than the preset second matching degree threshold, then the duplicate alarm identification result of the alarm information to be processed is determined to be a new alarm; the second matching degree threshold is less than the first matching degree threshold.

[0199] In some optional embodiments, the process of determining the duplicate alarm identification result of the alarm information to be processed based on the comprehensive matching similarity information and preset matching conditions in step S105 above, such as... Figure 6 As shown, this can be achieved through the following steps:

[0200] Step S601: If the comprehensive matching similarity information is greater than or equal to the preset first matching degree threshold, then the duplicate alarm identification result of the alarm information to be processed is determined to be a duplicate alarm.

[0201] Step S602: If the overall matching similarity information is less than the preset second matching degree threshold, then the duplicate alarm identification result of the alarm information to be processed is determined to be a new alarm; the second matching degree threshold is less than the first matching degree threshold.

[0202] Step S603: If the comprehensive matching similarity information is greater than the second matching degree threshold and less than the first matching degree threshold, then the duplicate alarm identification result of the alarm information to be processed is determined as an alarm to be verified; the alarm information to be verified needs to be additionally verified to determine whether the alarm information is duplicated.

[0203] In some embodiments of this application, the method further includes: aggregating heterogeneous alarm data sources to obtain multi-dimensional alarm information to be processed.

[0204] In practice, heterogeneous alarm data sources are aggregated to construct a multi-dimensional perception channel, obtaining multi-dimensional alarm information to be processed. By establishing a synchronous timestamp management mechanism, the spatiotemporal alignment of multimodal data is ensured, while the multi-dimensional perception channel is constructed to receive various types of data, providing basic information for data processing in the identification of recurring alarms.

[0205] For example, in the process of aggregating heterogeneous alarm data sources, the input data types are divided into: voiceprint data, GIS information, and text data. Voiceprint data mainly includes the original recordings of the alarm caller's voice (including background noise, dialects, and interference from intoxication, etc.); GIS information refers to the alarm location coordinates (latitude and longitude); and text data refers to the textual description of the alarm content (including dialects and non-standard expressions). This multimodal data can reflect alarm information from different perspectives, providing rich evidence for accurately identifying duplicate alarms.

[0206] The duplicate alarm identification method provided in this application acquires multi-dimensional data information and first-moment information of the alarm information to be processed during the identification process. The multi-dimensional data information is alarm information of a preset benchmark data type. The benchmark information types include voiceprint data, location data, and text data. Based on the first-moment information and preset time period determination rules, the associated alarm information corresponding to the alarm information to be processed is determined. The second-moment information corresponding to the associated alarm information satisfies the time period determination rules with the first-moment information. The alarm information to be processed and the associated alarm information are respectively used as source information and input into a trained feature extraction network model to obtain multi-dimensional basic features. The multi-dimensional basic features include the first multi-dimensional basic features of the alarm information to be processed and the second multi-dimensional basic features of the associated alarm information. Based on the first multi-dimensional basic features and the second multi-dimensional basic features, the comprehensive matching similarity information between the alarm information to be processed and the associated alarm information is determined. The comprehensive matching similarity information is a feature value that characterizes the information similarity between the alarm information to be processed and the associated alarm information. Based on the comprehensive matching similarity information and preset matching conditions, the duplicate alarm identification result of the alarm information to be processed is determined. This method, through efficient correlation analysis of multi-dimensional information, can improve the accuracy of duplicate alarm identification, reduce the time required for duplicate alarm identification, and effectively improve the efficiency of duplicate alarm identification.

[0207] Based on the same inventive concept, this application also provides a duplicate alarm identification device. Since this device corresponds to the duplicate alarm identification method provided in this application, and the principle of solving the problem by this device is similar to that of the method, the implementation of this device can refer to the implementation of the above method, and the repeated parts will not be described again.

[0208] Figure 7 This paper shows a schematic diagram of the structure of a repeat alarm identification device provided in an embodiment of this application, as shown below. Figure 7 As shown, the duplicate alarm identification device includes a basic information acquisition module 701, a related alarm preliminary selection module 702, a multi-dimensional feature acquisition module 703, a multi-dimensional matching evaluation module 704, and an identification result determination module 705.

[0209] The basic information acquisition module 701 is used to acquire multi-dimensional data information and first-moment information of the alarm information to be processed; the multi-dimensional data information is alarm information of a preset baseline data type; the baseline information type includes voiceprint data, location data and text data;

[0210] The associated alarm selection module 702 is used to determine the associated alarm information corresponding to the alarm information to be processed based on the first moment information and the preset time period determination rules; the second moment information corresponding to the associated alarm information and the first moment information satisfy the time period determination rules.

[0211] The multi-dimensional feature acquisition module 703 is used to take the alarm information to be processed and the associated alarm information as source information, respectively, and input them into the trained feature extraction network model to obtain multi-dimensional basic features; the multi-dimensional basic features include the first multi-dimensional basic features of the alarm information to be processed and the second multi-dimensional basic features of the associated alarm information.

[0212] The multidimensional matching evaluation module 704 is used to determine the comprehensive matching similarity information between the alarm information to be processed and the associated alarm information based on the first multidimensional basic features and the second multidimensional basic features; the comprehensive matching similarity information is a feature value that represents the information similarity between the alarm information to be processed and the associated alarm information.

[0213] The identification result determination module 705 is used to determine the duplicate alarm identification result of the alarm information to be processed based on the comprehensive matching similar information and the preset matching conditions.

[0214] In an optional embodiment, the identification result determination module 705 is specifically used for:

[0215] If the overall matching similarity information is greater than or equal to the preset first matching degree threshold, then the duplicate alarm identification result of the alarm information to be processed is determined to be a duplicate alarm.

[0216] In an optional embodiment, the identification result determination module 705 is further configured to:

[0217] If the overall matching similarity information is less than the preset second matching degree threshold, then the duplicate alarm identification result of the alarm information to be processed is determined as a new alarm; the second matching degree threshold is less than the first matching degree threshold.

[0218] In an optional embodiment, the identification result determination module 705 is further configured to:

[0219] If the overall matching similarity information is greater than the second matching degree threshold and less than the first matching degree threshold, then the duplicate alarm identification result of the alarm information to be processed is determined as an alarm to be verified; the alarm information to be verified needs to be additionally verified to determine whether the alarm information is duplicated.

[0220] In one optional embodiment, the feature extraction network model includes a data preprocessing module and a feature extraction module; the multidimensional feature acquisition module 703 is specifically used for:

[0221] Select the alarm information to be processed and related alarm information one by one as the source information;

[0222] For each piece of source information obtained, the current source information is input into the data preprocessing module to obtain intermediate information;

[0223] The intermediate information is input into the feature extraction module to obtain the multi-dimensional basic features corresponding to the current source information.

[0224] In an optional embodiment, the data preprocessing module includes a first sub-module, a second sub-module, and a third sub-module; the multidimensional feature acquisition module 703 is specifically used for:

[0225] The first subclass information in the current source information is input into the first submodule for sound information standardization to obtain the first sub-information; the sound information standardization operation includes noise reduction; the base data type corresponding to the first subclass information is voiceprint data;

[0226] The second subclass information from the current source information is input into the second submodule for geographic information standardization to obtain the second sub-information. The geographic information standardization operation includes coordinate unification, outlier removal, and map matching. The base data type corresponding to the second subclass information is location data.

[0227] The third subclass information in the current source information is input into the third submodule for text information standardization to obtain the third sub-information; the text information standardization operation includes dialect translation; the base data type corresponding to the third subclass information is text data;

[0228] Based on the first sub-information, the second sub-information, and the third sub-information, intermediate information is obtained.

[0229] In an optional embodiment, the multidimensional matching evaluation module 704 is specifically used for:

[0230] Based on the first and second multi-dimensional basic features, multi-dimensional similarity information is determined; the multi-dimensional similarity information includes voiceprint similarity, spatial similarity, and text similarity.

[0231] Based on multidimensional data and first-moment information, determine the target scenario type corresponding to the alarm information to be processed;

[0232] Based on the mapping relationship between the target scene type and the preset scene type and multi-dimensional weight information, the target multi-dimensional weight information corresponding to the alarm information to be processed is determined; the multi-dimensional weight information includes voiceprint weight, spatial weight and text weight.

[0233] Based on multidimensional similarity information and target multidimensional weight information, comprehensive matching similarity information between the alarm information to be processed and the associated alarm information is obtained.

[0234] Based on the same inventive concept as the above-described method embodiments, this application also provides an electronic device. This electronic device can be used for repeat alarm detection. In this application embodiment, the electronic device can be a server or a terminal device. In one embodiment, the electronic device can be a terminal device. In this embodiment, the structure of the electronic device can be as follows... Figure 8As shown, it includes a memory 801, a communication module 803, and one or more processors 802.

[0235] The memory 801 is used to store computer programs executed by the processor 802. The memory 801 may mainly include a program storage area and a data storage area. The program storage area may store the operating system and programs required to run instant messaging functions, etc.; the data storage area may store various instant messaging information and operation instruction sets, etc.

[0236] Memory 801 may be volatile memory, such as random-access memory (RAM); memory 801 may also be non-volatile memory, such as read-only memory, flash memory, hard disk drive (HDD), or solid-state drive (SSD); or memory 801 may be any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto. Memory 801 may be a combination of the above-mentioned memories.

[0237] The processor 802 may include one or more central processing units (CPUs) or digital processing units, etc. The processor 802 is used to implement the aforementioned duplicate alarm identification method when it calls the computer program stored in the memory 801.

[0238] The communication module 803 is used to communicate with terminal devices and other servers.

[0239] This application embodiment does not limit the specific connection medium between the memory 801, communication module 803, and processor 802 described above. This application embodiment... Figure 8 The memory 801 and the processor 802 are connected via a bus 804, and the bus 804 is in Figure 8 The connections between other components are indicated by thick lines and are for illustrative purposes only, not as limiting information. The 804 bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, Figure 8 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0240] According to one aspect of this application, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the repeat alarm identification method described in the above embodiments. The program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof.

[0241] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for identifying repeated alarm incidents, characterized in that, The method includes: Acquire multidimensional data information and first-moment information of the alarm information to be processed; the multidimensional data information is alarm information of a preset baseline data type; the baseline data type includes voiceprint data, location data and text data; Based on the first time information and the preset time period determination rule, the associated alarm information corresponding to the alarm information to be processed is determined; the second time information corresponding to the associated alarm information satisfies the time period determination rule with the first time information. The alarm information to be processed and the associated alarm information are respectively used as source information and input into the trained feature extraction network model to obtain multi-dimensional basic features; the multi-dimensional basic features include the first multi-dimensional basic features of the alarm information to be processed and the second multi-dimensional basic features of the associated alarm information. Based on the first multi-dimensional basic features and the second multi-dimensional basic features, comprehensive matching similarity information between the alarm information to be processed and the associated alarm information is determined; the comprehensive matching similarity information is a feature value that characterizes the information similarity between the alarm information to be processed and the associated alarm information. Based on the comprehensive matching similarity information and the preset matching conditions, the duplicate alarm identification result of the alarm information to be processed is determined; The step of determining the comprehensive matching similarity information between the alarm information to be processed and the associated alarm information based on the first multi-dimensional basic features and the second multi-dimensional basic features includes: Based on the first multi-dimensional basic features and the second multi-dimensional basic features, multi-dimensional similarity information is determined; the multi-dimensional similarity information includes voiceprint similarity, spatial similarity and text similarity; Based on the multidimensional data information and the first time information, the target scene type corresponding to the alarm information to be processed is determined; Based on the target scene type and the preset mapping relationship between scene type and multidimensional weight information, the target multidimensional weight information corresponding to the alarm information to be processed is determined; the multidimensional weight information includes voiceprint weight, spatial weight and text weight. Based on the multidimensional similarity information and the target multidimensional weight information, comprehensive matching similarity information between the alarm information to be processed and the associated alarm information is obtained.

2. The method according to claim 1, characterized in that, The step of determining the duplicate alarm identification result of the alarm information to be processed based on the comprehensive matching similarity information and preset matching conditions includes: If the comprehensive matching similarity information is greater than or equal to the preset first matching degree threshold, then the duplicate alarm identification result of the alarm information to be processed is determined to be a duplicate alarm.

3. The method according to claim 2, characterized in that, The step of determining the duplicate alarm identification result of the alarm information to be processed based on the comprehensive matching similarity information and the preset matching conditions further includes: If the comprehensive matching similarity information is less than the preset second matching degree threshold, then the duplicate alarm identification result of the alarm information to be processed is determined to be a new alarm; the second matching degree threshold is less than the first matching degree threshold.

4. The method according to claim 3, characterized in that, The step of determining the duplicate alarm identification result of the alarm information to be processed based on the comprehensive matching similarity information and the preset matching conditions further includes: If the comprehensive matching similarity information is greater than the second matching degree threshold and less than the first matching degree threshold, then the duplicate alarm identification result of the alarm information to be processed is determined as an alarm to be verified; the alarm information to be verified needs to be additionally verified to determine whether the alarm information is duplicated.

5. The method according to claim 1, characterized in that, The feature extraction network model includes a data preprocessing module and a feature extraction module; the alarm information to be processed and the associated alarm information are respectively used as source information and input into the trained feature extraction network model to obtain multi-dimensional basic features, including: Select the alarm information to be processed and the associated alarm information one by one as the source information; For each piece of source information obtained, the current source information is input into the data preprocessing module to obtain intermediate information; The intermediate information is input into the feature extraction module to obtain the multi-dimensional basic features corresponding to the current source information.

6. The method according to claim 5, characterized in that, The data preprocessing module includes a first submodule, a second submodule, and a third submodule; the step of inputting the current source information into the data preprocessing module to obtain intermediate information includes: The first sub-class information in the current source information is input into the first sub-module for sound information standardization to obtain the first sub-information; the sound information standardization operation includes noise reduction; the base data type corresponding to the first sub-class information is the voiceprint data; The second sub-class information from the current source information is input into the second sub-module for geographic information standardization to obtain the second sub-information; the geographic information standardization operation includes coordinate unification, outlier removal, and map matching; the base data type corresponding to the second sub-class information is the location data; The third subclass information in the current source information is input into the third submodule for text information standardization to obtain the third sub-information; the text information standardization operation includes dialect translation; the base data type corresponding to the third subclass information is the text data; The intermediate information is obtained based on the first sub-information, the second sub-information, and the third sub-information.

7. A repeat alarm detection device, characterized in that, include: The basic information acquisition module is used to acquire multi-dimensional data and first-moment information of the alarm information to be processed; The multidimensional data information is alarm information of a preset baseline data type; The baseline data types include voiceprint data, location data, and text data; The associated alarm selection module is used to determine the associated alarm information corresponding to the alarm information to be processed based on the first time information and the preset time period determination rules; The second time information corresponding to the associated alarm information satisfies the time period determination rule with the first time information; The multidimensional feature acquisition module is used to take the alarm information to be processed and the associated alarm information as source information, respectively, and input them into the trained feature extraction network model to obtain multidimensional basic features; the multidimensional basic features include the first multidimensional basic features of the alarm information to be processed and the second multidimensional basic features of the associated alarm information. The multidimensional matching evaluation module is used to determine the comprehensive matching similarity information between the alarm information to be processed and the associated alarm information based on the first multidimensional basic features and the second multidimensional basic features; The comprehensive matching similarity information represents the feature value of the information similarity between the alarm information to be processed and the associated alarm information; The identification result determination module is used to determine the duplicate alarm identification result of the alarm information to be processed based on the comprehensive matching similarity information and preset matching conditions; The multidimensional matching evaluation module is specifically used for: Based on the first multi-dimensional basic features and the second multi-dimensional basic features, multi-dimensional similarity information is determined; the multi-dimensional similarity information includes voiceprint similarity, spatial similarity and text similarity; Based on the multidimensional data information and the first time information, the target scene type corresponding to the alarm information to be processed is determined; Based on the target scene type and the preset mapping relationship between scene type and multidimensional weight information, the target multidimensional weight information corresponding to the alarm information to be processed is determined; The multidimensional weight information includes voiceprint weight, spatial weight, and text weight; Based on the multidimensional similarity information and the target multidimensional weight information, comprehensive matching similarity information between the alarm information to be processed and the associated alarm information is obtained.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, it implements the method of any one of claims 1 to 6.

9. An electronic device, characterized in that, It includes a memory and a processor, wherein the memory stores a computer program that can run on the processor, and when the computer program is executed by the processor, it implements the method of any one of claims 1 to 6.

Images