Baseline detection method, device, equipment, medium and product

Through the pattern recognition model and version determination model automatically handles the baseline detection version, the problem of low baseline detection efficiency in the existing technology is solved, and automated processing and efficiency improvement is achieved.

CN119063783APending Publication Date: 2024-12-03CHINA MOBILE INFORMATION TECHNOLOGY CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411020411.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-29
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

The existing baseline detection methods require manual selection or setting of detection strategies in advance, resulting in low efficiency and a large amount of manpower and material resources.

Method used

By obtaining the object parameters of the object to be detected, using the pattern recognition model and version determination model to automatically process it, obtain the baseline detection version, and then perform baseline detection to avoid manually selecting the detection strategy.

Benefits of technology

It reduces the waste of manpower and material resources, improves the efficiency of baseline detection, and realizes automated processing.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119063783A_ABST
    Figure CN119063783A_ABST
Patent Text Reader

Abstract

The embodiment of the invention provides a baseline detection method and device, equipment, a medium and a product. The method comprises the steps of obtaining object parameters of a to-be-detected object; inputting the object parameters of the to-be-detected object into the mode recognition model, and processing the object parameters of the to-be-detected object through the mode recognition model to obtain a mode type corresponding to the to-be-detected object; inputting the mode type corresponding to the to-be-detected object into a version determination model, and processing the mode type corresponding to the to-be-detected object through the version determination model to obtain a baseline detection version of the to-be-detected object; and performing baseline detection on the to-be-detected object based on the baseline detection version of the to-be-detected object. According to the embodiment of the invention, the waste of manpower and material resources is reduced, and the baseline detection efficiency is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application belongs to the field of baseline detection technology, and in particular, relates to a baseline detection method, device, equipment, medium and product. Background Art

[0002] Baseline testing refers to the multiple, comprehensive measurement and recording of various parameters of a system, environment or individual at a certain point in time or under a certain state, aiming to provide a comparative benchmark and reference for subsequent changes, interferences, and impacts, so that problems can be discovered and analyzed in a timely manner when changes occur. Obviously, baseline testing is particularly important in real-world applications.

[0003] In the prior art, baseline detection is generally performed by collecting configuration information of each module in the device and based on the configuration information of each module and a pre-set detection strategy. However, when performing baseline detection based on this method, it is necessary to manually select or set the corresponding detection strategy in advance, which not only consumes a lot of manpower and material resources, but also leads to low efficiency of baseline detection. Summary of the invention

[0004] The embodiments of the present application provide a baseline detection method, device, equipment, medium and product, which not only reduce the waste of manpower and material resources, but also improve the efficiency of baseline detection.

[0005] In a first aspect, an embodiment of the present application provides a baseline detection method, the method comprising:

[0006] Get the object parameters of the object to be detected;

[0007] Inputting the object parameters of the object to be detected into the pattern recognition model, processing the object parameters of the object to be detected by the pattern recognition model, and obtaining the pattern type corresponding to the object to be detected;

[0008] Inputting the pattern type corresponding to the object to be detected into the version determination model, processing the pattern type corresponding to the object to be detected by the version determination model, and obtaining a baseline detection version of the object to be detected;

[0009] A baseline detection is performed on the object to be detected based on the baseline detection version of the object to be detected.

[0010] In an optional implementation of the first aspect, obtaining object parameters of the object to be detected includes:

[0011] When it is detected that the current moment is in a target detection cycle among the multiple detection cycles, object parameters of the object to be detected in the target detection cycle are obtained.

[0012] In an optional implementation of the first aspect, inputting object parameters of the object to be detected into a pattern recognition model, and processing the object parameters of the object to be detected by the pattern recognition model to obtain a pattern type corresponding to the object to be detected, including:

[0013] Inputting the object parameters of the object to be detected into a pattern recognition model, wherein the pattern recognition model includes a correspondence between parameter value intervals and pattern types, and the correspondence between parameter value intervals and pattern types includes multiple parameter value intervals and a pattern type corresponding to each parameter value interval;

[0014] Determine, by a pattern recognition model, a target parameter value interval to which the object parameter of the object to be detected belongs from a plurality of parameter value intervals based on the correspondence between the parameter value interval and the pattern type and the object parameter of the object to be detected;

[0015] Based on the correspondence between the parameter value interval and the pattern type, the pattern recognition model is used to match the target parameter value interval to obtain the pattern type corresponding to the object to be detected.

[0016] In an optional implementation of the first aspect, the pattern recognition model includes N support vector machine SVM models, each SVM model includes two sub-correspondences, each sub-correspondence includes a parameter value space and a pattern type corresponding to the parameter value space, and the correspondence between the parameter value interval and the pattern type includes 2 N The correspondence between groups is that i is less than or equal to N, and both i and N are positive integers.

[0017] In an optional implementation of the first aspect, inputting a pattern type corresponding to the object to be detected into a version determination model, and processing the pattern type of the object to be detected by the version determination model to obtain a baseline detection version of the object to be detected includes:

[0018] Inputting the pattern type corresponding to the object to be detected into a version determination model, wherein the version determination model includes a mapping relationship between multiple pattern types and baseline detection versions;

[0019] The version determination model is used to match the pattern type corresponding to the object to be detected based on the mapping relationship between multiple pattern types and baseline detection versions, so as to obtain the baseline detection version of the object to be detected.

[0020] In an optional implementation of the first aspect, before inputting the pattern type of the object to be detected into the version determination model, the method further includes:

[0021] Acquire multiple mapping relationship groups, each mapping relationship group includes multiple pending mapping relationships and a relationship label for each pending mapping relationship, each pending mapping relationship includes a pattern type and a pending baseline detection version, pattern type vectors in different pending mapping relationships in each mapping relationship group, and different pending mapping relationships in each mapping relationship group have different pending baseline detection versions, and the relationship label for each pending mapping relationship is used to characterize the degree of matching between the pattern type in the pending mapping relationship and the pending baseline detection version;

[0022] For each mapping relationship group, perform the following steps:

[0023] For each pending mapping relationship, the pattern type in the pending mapping relationship is processed P times based on the first model to obtain P pattern vectors, and the pending baseline detection version in the pending mapping relationship is processed Q times based on the first model to obtain Q version vectors, where P and Q are both positive integers;

[0024] Based on the second model, for each vector group in the P×Q vector groups, calculating the similarity between the pattern vector and the version vector in the vector group, where the P×Q vector groups are determined based on the P pattern vectors and the Q version vectors;

[0025] Determine a total loss score of the pending mapping relationship based on the similarity of each vector group in the similarities of the P×Q vector groups corresponding to the pending mapping relationship, the relationship label of the pending mapping relationship, the P pattern vectors, and the Q version vectors;

[0026] A pending mapping relationship with the smallest total loss score is determined from multiple pending mapping relationships, which is the mapping relationship corresponding to the mapping relationship group.

[0027] In an optional implementation of the first aspect, determining a total loss score of the pending mapping relationship based on the similarity of each vector group in the similarities of the P×Q vector groups corresponding to the pending mapping relationship, the relationship label of the pending mapping relationship, the P pattern vectors, and the Q version vectors includes:

[0028] Calculate P×Q first loss scores based on the similarity of each vector group in the similarities of the P×Q vector groups corresponding to the pending mapping relationship and the relationship label of the pending mapping relationship;

[0029] Determine a second loss score of the pending mapping relationship based on the KL divergence loss values ​​of the P pattern vectors and the KL divergence loss values ​​of the Q version vectors;

[0030] Based on the P×Q first loss scores and the second loss scores, a total loss score of the pending mapping relationship is determined.

[0031] In an optional implementation of the first aspect, before inputting the object parameters of the object to be detected into the pattern recognition model, the method further includes:

[0032] Acquire a training sample set, the training sample set including a plurality of object parameter samples and a sample label corresponding to each object parameter sample, the sample label corresponding to each object parameter sample being used to indicate a mode type corresponding to the object parameter sample;

[0033] For each object parameter sample in the plurality of object parameter samples, perform the following steps:

[0034] Inputting the object parameter sample into a preset pattern recognition model, the preset pattern recognition model includes a correspondence between a preset parameter value interval and a pattern type, and the correspondence between the preset parameter value interval and the pattern type includes a plurality of preset parameter value intervals and a pattern type corresponding to each preset parameter value interval;

[0035] Determine, by a preset pattern recognition model, a reference parameter value interval to which the object parameter sample belongs from a plurality of parameter value intervals based on a preset correspondence between the parameter value interval and the pattern type and the object parameter sample;

[0036] By using a preset pattern recognition model, based on the correspondence between the preset parameter value interval and the pattern type, the reference parameter value interval is matched to determine the pattern type of the object parameter sample;

[0037] Processing the pattern type corresponding to the target object parameter sample and the sample label corresponding to the target training sample based on the loss function to obtain a loss function value of a preset pattern recognition model, wherein the target object parameter sample is any one of the multiple object parameter samples;

[0038] Based on the loss function value of the preset pattern recognition model, the preset pattern recognition model is trained using multiple object parameter samples to obtain a trained pattern recognition model.

[0039] In a second aspect, an embodiment of the present application provides a baseline detection device, the device comprising:

[0040] An acquisition module, used for acquiring object parameters of the object to be detected;

[0041] A processing module, used for inputting the object parameters of the object to be detected into the pattern recognition model, processing the object parameters of the object to be detected by the pattern recognition model, and obtaining the pattern type corresponding to the object to be detected;

[0042] The processing module is further used to input the pattern type corresponding to the object to be detected into the version determination model, and process the pattern type corresponding to the object to be detected through the version determination model to obtain a baseline detection version of the object to be detected;

[0043] The baseline detection module is used to perform a baseline detection on the object to be detected based on the baseline detection version of the object to be detected.

[0044] In a third aspect, an electronic device is provided, comprising: a memory for storing computer program instructions; and a processor for reading and running the computer program instructions stored in the memory to execute the baseline detection method provided by any optional implementation in the first aspect.

[0045] In a fourth aspect, a computer storage medium is provided, on which computer program instructions are stored, and when the computer program instructions are executed by a processor, the baseline detection method provided by any optional implementation in the first aspect is implemented.

[0046] In a fifth aspect, a computer program product is provided. The computer program product includes a computer program. When the computer program is executed by a processor, the baseline detection method provided by any optional implementation in the first aspect is implemented.

[0047] In an embodiment of the present application, it is possible to obtain the object parameters of the object to be detected, process the object parameters of the object to be detected through a pattern recognition model, obtain the pattern type corresponding to the object to be detected, and process the pattern type corresponding to the object to be detected through a version determination model to obtain a baseline detection version of the object to be detected, and then perform a baseline detection on the object to be detected based on the baseline detection version of the object to be detected. In this way, after obtaining the object parameters of the object to be detected, it is possible to automatically process through a pattern recognition model and a version determination model to obtain a baseline detection version for performing a baseline detection on the object to be detected, thereby avoiding the method of manually selecting or setting the corresponding detection strategy in advance, thereby not only reducing the waste of manpower and material resources, but also improving the efficiency of baseline detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solution of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0049] Figure 1 This is one of the flowcharts of a baseline detection method provided in an embodiment of the present application;

[0050] Figure 2 This is the second flow chart of a baseline detection method provided in an embodiment of the present application;

[0051] Figure 3 This is a flowchart of a baseline detection method provided in an embodiment of the present application;

[0052] Figure 4 This is a fourth flowchart of a baseline detection method provided in an embodiment of the present application;

[0053] Figure 5 It is a flow chart of a method for training a pattern recognition model provided in an embodiment of the present application;

[0054] Figure 6 is a structural schematic diagram of a baseline detection device provided in an embodiment of the present application;

[0055] Figure 7 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0056] The features and exemplary embodiments of various aspects of the present application will be described in detail below. In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, rather than to limit the present application. For those skilled in the art, the present application can be implemented without the need for some of these specific details. The following description of the embodiments is only to provide a better understanding of the present application by illustrating the examples of the present application.

[0057] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "include..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0058] The term "and / or" in this article is merely a description of the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone.

[0059] As the background technology, in order to solve the problem of low efficiency of baseline detection in the prior art, the embodiments of the present application provide a baseline detection method, device, equipment, medium and product, which can obtain the object parameters of the object to be detected, process the object parameters of the object to be detected through the pattern recognition model, obtain the pattern type corresponding to the object to be detected, and process the pattern type corresponding to the object to be detected through the version determination model to obtain the baseline detection version of the object to be detected, and then the baseline detection of the object to be detected can be performed based on the baseline detection version of the object to be detected. In this way, after obtaining the object parameters of the object to be detected, it can be automatically processed through the pattern recognition model and the version determination model to obtain the baseline detection version used for the baseline detection of the object to be detected, avoiding the method of manually selecting or setting the corresponding detection strategy in advance, thereby not only reducing the waste of manpower and material resources, but also improving the efficiency of baseline detection.

[0060] It should be noted that the execution subject of a baseline detection method provided in an embodiment of the present application may be a baseline detection device, or a control module in a baseline detection device for executing the baseline detection method. The present embodiment of the present application takes the baseline detection method executed by a baseline detection device as an example to introduce the baseline detection method provided in an embodiment of the present application in detail.

[0061] The baseline detection method provided in the embodiment of the present application is described in detail below through specific examples in conjunction with the accompanying drawings.

[0062] Figure 1 It is a flowchart of a baseline detection method provided in an embodiment of the present application.

[0063] like Figure 1 As shown, the execution subject of the method may be a baseline detection device. Based on this, the baseline detection method may specifically include the following steps:

[0064] S110, obtaining object parameters of the object to be detected.

[0065] Among them, the objects to be detected can be objects such as power equipment, databases, network equipment, middleware, and hosts, and the embodiments of the present application do not make specific limitations on this.

[0066] Based on this, the object parameter of the object to be detected can be the relevant parameter information generated during the operation of the above object, which is not specifically limited here. It should be noted that the object parameter of the object to be detected can include one parameter or multiple parameters, which is not specifically limited here.

[0067] S120, inputting the object parameters of the object to be detected into a pattern recognition model, and processing the object parameters of the object to be detected by the pattern recognition model to obtain a pattern type of the object to be detected.

[0068] The pattern recognition model may be a pre-trained related model for determining the pattern type of baseline detection, which is not specifically limited here.

[0069] Specifically, after acquiring the object parameters of the object to be detected, the baseline detection device can input the object parameters of the object to be detected into the pattern recognition model, and process the object parameters of the object to be detected through the pattern recognition model to obtain the pattern type corresponding to the object to be detected.

[0070] S130, inputting the pattern type corresponding to the object to be detected into a version determination model, processing the pattern type corresponding to the object to be detected by the version determination model, and obtaining a baseline detection version of the object to be detected.

[0071] Specifically, after obtaining the pattern type corresponding to the object to be detected, the baseline detection device can input the pattern type corresponding to the object to be detected into a version determination model, process the pattern type corresponding to the object to be detected through the version determination model, and obtain a baseline detection version of the object to be detected.

[0072] S140, performing a baseline detection on the object to be detected based on the baseline detection version of the object to be detected.

[0073] Specifically, after the pattern type corresponding to the object to be detected is processed through the version determination model to obtain a baseline detection version of the object to be detected, the baseline detection version can perform baseline detection processing on the object to be detected based on the baseline detection version of the object to be detected.

[0074] In an embodiment of the present application, it is possible to obtain the object parameters of the object to be detected, process the object parameters of the object to be detected through a pattern recognition model, obtain the pattern type corresponding to the object to be detected, and process the pattern type corresponding to the object to be detected through a version determination model to obtain a baseline detection version of the object to be detected, and then perform a baseline detection on the object to be detected based on the baseline detection version of the object to be detected. In this way, after obtaining the object parameters of the object to be detected, it is possible to automatically process through a pattern recognition model and a version determination model to obtain a baseline detection version for performing a baseline detection on the object to be detected, thereby avoiding the method of manually selecting or setting the corresponding detection strategy in advance, thereby not only reducing the waste of manpower and material resources, but also improving the efficiency of baseline detection.

[0075] In order to obtain the object parameters of the object to be processed so as to accurately determine the baseline detection version to perform baseline detection on the object to be detected, in one embodiment, the step of obtaining the object parameters of the object to be detected may specifically include the following steps:

[0076] Acquire dynamic signals of the object to be detected;

[0077] Preprocessing the dynamic signal of the object to be detected to obtain the preprocessed dynamic signal of the object to be detected;

[0078] The object parameters of the object to be detected are extracted from the preprocessed dynamic signal of the object to be detected.

[0079] The dynamic signal of the object to be detected may be a signal generated by the object to be detected during actual operation, which is not specifically limited here.

[0080] Specifically, the baseline detection device can obtain the dynamic signal of the object to be detected, and preprocess the dynamic signal of the object to be detected to eliminate the noise and other interference factors existing in the dynamic signal of the object to be detected, and obtain the preprocessed dynamic signal of the object to be detected, and then the object parameters of the object to be detected can be extracted from the preprocessed dynamic signal of the object to be detected.

[0081] In this embodiment, the dynamic signal of the object to be detected can be obtained and preprocessed to extract the object parameters of the object to be detected from the preprocessed dynamic signal of the object to be detected. In this way, the acquired dynamic signal of the object to be detected can be preprocessed to eliminate noise and other interference factors in the dynamic signal, thereby improving the accuracy of obtaining the object parameters of the object to be detected.

[0082] Considering that the object parameters of the object to be detected are not fixed but constantly changing, in order to ensure that the most appropriate baseline detection version is used to perform baseline detection on the object to be detected in different time periods, in one embodiment, the above S110 may specifically include the following steps:

[0083] When it is detected that the current moment is in a target detection cycle among the multiple detection cycles, object parameters of the object to be detected in the target detection cycle are obtained.

[0084] The target detection period may be any one of the multiple detection periods and is not specifically limited here.

[0085] Specifically, the baseline detection device can set multiple detection cycles, and when it is detected that the current moment is in a target detection cycle among the multiple detection cycles, obtain the object parameters of the object to be detected in the target detection cycle.

[0086] It should be noted that each of the multiple detection cycles may include a start time and an end time, and the duration from the start time to the end time is the duration of the detection cycle. The duration of each detection cycle can be pre-set based on actual experience or circumstances, and is not specifically limited here.

[0087] Based on this, it is also necessary to explain that the baseline detection device can set multiple detection cycles, use time calibration for the start and end times of each detection cycle, and set corresponding labels for each detection cycle based on the time sequence. In this way, the target detection cycle that the current moment falls into can be determined based on the current moment and the start and end times of each detection cycle, and it can be determined whether the target detection cycle is the same as the label of the target cycle detected last time. If they are the same, no processing is performed. If they are different, the object parameters of the object to be detected within the target cycle are re-acquired.

[0088] In this embodiment, when it is detected that the current moment is in a target detection cycle among multiple detection cycles, the object parameters of the object to be detected in the target detection cycle can be obtained. In this way, it can be ensured that the most appropriate baseline detection version is used to perform baseline detection on the object to be detected in different time periods.

[0089] In order to accurately obtain the pattern type corresponding to the object to be detected, in one embodiment, Figure 2 As shown, the above S120 may specifically include the steps of:

[0090] S210: Inputting object parameters of the object to be detected into a pattern recognition model.

[0091] In some embodiments, the pattern recognition model may include a correspondence between parameter value intervals and pattern types. The correspondence may include multiple parameter value intervals and pattern types corresponding to each parameter value interval. The parameter value interval refers to the value interval of the object parameter, which is not limited here.

[0092] S220, determining a target parameter value interval to which the object parameter of the object to be detected belongs from multiple parameter value intervals through a pattern recognition model based on the correspondence between the parameter value interval and the pattern type and the object parameter of the object to be detected.

[0093] Specifically, after the object parameters of the object to be detected are input into the pattern recognition model, since the pattern recognition model can include a correspondence between parameter value intervals and pattern types, and the correspondence can include multiple parameter value intervals, the baseline detection device can determine the target parameter value interval to which the object parameters of the object to be detected belong from multiple parameter value intervals based on the correspondence between the parameter value intervals and the pattern type and the object parameters of the object to be detected.

[0094] S230, matching the target parameter interval based on the correspondence between the parameter value interval and the pattern type through the pattern recognition model to obtain the pattern type corresponding to the object to be detected.

[0095] Specifically, after determining the target parameter value interval to which the object parameter of the object to be detected belongs, the target parameter value interval to which the object parameter of the object to be detected belongs can be matched through the pattern recognition model based on the correspondence between the parameter value interval and the pattern type to obtain the corresponding pattern type of the object to be detected.

[0096] In this embodiment, since the pattern recognition model includes the correspondence between the parameter value interval and the pattern type, the target parameter value interval to which the object parameter of the object to be detected belongs can be first determined by the pattern recognition model based on the correspondence between the parameter value interval and the pattern type and the object parameter of the object to be detected, and then the pattern type corresponding to the object to be detected can be obtained. In this way, when the object parameter of the object to be detected is input into the pattern recognition model, the object parameter of the object to be detected can be automatically identified by the pattern recognition model, thereby improving the efficiency of obtaining the pattern type corresponding to the object to be detected.

[0097] In order to describe the baseline detection method provided by the embodiment of the present application in more detail, in one embodiment, the above-mentioned pattern recognition model may include N layer structures, and the i-th layer structure may include 2 i-1 A Support Vector Machine (SVM) model.

[0098] Since each SVM model is a two-class classifier, each SVM model may include two sub-correspondences, and each sub-correspondence includes a parameter value space and a pattern type corresponding to the parameter value space.

[0099] Thus, the Nth layer structure can include N-1 SVM models, and 2 N Therefore, the correspondence between the above parameter value range and mode type can include 2 N The corresponding relationship of groups, where i is less than or equal to N, and both i and N are positive integers.

[0100] In this embodiment, since the pattern recognition model may include N layer structures, the i-th layer structure may include 2 i SVM model, and SVM model is a two-class classifier, so the pattern recognition model can finally get 2 N In this way, the corresponding relationship between the parameter value range and the mode type can be accurately obtained.

[0101] In order to accurately obtain the baseline detection version of the object to be detected, in some embodiments, the version determination model may include a mapping relationship between multiple pattern types and baseline detection versions. Figure 3 As shown, the above S130 may include the following steps:

[0102] S310: Input the pattern type corresponding to the object to be detected into the version determination model.

[0103] In some embodiments, the version determination model may include a mapping relationship between multiple mode types and baseline detection versions.

[0104] S320, matching the pattern type corresponding to the object to be detected based on the mapping relationship between multiple pattern types and baseline detection versions through the version determination model to obtain the baseline detection version of the object to be detected.

[0105] Specifically, since the version determination model can include a mapping relationship between multiple pattern types and baseline detection versions, the baseline detection device can input the pattern type corresponding to the object to be detected into the version determination model after obtaining the pattern type corresponding to the object to be detected, and then the version determination model can be used to match the pattern type of the object to be detected based on the mapping relationship between the multiple pattern types and the baseline detection version included in the version determination model to obtain the baseline detection version of the object to be detected.

[0106] In this embodiment, the mapping relationship between multiple pattern types and baseline detection versions included in the version determination model can be determined. After the pattern type corresponding to the object to be detected is obtained, the pattern type corresponding to the object to be detected is automatically matched to obtain the baseline detection version of the object to be detected, thereby improving the efficiency of obtaining the basic detection version of the object to be detected.

[0107] Since the version determination model may include mapping relationships between multiple mode types and baseline detection versions, based on this, in order to accurately obtain the mapping relationships between multiple mode types and baseline detection versions, in one embodiment, Figure 4 As shown, before the pattern type of the object to be detected is input into the version determination model, the above-mentioned other methods may also include the following steps:

[0108] S410: Acquire multiple mapping relationship groups.

[0109] In some embodiments, each mapping relationship group may include multiple pending mapping relationships and a relationship tag for each pending mapping relationship. Each pending mapping relationship may include a pattern type and a pending baseline detection version. It should be noted that the pattern types in different pending mapping relationships in each mapping relationship group are the same, and the pending baseline detection versions in different pending mapping relationships in each mapping relationship group are different.

[0110] In addition, the relationship label of each pending mapping relationship can be used to characterize the degree of match between the pattern type in the pending mapping relationship and the pending baseline detection version. Exemplarily, if the relationship label of a pending mapping relationship is 0, it can be characterized that the pattern type in the pending mapping relationship in the mapping relationship does not match the pending baseline detection version; if the relationship label of a pending mapping relationship is 1, it can be characterized that the pattern type in the pending mapping relationship matches the pending baseline detection version. It should be noted that the specific label content of the above-mentioned relationship label can also be represented by other numerical values ​​that can characterize the degree of match between the pattern type and the pending baseline detection version. The embodiments of the present application do not specifically limit the specific label content of the relationship label.

[0111] S420, for each pending mapping relationship in each mapping relationship group, the pattern type in the pending mapping relationship is processed P times based on the first model to obtain P pattern vectors, and the pending baseline detection version in the pending mapping relationship is processed Q times based on the first model to obtain Q version vectors.

[0112] Among them, the above-mentioned first model can be a Transformer-based bidirectional encoding representation (BERT, Bidirectional Encoder Representations from Transformers) model, which is not specifically limited here.

[0113] In addition, the P pattern vectors may be different vectors, and correspondingly, the Q version vectors may be Q different version vectors. It should also be noted that the specific values ​​of P and Q may be the same or different, and are not specifically limited here. P and Q are both positive integers.

[0114] Specifically, after acquiring multiple mapping relationship groups, the baseline detection device can, for each pending mapping relationship in each mapping relationship group, where the pending mapping relationship includes a pattern type and a pending baseline detection version, perform P processing on the pattern type in the pending mapping relationship based on the first model to obtain P pattern vectors, and perform Q processing on the pending baseline detection version in the pending mapping relationship based on the first model to obtain Q version vectors.

[0115] It should be noted that in the process of obtaining P pattern vectors, the pattern type can be segmented into words to obtain a feature list corresponding to the pattern type, and the word segmenter of the BERT model is used to convert the feature list corresponding to the pattern type into a feature word code corresponding to the pattern type, and the feature word code corresponding to the pattern type is input into the BERT model P times. The BERT model randomly removes some neural units each time, and finally obtains P different pattern vectors. It should also be noted that the process of determining Q version vectors is similar to the process of determining P pattern vectors mentioned above, and will not be elaborated here.

[0116] S430 , based on the second model, for each vector group in the P×Q vector groups, calculate the similarity between the pattern vector and the version vector in the vector group.

[0117] The second model may be an Enhanced Sequential Inference Model (ESIM), which is not specifically limited here.

[0118] In some embodiments, P×Q vector groups are determined based on P pattern vectors and Q version vectors. For example, if the baseline detection device can obtain two pattern vectors, namely X1 and X2, and two version vectors, namely Y1 and Y2, the baseline detection device can obtain four vector groups, namely (X1, Y1), (X1, Y2), (X2, Y1), and (X2, Y2).

[0119] Specifically, after obtaining P pattern vectors and Q version vectors, the baseline detection device can obtain P×Q vector groups based on the P pattern vectors and Q version vectors, and then calculate the similarity of the pattern vector and the version vector in each of the P×Q vector groups through a second model.

[0120] S440, determining a total loss score of the pending mapping relationship based on the similarity of each vector group among the similarities of the P×Q vector groups corresponding to the pending mapping relationship, the relationship label of the pending mapping relationship, the P pattern vectors, and the Q version vectors.

[0121] The total loss score of the above-mentioned pending mapping relationships is used to select the most appropriate mapping relationship from each mapping relationship group.

[0122] Specifically, after determining the similarity between the pattern vector and the version vector in each vector group, the baseline detection device can determine the total loss score of the pending mapping relationship based on the similarity of each vector group in the P×Q vector groups corresponding to the pending mapping relationship, the relationship label of the pending mapping relationship, the P pattern vectors and the Q version vectors.

[0123] S450, determining a pending mapping relationship with a minimum total loss score from a plurality of pending mapping relationships, as the mapping relationship corresponding to the mapping relationship group.

[0124] Specifically, after obtaining the total loss score of each pending mapping relationship in a plurality of pending mapping relationships, the baseline detection device can determine the pending mapping relationship with the smallest total loss score from the plurality of pending mapping relationships based on the total loss score of each pending mapping relationship, which is the mapping relationship corresponding to the mapping relationship group.

[0125] In this embodiment, multiple mapping relationship groups can be obtained, and the multiple mapping relationships include multiple pending mapping relationships and a relationship label for each pending mapping relationship. Each pending mapping relationship includes a pattern type and a pending baseline detection version. Different pending mapping relationships in each mapping relationship group have the same pattern type and different pending baseline detection versions. In this way, by calculating the total loss score of each pending mapping relationship for each mapping relationship, the mapping relationship corresponding to the mapping relationship group can be accurately determined from the multiple pending mapping relationships, that is, the baseline detection version that is most suitable for the current pattern type can be determined from the multiple pending baseline detection versions, thereby improving the accuracy of obtaining the mapping relationships between the multiple pattern types and the baseline detection versions included in the version determination model.

[0126] In order to accurately calculate the total loss score, so as to accurately determine the mapping relationship in the version determination model, in one embodiment, the above S440 may specifically include the following steps:

[0127] Calculate P×Q first loss scores based on the similarity of each vector group in the similarities of the P×Q vector groups corresponding to the pending mapping relationship and the relationship label of the pending mapping relationship;

[0128] Determine a second loss score of the pending mapping relationship based on the KL divergence loss values ​​of the P pattern vectors and the KL divergence loss values ​​of the Q version vectors;

[0129] Based on the P×Q first loss scores and the second loss scores, a total loss score of the pending mapping relationship is determined.

[0130] Specifically, the baseline detection device can calculate P×Q first loss scores based on the similarity of each vector group among the P×Q vector groups corresponding to the pending mapping relationship and the relationship label of the pending mapping relationship, and can determine the second loss score of the pending mapping relationship based on the KL divergence loss value of the P pattern vectors and the KL divergence loss value of the Q version vectors, and then can determine the total loss score of the pending mapping relationship based on the first loss score and the second loss score.

[0131] It should be noted that, in the process of calculating the first loss score, the cross entropy loss function can be used to process the similarity of each vector group and the relationship label of the pending mapping relationship to obtain the corresponding first loss score. In the process of calculating the second loss score, the KL (Kullback-Leibler) loss function can be used to process the P pattern vectors to obtain the KL divergence loss value of the P pattern vectors, and the KL loss score can also be used to process the Q version vectors. The KL loss value of the Q version vectors is not specifically limited here.

[0132] In one embodiment, determining the total loss score of the pending mapping relationship based on the P×Q first loss scores and the second loss scores includes:

[0133] Based on the preset weights, a weighted sum of P×Q first loss scores and second loss scores is performed to obtain a total loss score of the pending mapping relationship.

[0134] The preset weights may be preset based on actual experience or circumstances, and are not specifically limited here.

[0135] Specifically, after obtaining P×Q first loss scores and second loss scores corresponding to the pending mapping relationship, the baseline detection device can perform weighted summation on the P×Q first loss scores and second loss scores based on preset weights to obtain the total loss score of the pending mapping relationship.

[0136] In this embodiment, P×Q first loss scores can be obtained based on the similarity of each vector group in the similarities corresponding to the pending mapping relationship and the relationship label of the pending mapping relationship, and the second loss score of the pending mapping relationship can be determined based on the KL divergence loss value of the P mode vector and the KL divergence loss value of the Q version vectors. Furthermore, the total loss score of the pending mapping relationship can be accurately determined based on the P×Q first loss scores and the second loss scores, so that the mapping relationship corresponding to each mapping relationship group can be accurately determined subsequently.

[0137] It should be noted that the baseline detection method provided in the embodiment of the present application needs to use a pre-trained pattern recognition model to process the object parameters of the object to be detected. Therefore, before S120, the pattern recognition model needs to be trained first. Therefore, the specific implementation method of the pattern recognition model training method provided in the embodiment of the present application is described below in conjunction with the accompanying drawings.

[0138] Figure 5 It is a flow chart of a method for training a pattern recognition model provided in an embodiment of the present application.

[0139] like Figure 5 As shown, the execution subject of the method may be a baseline detection device, based on which, the method may include the following steps:

[0140] S510, obtaining a training sample set.

[0141] Before introducing this step, it should be noted that the service quality evaluation model needs to be iterated multiple times to adjust its loss function value until the loss function meets the training stop condition and obtains the trained pattern recognition model. However, in each iterative training, if only one training sample is input, the sample size is too small to use the training adjustment of the pattern recognition model, that is, in the process of training the pattern recognition model, multiple training samples are required to iteratively process the pattern recognition model.

[0142] Therefore, the above training sample set may include multiple object parameter samples and a sample label corresponding to each object parameter sample, and the sample label corresponding to each object parameter sample is used to indicate the mode type corresponding to the object parameter sample.

[0143] S520: For each object parameter sample among the multiple object parameter samples, input the object parameter sample into a preset pattern recognition model.

[0144] The preset pattern recognition model may include a correspondence between a preset parameter value interval and a pattern type, and the correspondence between a preset parameter value interval and a pattern type may include multiple preset parameter value intervals and a pattern type corresponding to each preset parameter value interval. The preset parameter value interval may be a preset value interval of an object parameter, which is not specifically limited here.

[0145] S530, determining a reference parameter value interval to which the object parameter sample belongs from a plurality of parameter value intervals through a preset pattern recognition model based on a preset correspondence between the parameter value interval and the pattern type and the object parameter sample.

[0146] Specifically, after the baseline detection device inputs the object parameter sample into the preset pattern recognition model, since the preset pattern recognition model includes a correspondence between a preset parameter value interval and a pattern type, and the correspondence between the preset parameter value interval and the pattern type may include multiple preset parameter value intervals, the baseline detection device can determine the reference parameter value interval to which the object parameter sample belongs from multiple preset parameter value intervals based on the preset pattern recognition model and the correspondence between the preset parameter value interval and the pattern type and the object parameter sample.

[0147] S540, matching the reference parameter value interval based on the preset correspondence between the parameter value interval and the pattern type through the preset pattern recognition model to determine the pattern type of the object parameter sample.

[0148] Specifically, after determining the reference parameter value interval to which the object parameter sample belongs, the reference parameter value interval can be matched through a preset pattern recognition model based on the correspondence between the preset parameter value interval and the pattern type to determine the pattern type of the object parameter sample.

[0149] S550, processing the pattern type corresponding to the target object parameter sample and the sample label corresponding to the target training sample based on the loss function to obtain the loss function value of the preset pattern recognition model.

[0150] The target object parameter sample is any one of multiple object parameter samples.

[0151] Specifically, the baseline detection device can process the pattern type corresponding to the target object parameter sample and the sample label corresponding to the target training sample based on the loss function to determine the loss function value of the preset pattern recognition model.

[0152] S560, based on the loss function value of the preset pattern recognition model, use multiple object parameter samples to train the preset pattern recognition model to obtain a trained pattern recognition model.

[0153] Specifically, in order to obtain a better trained pattern recognition model, when the target loss function value does not meet the training stop condition, the model parameters of the preset pattern recognition model are adjusted, and the pattern recognition model with adjusted parameters is continuously trained using multiple training samples until the loss function value meets the training stop condition, thereby obtaining the trained pattern recognition model. The training stop condition may be determined based on actual experience and is not specifically limited here.

[0154] In this embodiment, a training sample set can be obtained, which includes multiple object parameter samples and a sample label corresponding to each object parameter sample. For each object parameter sample, the object parameter sample is input into a preset pattern recognition model, and the object parameter sample is processed by the preset pattern recognition model to obtain the pattern type of the object parameter sample. Then, based on the pattern type and sample label of the object parameter sample, the loss function value of the preset pattern recognition model can be obtained. Then, based on the loss function value of the preset pattern recognition model, multiple object parameter samples can be utilized.

[0155] Based on the same inventive concept, the present application embodiment also provides a baseline detection device. Figure 6 The baseline detection device provided in the embodiment of the present application is described in detail.

[0156] Figure 6 It is a structural schematic diagram of a baseline detection device provided in an embodiment of the present application.

[0157] like Figure 6 As shown, the baseline detection device 600 may include:

[0158] An acquisition module 610 is used to acquire object parameters of the object to be detected;

[0159] The processing module 620 is used to input the object parameters of the object to be detected into the pattern recognition model, and process the object parameters of the object to be detected through the pattern recognition model to obtain the pattern type corresponding to the object to be detected;

[0160] The processing module 620 is further used to input the pattern type corresponding to the object to be detected into the version determination model, and process the pattern type corresponding to the object to be detected through the version determination model to obtain a baseline detection version of the object to be detected;

[0161] The baseline detection module 630 is used to perform a baseline detection on the object to be detected based on the baseline detection version of the object to be detected.

[0162] In one embodiment, the acquisition module is specifically used to:

[0163] When it is detected that the current moment is in a target detection cycle among the multiple detection cycles, object parameters of the object to be detected in the target detection cycle are obtained.

[0164] In one embodiment, the processing module mentioned above is specifically used for:

[0165] Inputting the object parameters of the object to be detected into a pattern recognition model, wherein the pattern recognition model includes a correspondence between parameter value intervals and pattern types, and the correspondence between parameter value intervals and pattern types includes multiple parameter value intervals and a pattern type corresponding to each parameter value interval;

[0166] Determine, by a pattern recognition model, a target parameter value interval to which the object parameter of the object to be detected belongs from a plurality of parameter value intervals based on the correspondence between the parameter value interval and the pattern type and the object parameter of the object to be detected;

[0167] Based on the correspondence between the parameter value interval and the pattern type, the pattern recognition model is used to match the target parameter value interval to obtain the pattern type corresponding to the object to be detected.

[0168] In some embodiments, the pattern recognition model includes N support vector machine SVM models, each SVM model includes two sub-correspondences, each sub-correspondence includes a parameter value space and a pattern type corresponding to the parameter value space, and the correspondence between the parameter value interval and the pattern type includes 2 N The correspondence between groups is that i is less than or equal to N, and both i and N are positive integers.

[0169] In one embodiment, the processing module is specifically used to:

[0170] Inputting the pattern type corresponding to the object to be detected into a version determination model, wherein the version determination model includes a mapping relationship between multiple pattern types and baseline detection versions;

[0171] The version determination model is used to match the pattern type corresponding to the object to be detected based on the mapping relationship between multiple pattern types and baseline detection versions, so as to obtain the baseline detection version of the object to be detected.

[0172] In one embodiment, the baseline detection device may further include a calculation module and a determination module.

[0173] The acquisition module is further used to acquire multiple mapping relationship groups, each mapping relationship group includes multiple pending mapping relationships and a relationship label of each pending mapping relationship, each pending mapping relationship includes a pattern type and a pending baseline detection version, the pattern type vectors in different pending mapping relationships in each mapping relationship group, and the pending baseline detection versions in different pending mapping relationships in each mapping relationship group are different, and the relationship label of each pending mapping relationship is used to characterize the matching degree between the pattern type in the pending mapping relationship and the pending baseline detection version;

[0174] The processing module is further used to process the pattern type in the pending mapping relationship P times based on the first model for each pending mapping relationship in each mapping relationship group to obtain P pattern vectors, and process the pending baseline detection version in the pending mapping relationship Q times based on the first model to obtain Q version vectors, where P and Q are both positive integers;

[0175] A calculation module, configured to calculate, based on the second model, for each vector group in the P×Q vector groups, a similarity between a pattern vector and a version vector in the vector group, where the P×Q vector groups are determined based on the P pattern vectors and the Q version vectors;

[0176] A determination module, configured to determine a total loss score of the pending mapping relationship based on the similarity of each vector group among the similarities of the P×Q vector groups corresponding to the pending mapping relationship, the relationship label of the pending mapping relationship, the P pattern vectors, and the Q version vectors;

[0177] The determination module is also used to determine the pending mapping relationship with the smallest total loss score from multiple pending mapping relationships, which is the mapping relationship corresponding to the mapping relationship group.

[0178] In one embodiment, the above-mentioned determination module is specifically used to:

[0179] Calculate P×Q first loss scores based on the similarity of each vector group in the similarities of the P×Q vector groups corresponding to the pending mapping relationship and the relationship label of the pending mapping relationship;

[0180] Determine a second loss score of the pending mapping relationship based on the KL divergence loss values ​​of the P pattern vectors and the KL divergence loss values ​​of the Q version vectors;

[0181] Based on the P×Q first loss scores and the second loss scores, a total loss score of the pending mapping relationship is determined.

[0182] In one embodiment, the baseline detection device may further include a training module. The training module is specifically used to:

[0183] Acquire a training sample set, the training sample set including a plurality of object parameter samples and a sample label corresponding to each object parameter sample, the sample label corresponding to each object parameter sample being used to indicate a mode type corresponding to the object parameter sample;

[0184] For each object parameter sample in the plurality of object parameter samples, perform the following steps:

[0185] Inputting the object parameter sample into a preset pattern recognition model, the preset pattern recognition model includes a correspondence between a preset parameter value interval and a pattern type, and the correspondence between the preset parameter value interval and the pattern type includes a plurality of preset parameter value intervals and a pattern type corresponding to each preset parameter value interval;

[0186] Determine, by a preset pattern recognition model, a reference parameter value interval to which the object parameter sample belongs from a plurality of parameter value intervals based on a preset correspondence between the parameter value interval and the pattern type and the object parameter sample;

[0187] By using a preset pattern recognition model, based on the correspondence between the preset parameter value interval and the pattern type, the reference parameter value interval is matched to determine the pattern type of the object parameter sample;

[0188] Processing the pattern type corresponding to the target object parameter sample and the sample label corresponding to the target training sample based on the loss function to obtain a loss function value of a preset pattern recognition model, wherein the target object parameter sample is any one of the multiple object parameter samples;

[0189] Based on the loss function value of the preset pattern recognition model, the preset pattern recognition model is trained using multiple object parameter samples to obtain a trained pattern recognition model.

[0190] In an embodiment of the present application, it is possible to obtain the object parameters of the object to be detected, process the object parameters of the object to be detected through a pattern recognition model, obtain the pattern type corresponding to the object to be detected, and process the pattern type corresponding to the object to be detected through a version determination model to obtain a baseline detection version of the object to be detected, and then perform a baseline detection on the object to be detected based on the baseline detection version of the object to be detected. In this way, after obtaining the object parameters of the object to be detected, it is possible to automatically process through a pattern recognition model and a version determination model to obtain a baseline detection version for performing a baseline detection on the object to be detected, thereby avoiding the method of manually selecting or setting the corresponding detection strategy in advance, thereby not only reducing the waste of manpower and material resources, but also improving the efficiency of baseline detection.

[0191] Each module in the baseline detection device provided in the embodiment of the present application can be implemented Figures 1 to 5 The method steps of any of the embodiments shown in the figure can achieve the corresponding technical effects, which will not be repeated here for the sake of brevity.

[0192] Figure 7 A schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present application is shown.

[0193] The electronic device may include a processor 701 and a memory 702 storing computer program instructions.

[0194] Specifically, the processor 701 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0195] The memory 702 may include a large capacity memory for data or instructions. By way of example and not limitation, the memory 702 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive or a combination of two or more of these. In appropriate cases, the memory 702 may include a removable or non-removable (or fixed) medium. In appropriate cases, the memory 702 may be inside or outside the electronic device. In a particular embodiment, the memory 702 is a non-volatile solid-state memory.

[0196] The memory may include read-only memory (ROM), random access memory (RAM), magnetic disk storage media devices, optical storage media devices, flash memory devices, electrical, optical or other physical / tangible memory storage devices. Thus, typically, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method according to an aspect of the present disclosure.

[0197] The processor 701 implements any one of the baseline detection methods in the above embodiments by reading and executing computer program instructions stored in the memory 702 .

[0198] In one example, the electronic device may further include a communication interface 703 and a bus 710. Figure 7 As shown, the processor 701, the memory 702, and the communication interface 703 are connected via a bus 710 and communicate with each other.

[0199] The communication interface 703 is mainly used to implement communication between various modules, devices, units and / or equipment in the embodiments of the present application.

[0200] Bus 710 includes hardware, software or both, and the parts of online data flow billing equipment are coupled to each other. For example, but not limitation, bus may include accelerated graphics port (AGP) or other graphics bus, enhanced industrial standard architecture (EISA) bus, front-end bus (FSB), hypertransport (HT) interconnection, industrial standard architecture (ISA) bus, infinite bandwidth interconnection, low pin count (LPC) bus, memory bus, micro channel architecture (MCA) bus, peripheral component interconnection (PCI) bus, PCI-Express (PCI-X) bus, serial advanced technology attachment (SATA) bus, video electronics standard association local (VLB) bus or other suitable bus or two or more of these combinations. In appropriate cases, bus 710 may include one or more buses. Although the present application embodiment describes and shows a specific bus, the present application considers any suitable bus or interconnection.

[0201] In addition, in combination with the baseline detection method in the above embodiment, the embodiment of the present application can provide a computer storage medium for implementation. The computer storage medium stores computer program instructions; when the computer program instructions are executed by the processor, the baseline detection method provided in the embodiment of the present application is implemented.

[0202] An embodiment of the present application also provides a computer program product, which includes a computer program, and the computer program is executed by a processor to implement the baseline detection method provided in the embodiment of the present application.

[0203] It should be clear that the present application is not limited to the specific configuration and processing described above and shown in the figures. For the sake of simplicity, a detailed description of the known method is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present application is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between the steps after understanding the spirit of the present application.

[0204] The functional blocks shown in the above block diagram can be implemented as hardware, software, firmware or a combination thereof. When implemented in hardware, it can be, for example, an electronic circuit, an application specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc. When implemented in software, the elements of the present application are programs or code segments that are used to perform the required tasks. The program or code segment can be stored in a machine-readable medium, or transmitted on a transmission medium or a communication link by a data signal carried in a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, optical fiber media, radio frequency (RF) links, etc. The code segment can be downloaded via a computer network such as the Internet, an intranet, etc.

[0205] It should also be noted that the exemplary embodiments mentioned in this application describe some methods or systems based on a series of steps or devices. However, this application is not limited to the order of the above steps, that is, the steps can be performed in the order mentioned in the embodiment, or in a different order from the embodiment, or several steps can be performed simultaneously.

[0206] Aspects of the present disclosure are described above with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present disclosure. It should be understood that each box in the flowchart and / or block diagram and the combination of boxes in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable baseline detection device to produce a machine so that these instructions executed by the processor of the computer or other programmable baseline detection device enable the implementation of the functions / actions specified in one or more boxes of the flowchart and / or block diagram. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field programmable logic circuit. It can also be understood that each box in the block diagram and / or flowchart and the combination of boxes in the block diagram and / or flowchart can also be implemented by dedicated hardware that performs the specified function or action, or can be implemented by a combination of dedicated hardware and computer instructions.

[0207] The above are only specific implementation methods of the present application. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the described systems, modules and units can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here. It should be understood that the protection scope of the present application is not limited to this. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in this application, and these modifications or replacements should be included in the protection scope of this application.

Claims

1. A baseline detection method, characterized in that: The method comprises: Get the object parameters of the object to be detected; Inputting the object parameters of the object to be detected into a pattern recognition model, and processing the object parameters of the object to be detected by the pattern recognition model to obtain a pattern type corresponding to the object to be detected; Inputting the pattern type corresponding to the object to be detected into a version determination model, and processing the pattern type corresponding to the object to be detected by the version determination model to obtain a baseline detection version of the object to be detected; A baseline detection is performed on the object to be detected based on the baseline detection version of the object to be detected.

2. The method according to claim 1, characterized in that Get the object parameters of the object to be detected, including: When it is detected that the current moment is in a target detection cycle among multiple detection cycles, object parameters of the object to be detected in the target detection cycle are obtained.

3. The method according to claim 1, characterized in that The step of inputting the object parameters of the object to be detected into a pattern recognition model, and processing the object parameters of the object to be detected by the pattern recognition model to obtain a pattern type corresponding to the object to be detected includes: Inputting the object parameter of the object to be detected into a pattern recognition model, wherein the pattern recognition model includes a correspondence between parameter value intervals and pattern types, wherein the correspondence between parameter value intervals and pattern types includes multiple parameter value intervals and a pattern type corresponding to each parameter value interval; Determining, by the pattern recognition model, a target parameter value interval to which the object parameter of the object to be detected belongs from the multiple parameter value intervals based on the correspondence between the parameter value intervals and the pattern type and the object parameter of the object to be detected; The pattern recognition model is used to match the target parameter value interval based on the correspondence between the parameter value interval and the pattern type to obtain the pattern type corresponding to the object to be detected.

4. The method according to claim 3, characterized in that The pattern recognition model includes N layer structures, and the i-th layer structure includes 2 i-1 Support vector machine SVM models, each of which includes two sub-corresponding relationships, each of which includes a parameter value space and a mode type corresponding to the parameter value space, and the corresponding relationship between the parameter value interval and the mode type includes 2 N The correspondence between groups is that i is less than or equal to N, and both i and N are positive integers.

5. The method according to claim 1, characterized in that The step of inputting the pattern type corresponding to the object to be detected into a version determination model, and processing the pattern type of the object to be detected by the version determination model to obtain a baseline detection version of the object to be detected includes: Inputting the pattern type corresponding to the object to be detected into a version determination model, wherein the version determination model includes a mapping relationship between multiple pattern types and baseline detection versions; The version determination model is used to match the pattern type corresponding to the object to be detected based on the mapping relationship between the multiple pattern types and the baseline detection version to obtain the baseline detection version of the object to be detected.

6. The method according to claim 1 or 5, characterized in that: Before inputting the pattern type of the object to be detected into the version determination model, the method further includes: Acquire multiple mapping relationship groups, each of the mapping relationship groups includes multiple pending mapping relationships and a relationship tag for each of the pending mapping relationships, each of the pending mapping relationships includes a pattern type and a pending baseline detection version, the pattern types in different pending mapping relationships in each of the mapping relationship groups are the same, and the pending baseline detection versions in different pending mapping relationships in each of the mapping relationship groups are different, and the relationship tag for each of the pending mapping relationships is used to characterize the degree of matching between the pattern type in the pending mapping relationship and the pending baseline detection version; For each mapping relationship group, perform the following steps: For each of the pending mapping relationships, the pattern type in the pending mapping relationship is processed P times based on the first model to obtain P pattern vectors, and the pending baseline detection version in the pending mapping relationship is processed Q times based on the first model to obtain Q version vectors, where P and Q are both positive integers; Based on the second model, for each vector group in P×Q vector groups, calculating the similarity between the pattern vector and the version vector in the vector group, wherein the P×Q vector groups are determined based on the P pattern vectors and the Q version vectors; Determine a total loss score of the pending mapping relationship based on the similarity of each vector group among the similarities of the P×Q vector groups corresponding to the pending mapping relationship, the relationship label of the pending mapping relationship, the P pattern vectors, and the Q version vectors; A pending mapping relationship with the smallest total loss score is determined from the multiple pending mapping relationships, as the mapping relationship corresponding to the mapping relationship group.

7. The method according to claim 6, characterized in that The determining, based on the similarities of each vector group among the P×Q vector groups corresponding to the pending mapping relationship, the relationship label of the pending mapping relationship, the P pattern vectors, and the Q version vectors, a total loss score of the pending mapping relationship comprises: Calculate P×Q first loss scores based on the similarity of each vector group among the similarities of the P×Q vector groups corresponding to the pending mapping relationship and the relationship label of the pending mapping relationship; Determine a second loss score of the pending mapping relationship based on the KL divergence loss values ​​of the P pattern vectors and the KL divergence loss values ​​of the Q version vectors; Based on the P×Q first loss scores and the second loss scores, a total loss score of the pending mapping relationship is determined.

8. The method according to any one of claims 1, 3 and 4, characterized in that: Before inputting the object parameters of the object to be detected into the pattern recognition model, the method further comprises: Acquire a training sample set, the training sample set comprising a plurality of object parameter samples and a sample label corresponding to each of the object parameter samples, the sample label corresponding to each of the object parameter samples being used to indicate a mode type corresponding to the object parameter sample; For each object parameter sample in the plurality of object parameter samples, perform the following steps: Inputting the object parameter sample into a preset pattern recognition model, wherein the preset pattern recognition model includes a correspondence between a preset parameter value interval and a pattern type, wherein the correspondence between the preset parameter value interval and the pattern type includes a plurality of preset parameter value intervals and a pattern type corresponding to each of the preset parameter value intervals; Determining, by the preset pattern recognition model, a reference parameter value interval to which the object parameter sample belongs from the plurality of preset parameter value intervals based on the correspondence between the preset parameter value intervals and the pattern type and the object parameter sample; The preset pattern recognition model is used to match the reference parameter value interval based on the corresponding relationship between the preset parameter value interval and the pattern type to determine the pattern type of the object parameter sample; Processing the pattern type corresponding to the target object parameter sample and the sample label corresponding to the target object parameter sample based on the loss function to obtain a loss function value of a preset pattern recognition model, wherein the target object parameter sample is any one of the multiple object parameter samples; Based on the loss function value of the preset pattern recognition model, the preset pattern recognition model is trained using a plurality of object parameter samples to obtain a trained pattern recognition model.

9. A baseline detection device, characterized in that: The device comprises: An acquisition module, used for acquiring object parameters of the object to be detected; A processing module, used for inputting the object parameters of the object to be detected into a pattern recognition model, processing the object parameters of the object to be detected by the pattern recognition model, and obtaining a pattern type corresponding to the object to be detected; The processing module is further used to input the pattern type corresponding to the object to be detected into a version determination model, and process the pattern type corresponding to the object to be detected through the version determination model to obtain a baseline detection version of the object to be detected; The baseline detection module is used to perform a baseline detection on the object to be detected based on the baseline detection version of the object to be detected.

10. An electronic device, characterized in that: The device comprises: a processor and a memory storing computer program instructions; The processor reads and executes the computer program instructions to implement the baseline detection method as described in any one of claims 1-8.

11. A computer storage medium, characterized in that: The computer storage medium stores computer program instructions, and when the computer program instructions are executed by the processor, the baseline detection method according to any one of claims 1 to 8 is implemented.

12. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the baseline detection method according to any one of claims 1 to 8 is implemented.