Dynamic scale anomaly detection method and device and readable storage medium

Abstract

The application provides a dynamic scale anomaly detection method, device and readable storage medium. The method comprises the following steps: acquiring complex weight information of a plurality of dynamic scales in a preset time range, wherein the complex weight information comprises identification information of each dynamic scale in the plurality of dynamic scales, total quantity information of the plurality of dynamic scales, and weight information of a package weighed by each dynamic scale and shipping order information corresponding to the package; preprocessing the complex weight information to obtain effective complex weight information of the plurality of dynamic scales; calculating complex weight deviation information between each dynamic scale and other dynamic scales in the plurality of dynamic scales based on the effective complex weight information; and determining a target abnormal dynamic scale from the plurality of dynamic scales based on the complex weight deviation information. Compared with the prior art, the embodiment of the application can quickly determine a target abnormal dynamic scale with more abnormal possibilities from the plurality of dynamic scales, avoids the case that a large area of manual detection is performed on each dynamic scale, saves the cost, and reduces the mis-sorting risk.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of logistics, in particular to a dynamic scale abnormality detection method and device and a readable storage medium. BACKGROUND

[0002] With the development of science and technology and the acceleration of life rhythm, the express industry and people's life have become more and more inseparable, and a large number of packages will exist in the express company every day.

[0003] In the sorting process, dynamic scales are used to dynamically weigh packages to control the risk of underestimating the weight of express items caused by irregular operations or malicious behavior. However, as mechanical equipment, dynamic scales may have abnormal situations, which may lead to inaccurate subsequent control. In addition, due to the large number of sorting equipment nationwide, it is impossible to effectively detect the abnormality of each dynamic scale.

[0004] Therefore, how to effectively detect whether the dynamic scale is abnormal is a technical problem that needs to be solved in the current logistics technical field. SUMMARY

[0005] The present application provides a dynamic scale abnormality detection method, device and readable storage medium, which aims to solve the technical problem of how to effectively detect whether the dynamic scale is abnormal.

[0006] In one aspect, the present application provides a dynamic scale abnormality detection method, which comprises:

[0007] Obtaining complex weight information of a plurality of dynamic scales in a preset time range, the complex weight information comprising identification information of each dynamic scale in the plurality of dynamic scales, total quantity information of the plurality of dynamic scales, and weight information of a package weighed by the each dynamic scale and shipping order information corresponding to the package;

[0008] Pretreating the complex weight information to obtain effective complex weight information of the plurality of dynamic scales;

[0009] Based on the effective complex weight information, calculating complex weight deviation information between the each dynamic scale and other dynamic scales in the plurality of dynamic scales;

[0010] Based on the complex weight deviation information, determining a target abnormal dynamic scale from the plurality of dynamic scales.

[0011] In one possible implementation of the present application, based on the complex weight deviation information, the target abnormal dynamic scale is determined from the plurality of dynamic scales, comprising:

[0012] Based on the complex weight deviation information, calculating an abnormal score value of the each dynamic scale;

[0013] determine a target abnormal dynamic scale from the plurality of dynamic scales based on the abnormal score value.

[0014] In a possible implementation of the present application, the pre-processing of the complex weight information to obtain the valid complex weight information of the plurality of dynamic scales comprises:

[0015] filtering the complex weight information with the weight information greater than a preset weight threshold to obtain first complex weight information;

[0016] identifying invalid shipping order information and superimposed shipping order information in the shipping order information, wherein the superimposed shipping order information is the same shipping order information corresponding to the logistics information of at least two packages;

[0017] deleting the invalid shipping order information and the superimposed shipping order information from the first complex weight information to obtain second complex weight information;

[0018] deleting the complex weight information only passing through one dynamic scale for weighing in the second complex weight information to obtain the valid complex weight information of the plurality of dynamic scales.

[0019] In a possible implementation of the present application, the calculating of the complex weight deviation information between each dynamic scale and other dynamic scales in the plurality of dynamic scales based on the valid complex weight information comprises:

[0020] calculating a complex weight deviation ratio between each dynamic scale and other dynamic scales in the plurality of dynamic scales based on the shipping order information, the weight information and the identification information;

[0021] determining complex weight deviation feature information based on the complex weight deviation ratio;

[0022] determining the complex weight deviation information between each dynamic scale and other dynamic scales in the plurality of dynamic scales based on a preset time window mode and the complex weight deviation feature information.

[0023] In a possible implementation of the present application, the calculating of the abnormal score value of each dynamic scale based on the complex weight deviation information comprises:

[0024] inputting the complex weight deviation information into a pre-trained isolation forest model to obtain the abnormal score value of each dynamic scale.

[0025] In a possible implementation of the present application, the determining of a target abnormal dynamic scale from the plurality of dynamic scales based on the abnormal score value comprises:

[0026] sorting the plurality of dynamic scales in descending order of abnormal risk based on the abnormal score value to obtain dynamic scale descending order sorting information;

[0027] determine a target abnormal dynamic scale quantity according to the preset abnormal proportion hyperparameter and the total quantity information;

[0028] determine a target abnormal dynamic scale from the plurality of dynamic scales based on the target abnormal dynamic scale quantity, the identification information, and the dynamic scale descending order sorting information.

[0029] In a possible implementation of the present application, after determining a target abnormal dynamic scale from the plurality of dynamic scales based on the abnormal score value, the method further includes:

[0030] statistically determine actual abnormal information in the target abnormal dynamic scale;

[0031] calculate an abnormal detection accuracy based on the actual abnormal information, the total quantity information, and the hyperparameter;

[0032] adjust the hyperparameter based on the abnormal detection accuracy and a preset accuracy threshold.

[0033] In another aspect, the present application provides a dynamic scale abnormality detection device, which includes:

[0034] a first acquisition unit configured to acquire complex weight information of a plurality of dynamic scales in a preset time range, the complex weight information including identification information of each dynamic scale in the plurality of dynamic scales, total quantity information of the plurality of dynamic scales, and weight information of a package weighed by the each dynamic scale and shipping order information corresponding to the package;

[0035] a first preprocessing unit configured to preprocess the complex weight information to obtain effective complex weight information of the plurality of dynamic scales;

[0036] a first calculation unit configured to calculate complex weight deviation information between the each dynamic scale and other dynamic scales in the plurality of dynamic scales based on the effective complex weight information;

[0037] a first determination unit configured to determine a target abnormal dynamic scale from the plurality of dynamic scales based on the complex weight deviation information.

[0038] In a possible implementation of the present application, the first determination unit specifically includes:

[0039] a first calculation unit configured to calculate an abnormal score value of the each dynamic scale based on the complex weight deviation information;

[0040] a second determination unit configured to determine a target abnormal dynamic scale from the plurality of dynamic scales based on the abnormal score value.

[0041] In a possible implementation of the present application, the pre-processing of the complex weight information to obtain the effective complex weight information of the plurality of dynamic scales is specifically used for:

[0042] The pre-processing of the complex weight information to obtain the effective complex weight information of the plurality of dynamic scales includes:

[0043] Filtering the complex weight information with the weight information greater than a preset weight threshold to obtain first complex weight information;

[0044] Identifying invalid shipping order information and superimposed shipping order information in the shipping order information, wherein the superimposed shipping order information is the same shipping order information corresponding to the logistics information of at least two packages;

[0045] Deleting the invalid shipping order information and the superimposed shipping order information from the first complex weight information to obtain second complex weight information;

[0046] Deleting the complex weight information only passing through one dynamic scale for weighing in the second complex weight information to obtain the effective complex weight information of the plurality of dynamic scales.

[0047] In a possible implementation of the present application, the first calculation unit is specifically used for:

[0048] Calculating a complex weight deviation ratio between each dynamic scale and other dynamic scales in the plurality of dynamic scales based on the shipping order information, the weight information, and the identification information;

[0049] Determining complex weight deviation feature information based on the complex weight deviation ratio;

[0050] Determining complex weight deviation information between each dynamic scale and other dynamic scales in the plurality of dynamic scales based on a preset time window mode and the complex weight deviation feature information.

[0051] In a possible implementation of the present application, the second calculation unit is specifically used for:

[0052] Inputting the complex weight deviation information into a pre-trained isolation forest model to obtain an abnormal score value of each dynamic scale.

[0053] In a possible implementation of the present application, the second determination unit is specifically used for:

[0054] Based on the abnormal score value, performing abnormal risk descending sorting on the plurality of dynamic scales to obtain dynamic scale descending sorting information;

[0055] Determining a target abnormal dynamic scale quantity according to a pre-set abnormal proportion hyperparameter and the total quantity information;

[0056] determine a target abnormal dynamic scale from the plurality of dynamic scales based on the target abnormal dynamic scale quantity, the identification information, and the dynamic scale descending order sorting information.

[0057] In a possible implementation of the present application, after determining the target abnormal dynamic scale from the plurality of dynamic scales based on the abnormal score value, the device is further configured to:

[0058] count actual abnormal information in the target abnormal dynamic scale;

[0059] calculate an abnormal detection accuracy based on the actual abnormal information, the total quantity information, and the hyperparameter;

[0060] adjust the hyperparameter based on the abnormal detection accuracy and a preset accuracy threshold.

[0061] In another aspect, the present application also provides a computer device, which comprises:

[0062] one or more processors;

[0063] a memory; and

[0064] one or more application programs, wherein the one or more application programs are stored in the memory and configured to be executed by the processor to implement the dynamic scale abnormal detection method.

[0065] In another aspect, the present application also provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is loaded by a processor to execute the steps in the dynamic scale abnormal detection method.

[0066] The application provides a dynamic scale anomaly detection method, which comprises the following steps: acquiring complex weight information of a plurality of dynamic scales in a preset time range, wherein the complex weight information comprises identification information of each dynamic scale in the plurality of dynamic scales, total quantity information of the plurality of dynamic scales, and weight information of a package weighed by each dynamic scale and shipping order information corresponding to the package; preprocessing the complex weight information to obtain effective complex weight information of the plurality of dynamic scales; calculating complex weight deviation information between each dynamic scale and other dynamic scales in the plurality of dynamic scales based on the effective complex weight information; and determining a target abnormal dynamic scale from the plurality of dynamic scales based on the complex weight deviation information. Compared with the prior art, the application calculates the complex weight deviation information between each dynamic scale and other dynamic scales in the plurality of dynamic scales based on the effective complex weight information of the plurality of dynamic scales obtained by preprocessing the complex weight information of the plurality of dynamic scales in the preset time range, and then determines the target abnormal dynamic scale that is more likely to be abnormal from the plurality of dynamic scales based on the complex weight deviation information, thereby avoiding the condition that a large area is manually detected for each dynamic scale, saving the cost, and reducing the risk of mis-sorting. BRIEF DESCRIPTION OF DRAWINGS

[0067] In order to more clearly illustrate the technical solutions in the embodiments of the application, the drawings needed to be used in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative effort on the basis of these drawings.

[0068] Figure 1 is a scene schematic diagram of a dynamic scale anomaly detection system provided by the embodiments of the application;

[0069] Figure 2 is an embodiment flowchart of a dynamic scale anomaly detection method provided in the embodiments of the application;

[0070] Figure 3 is an embodiment flowchart of step 204 in the embodiments of the application;

[0071] Figure 4 is an embodiment flowchart of step 202 in the embodiments of the application;

[0072] Figure 5 is an embodiment flowchart of step 203 in the embodiments of the application

[0073] Figure 6 is an embodiment flowchart of step 302 in the embodiments of the application

[0074] Figure 7is another embodiment flow diagram of the dynamic scale abnormality detection method provided in the embodiments of the present application

[0075] Figure 8 is an embodiment structure diagram of the dynamic scale abnormality detection device provided in the embodiments of the present application

[0076] Figure 9 is an embodiment structure diagram of the computer device provided in the embodiments of the present application. DETAILED DESCRIPTION

[0077] The technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0078] In the description of the present application, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application. In addition, the terms "first", "second" are only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Therefore, the features limited by "first", "second" can explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "multiple" is two or more, unless otherwise explicitly and specifically limited.

[0079] In the present application, the word "exemplary" is used to mean "serving as an example, instance, or illustration". Any implementation described as "exemplary" in the present application is not necessarily to be construed as preferred or advantageous over other implementations. The following description is presented to enable any person skilled in the art to make and use the present application. In the following description, for the purpose of explanation, details are set forth. It is apparent to those skilled in the art that the present application can be practiced without using these specific details. In other instances, well-known structures and processes are not described in detail in order to avoid obscuring the description of the present application. Thus, the present application is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed herein.

[0080] This application provides a dynamic scale anomaly detection method, device, and readable storage medium, which will be described in detail below.

[0081] like Figure 1 As shown, Figure 1 This is a schematic diagram of a dynamic scale anomaly detection system provided in an embodiment of this application. The system may include multiple terminals 100 and a server 200, which are network-connected. The server 200 integrates a dynamic scale anomaly detection device, such as... Figure 1 In the server, terminal 100 can access server 200.

[0082] In this embodiment, the server 200 is mainly used to obtain the re-weighing information of multiple dynamic scales within a preset time range. The re-weighing information includes the identification information of each dynamic scale, the total number of dynamic scales, the weight information of the package weighed by each dynamic scale, and the corresponding waybill information of the package. The re-weighing information is preprocessed to obtain the effective re-weighing information of the multiple dynamic scales. Based on the effective re-weighing information, the re-weighing deviation information between each dynamic scale and other dynamic scales in the multiple dynamic scales is calculated. Based on the re-weighing deviation information, the target abnormal dynamic scale is determined from the multiple dynamic scales.

[0083] In this embodiment, the server 200 can be a standalone server, a server network, or a server cluster. For example, the server 200 described in this embodiment includes, but is not limited to, a computer, a network terminal, a single network server, a set of multiple network servers, or a cloud server composed of multiple servers. The cloud server is composed of a large number of computers or network servers based on cloud computing. In this embodiment, communication between the server and the terminal can be achieved through any communication method, including but not limited to, mobile communication based on the 3rd Generation Partnership Project (3GPP), Long Term Evolution (LTE), and Worldwide Interoperability for Microwave Access (WiMAX), or computer network communication based on the TCP / IP Protocol Suite (TCP / IP) and User Datagram Protocol (UDP).

[0084] It can be understood that the terminal 100 used in the embodiments of the present application can be a device that includes receiving and transmitting hardware, and has receiving and transmitting hardware capable of performing bidirectional communication on a bidirectional communication link. Such a terminal can include a cellular or other communication device with a single-line display or a multi-line display or a cellular or other communication device without a multi-line display. The terminal 100 can be a desktop terminal or a mobile terminal, and can be one of a mobile phone, a tablet computer, a notebook computer, etc.

[0085] Those skilled in the art can understand that Figure 1 The application environment shown in the above Figure 1 The application environment shown in the above Figure 1 The application environment shown in the above

[0086] In addition, as shown in the above Figure 1 The dynamic scale abnormality detection system can further include a memory 300 for storing data, such as dynamic scale weight data and dynamic scale abnormality detection data, for example, dynamic scale abnormality detection data when the dynamic scale abnormality detection system is running.

[0087] It should be noted that Figure 1 The scene diagram of the dynamic scale abnormality detection system shown in the above is only an example, and the dynamic scale abnormality detection system and the scene described in the embodiments of the present application are used to more clearly illustrate the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art can know that, as the dynamic scale abnormality detection system evolves and new business scenarios appear, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

[0088] Next, the dynamic scale abnormality detection method provided by the embodiments of the present application is introduced.

[0089] In an embodiment of the dynamic scale abnormality detection method provided in the present application, a dynamic scale abnormality detection device is taken as an execution subject. In order to simplify and facilitate description, the execution subject will be omitted in subsequent method embodiments. The dynamic scale abnormality detection device is applied to a computer device, and the method comprises: acquiring complex weight information of a plurality of dynamic scales in a preset time range, the complex weight information comprising identification information of each dynamic scale in the plurality of dynamic scales, total quantity information of the plurality of dynamic scales, and weight information of a package weighed by each dynamic scale and shipping order information corresponding to the package; preprocessing the complex weight information to obtain effective complex weight information of the plurality of dynamic scales; based on the effective complex weight information, calculating complex weight deviation information between each dynamic scale and other dynamic scales in the plurality of dynamic scales; and determining a target abnormal dynamic scale from the plurality of dynamic scales based on the complex weight deviation information.

[0090] Please refer to Figures 2 to 7 , Figure 2 An embodiment flowchart of the dynamic scale abnormality detection method provided in the present application is shown. The dynamic scale abnormality detection method comprises steps 201 to 204.

[0091] 201. Acquire complex weight information of a plurality of dynamic scales in a preset time range.

[0092] When a user sends a package, a staff member will weigh the package for the first time to obtain weight information of the package. At the same time, the user will also fill in relevant sending information on a shipping order corresponding to the package. The sending information generally includes a sending address, a receiving address, a sender's name, a sender's contact information, a consignee, and a consignee's contact information. Thus, the staff member can determine a sending fee according to the weight information of the package, the sending address, and the receiving address.

[0093] The complex weight information refers to at least second repeated weighing information relative to the first weighing of the package. Specifically, the complex weight information can comprise identification information of each dynamic scale in the plurality of dynamic scales, total quantity information of the plurality of dynamic scales, and weight information of a package weighed by each dynamic scale and shipping order information corresponding to the package. The identification information of the dynamic scale is a marker for identifying the dynamic scale, which can be a device number of the dynamic scale. Generally, each logistics transfer station will be provided with a plurality of sorting devices, and at least one dynamic scale will be provided on each sorting device. However, the background of each logistics transfer station will record all device information, including the total quantity information of the plurality of dynamic scales. When a package is sorted by a sorting device, the shipping order information corresponding to the package will be read by a bar gun. Therefore, when the package is weighed by a dynamic scale, the weighing information of the package and the corresponding shipping order information will be recorded one by one.

[0094] Generally, the reweighing of the package is to control the risk of less weight of the express caused by illegal operation or malicious behavior, and the reweighing result of the dynamic scale is an important indicator. Therefore, the preset time range needs to ensure the timeliness as a prerequisite, for example, the interval between the reweighing of the package and the delivery of the package is 32h, and the preset time range should be within 32h. Of course, in actual business, the preset time range is generally within 24h or 12h, and the specific time range can be adjusted according to the actual situation.

[0095] 202、The reweighing information is preprocessed to obtain effective reweighing information of the plurality of dynamic scales.

[0096] In the embodiment of the application, the reweighing deviation information between a certain dynamic scale and other dynamic scales in the plurality of dynamic scales is mainly used for abnormal detection. Therefore, if a certain package only passes through one dynamic scale for reweighing, the reweighing information of the certain package cannot be used.

[0097] Further, the reweighing information in some special scenarios cannot be used, for example, in a test scenario, there are some test packages and corresponding test waybill information.

[0098] Before using the reweighing information, it is necessary to preprocess the reweighing information to obtain effective target weight information. Specifically, the preprocessing can be data cleaning of the reweighing information, and the specific preprocessing steps can be seen in the following embodiment.

[0099] 203、Based on the effective reweighing information, the reweighing deviation information between each dynamic scale and other dynamic scales in the plurality of dynamic scales is calculated.

[0100] The weight deviation information refers to the weight deviation information between the weight information of a certain dynamic scale and the weight information of other dynamic scales in the plurality of dynamic scales for the same package and corresponding waybill information.

[0101] The specific calculation method can be seen in the following embodiment.

[0102] 204、Based on the reweighing deviation information, a target abnormal dynamic scale is determined from the plurality of dynamic scales.

[0103] The application provides a dynamic scale anomaly detection method, which comprises the following steps: acquiring complex weight information of a plurality of dynamic scales in a preset time range, wherein the complex weight information comprises identification information of each dynamic scale in the plurality of dynamic scales, total quantity information of the plurality of dynamic scales, and weight information of a package weighed by each dynamic scale and shipping order information corresponding to the package; preprocessing the complex weight information to obtain effective complex weight information of the plurality of dynamic scales; calculating complex weight deviation information between each dynamic scale and other dynamic scales in the plurality of dynamic scales based on the effective complex weight information; and determining a target abnormal dynamic scale from the plurality of dynamic scales based on the complex weight deviation information. Compared with the prior art, the application calculates the complex weight deviation information between each dynamic scale and other dynamic scales in the plurality of dynamic scales based on the effective complex weight information of the plurality of dynamic scales obtained by preprocessing the complex weight information of the plurality of dynamic scales in the preset time range, and then determines the target abnormal dynamic scale that is more likely to be abnormal from the plurality of dynamic scales based on the complex weight deviation information, thereby avoiding the case that a large area of manual detection is performed on each dynamic scale, saving the cost and reducing the risk of mis-sorting.

[0104] In the embodiment of the application, as shown in Figure 3 Step 204 specifically comprises steps 301 and 302.

[0105] 301. Calculate an abnormal score value of each dynamic scale based on the complex weight deviation information.

[0106] 302. Determine a target abnormal dynamic scale from the plurality of dynamic scales based on the abnormal score value.

[0107] In step 301, the abnormal score value of each dynamic scale is calculated based on the complex weight deviation information. Specifically, the complex weight deviation information can be input into a pre-trained isolation forest model to obtain the abnormal score value of each dynamic scale.

[0108] Among all the dynamic scales, the dynamic scales in an abnormal state are usually a minority. Although the dynamic scale anomaly itself has many types and can be heavy, light or randomly fluctuating, it will be obviously different from the normal state, that is, it cannot weigh the correct weight. The isolation forest algorithm is an unsupervised anomaly detection method suitable for continuous data (Continuous numerical data). Anomaly refers to an outlier that is easy to be isolated. It can be understood as a point that is far away from the group with high density and sparse distribution, which is consistent with the situation of abnormal dynamic scales. The core of the isolation forest algorithm is to recursively randomly divide the data set until all sample points are isolated. Under this random division strategy, the abnormal points usually have a shorter path. Intuitively, the points that can be divided into leaf nodes for a smaller number of times are more likely to be abnormal data.

[0109] In the embodiments of the present application, before using the isolated forest model, the model can be pre-trained, specifically, the model can be iteratively trained and updated, the historical weight deviation information of a preset time period is used as a training set to construct an isolated forest, and an interface service is encapsulated for calling, wherein the average path length of the tree is:

[0110]

[0111] wherein n is the sample capacity of the training set, H(i) is a harmonic function, and c(n) is the average value of the path length when the sample number n is given, which is used to standardize the path length h(x) of the sample x.

[0112] Then, the dynamic scale feature data x of the day is input, and the anomaly score of each dynamic scale sample x is:

[0113]

[0114] In step 302, based on the anomaly score value, a target abnormal dynamic scale is determined from a plurality of dynamic scales, and the following specific embodiments can be seen.

[0115] In the embodiments of the present application, the data in the dynamic scale weight information belongs to industrial data, and there are more dirty data. In order to ensure the detection effect and the accuracy of the final conclusion, data cleaning is needed. As shown in the following table, step 202 specifically includes steps 401 to 404: Figure 4

[0116] 401, filtering the weight information greater than the preset weight threshold to obtain the first weight information.

[0117] 402, identifying invalid shipping information and superimposed shipping information in the shipping information.

[0118] The superimposed shipping information is the same shipping information corresponding to the logistics information of at least two packages. The logistics information can be weight information. That is, when two packages are not completely sorted in the sorting process, at least two packages overlap, which can be overlapped up and down or left and right, so that only one package of shipping information is scanned to the side of the shipping scanning system. When at least two packages overlap, the overlapping package is formed, and at this time, the weight information taken by the dynamic scale is actually the weight information of the overlapping package.

[0119] 403, deleting invalid shipping information and superimposed shipping information from the first weight information to obtain the second weight information.

[0120] ​404. Delete the re-weighing information in the second re-weighing information that has only been weighed by one dynamic scale, and obtain the valid re-weighing information of multiple dynamic scales.

[0121] In step 401, the preset weight threshold can be determined based on the normal weighing range of the dynamic scale. For example, under normal circumstances, the normal weighing range of dynamic scales in some transfer stations is within 60kg. Therefore, the preset weight threshold can be set to 60kg. It should be noted that in some special transfer stations, special dynamic scales can be configured according to the weighing object, and the normal weighing range of their dynamic scales can reach 100kg or higher. Therefore, the preset weight threshold can be set according to the normal weighing range of the dynamic scales actually configured in the transfer station.

[0122] In steps 402 and 403, as described above, duplicate information in certain special scenarios cannot be used. For example, in a test scenario, there may be test packages and corresponding test waybill information, which are invalid waybill information. This invalid waybill information is recorded in the waybill database; therefore, it can be identified and deleted (filtered) using waybill routing information. For overlapping waybill information, a configured visual recognition device can identify overlapping packages based on visual algorithms, thereby deleting (filtering) the overlapping waybill information.

[0123] In step 404, there are N scales in total. For a certain scale body 'a', 1k packages passed over it within a certain time period, generating 1k waybill numbers and weighing results records. Among these 1k packages, 0.2k packages may also have passed over any one or more of the remaining N-1 scales. The remaining 0.8k packages are considered duplicate weighing information, meaning they were weighed only by one dynamic scale.

[0124] In the embodiments of this application, such as Figure 5 As shown, step 203 specifically includes steps 501 to 503:

[0125] 501. Based on waybill information, weight information, and identification information, calculate the reweight deviation ratio between each dynamic scale and other dynamic scales among multiple dynamic scales.

[0126] 502. Based on the repetition deviation ratio, determine the repetition deviation characteristic information.

[0127] 503. Based on the preset time windowing method and reweight deviation characteristic information, determine the reweight deviation information between each dynamic scale and other dynamic scales among multiple dynamic scales.

[0128] In step 501, first, according to the identification information, a certain dynamic scale is determined, the first weight information of the package on the certain dynamic scale is read, then according to the waybill information of the package, the second weight information of the package on other dynamic scales in the plurality of dynamic scales is determined, and then the complex weight deviation ratio is calculated through the first weight information and the second weight information. For example, the complex weight deviation ratio of each ticket = |the weighing result of scale body a - the weighing result of any scale body x in the remaining N-1| / the weighing result of scale body a * 100%.

[0129] In step 502, the complex weight deviation feature information refers to the feature data of dividing the complex weight deviation ratio into a plurality of segment proportions, wherein the plurality of segment proportions can be set as required, and the plurality of segments are required to be not less than two. For example, five segment proportions are set, which are 0-5%, 5%-10%, 10%-20%, 20%-40%, and 40%-80%. Specifically, there are N scales in total, for a certain scale body a, in a certain time period, 1w tickets pass through it, generating 1w records of waybill numbers and weighing results. Among the 1w tickets, there may be 200 tickets that also pass through any one or more of the remaining N-1 scales. At this time, the weighing results of the 200 tickets on the scale body a and the remaining scales are counted, and the complex weight deviation ratio of each ticket = |the weighing result of scale body a - the weighing result of any scale body x in the remaining N-1| / the weighing result of scale body a * 100%. Thus, there are not less than 200 difference ratio results for scale body a (if one of the 200 tickets also passes through two scales, and the other 199 tickets only pass through one scale, there are 201 difference ratios). Here, it is assumed that there are only 200 difference ratio results, among which 100 difference ratios are less than 5%, 125 difference ratios are less than 10%, 150 difference ratios are less than 20%, 175 difference ratios are less than 40%, and 200 difference ratios are less than 80%. Therefore, for scale body a, the complex weight deviation ratio in the time range is divided into the feature values corresponding to the deviation of 5%, 10%, 20%, 40%, and 80% as (0.5, 0.625, 0.75, 0.875, 1).

[0130] In step 503, the preset time window mode can be a Fibonacci time window mode, specifically, 5 time ranges (1 day, 2 days, 3 days, 5 days, and 8 days) can be selected, so that the characteristic value of each scale body in the specified time range can be calculated. Each time period corresponds to 5-dimensional complex weight deviation characteristics, and there are 5 time ranges (1 day, 2 days, 3 days, 5 days, and 8 days), so the feature vector of each scale body finally has 25 dimensions. The above complex weight deviation characteristic information is calculated by dividing the data by time window, and the complex weight deviation information is obtained. For example, today (June 11), the time window is 1, and only the data of June 10 is taken to calculate; the time window is 8, and the data from June 3 to June 10 is taken to calculate the characteristic value. Finally, the 5-dimensional difference value characteristics calculated in each time window are spliced to become a 25-dimensional final dynamic scale feature. In this way, the cumulative reinforcement representation of the weighing difference of the dynamic scale in each deviation section and the cumulative reinforcement representation of the suddenness in time can be realized.

[0131] According to the foregoing calculation, each scale body can obtain a 25-dimensional feature every day. Assuming that there are a total of 200 scales, and the 15-day data set is a 3000*25 matrix. Each scale has 15 sample data, and each sample feature has 25 dimensions (5-dimensional difference value characteristics corresponding to each time window, and there are 5 time windows, so there are 25 dimensions in total).

[0132] In the embodiment of the application, as shown in Figure 6 Step 302 specifically includes steps 601 to 603:

[0133] 601. Based on the abnormal score value, the plurality of dynamic scales are sorted in descending order of abnormal risk to obtain dynamic scale descending order sorting information.

[0134] 602. According to the pre-set abnormal proportion hyperparameter and total quantity information, the number of target abnormal dynamic scales is determined.

[0135] 603. Based on the number of target abnormal dynamic scales, the identification information, and the dynamic scale descending order sorting information, the target abnormal dynamic scale is determined from the plurality of dynamic scales.

[0136] In step 602, according to the set hyperparameter contamination (the proportion of abnormal values in the data set), based on the total quantity information n of the dynamic scales, the number of target abnormal dynamic scales is selected = n*contamination. For example, the total quantity information is 1000 dynamic scales, the hyperparameter contamination is 0.005, and then the number of target abnormal dynamic scales = 1000*0.005 = 5. That is, the number of target abnormal dynamic scales is 5.

[0137] In step 603, when the number of target abnormal dynamic scales is 5, the first 5 in the dynamic scale descending order sorting information are selected to be determined as the target abnormal dynamic scales.

[0138] In the embodiments of the present application, as shown in step 204, the method further comprises steps 701 to 703: Figure 7

[0139] 701. Count the actual abnormal information in the target abnormal dynamic scale.

[0140] 702. Calculate the abnormal detection accuracy based on the actual abnormal information, the total quantity information, and the hyperparameter.

[0141] 703. Adjust the hyperparameter based on the abnormal detection accuracy and the preset accuracy threshold.

[0142] In step 701, after step 204, i.e., after determining the target abnormal dynamic scale from the multiple dynamic scales based on the complex weight deviation information, the subject will send the relevant information corresponding to the target abnormal dynamic scale to the terminal of the corresponding maintenance personnel, which can include abnormal information and identification information of the dynamic scale, so that the maintenance personnel can locate the target abnormal dynamic scale according to the relevant information corresponding to the target abnormal dynamic scale and perform abnormal detection on the target abnormal dynamic scale. It should be noted that the target abnormal dynamic scale may be abnormal or normal, for example, due to some external factors, the target abnormal dynamic scale may appear abnormal for a period of time, but after the external factors are removed, the target abnormal dynamic scale automatically recovers. Therefore, when the maintenance personnel perform abnormal detection on the target abnormal dynamic scale, whether the target abnormal dynamic scale is abnormal or not, the actual detection result should be recorded, for example, the actual detection result can be normal or abnormal, and when it is abnormal, the abnormal point can be recorded in more detail, such as sensor damage, dynamic scale skew, etc.

[0143] In step 702, the abnormal detection accuracy is calculated based on the actual abnormal information, the total quantity information, and the hyperparameter, which can be specifically P = actual abnormal information a / (total quantity information n * hyperparameter b) * 100%, example 1, when the total quantity is 1000 dynamic scales, the hyperparameter is 0.005, and the actual abnormal information is 2, the abnormal detection accuracy P = 2 / (1000 * 0.005) * 100% = 40%, example 2, when the total quantity is 1000 dynamic scales, the hyperparameter is 0.005, and the actual abnormal information is 5, the abnormal detection accuracy P = 5 / (1000 * 0.005) * 100% = 100%.

[0144] ​In step 703, based on the anomaly detection accuracy and the preset accuracy threshold, the hyperparameter is adjusted. Specifically, when the anomaly detection accuracy is less than the preset accuracy threshold, the hyperparameter is adjusted to be smaller, and vice versa, when the anomaly detection accuracy is greater than the preset accuracy threshold, the hyperparameter is adjusted to be larger. For example, in example 1 of step 702, when the total number of dynamic scales is 1000, the hyperparameter is 0.005, and the actual abnormal information is 2, the anomaly detection accuracy P = 2 / (1000*0.005)*100% = 40%, and the preset accuracy threshold is 60%. At this time, the anomaly detection accuracy P is 40%, which is less than the accuracy threshold of 60%. Therefore, the hyperparameter can be adjusted to be smaller according to the preset proportion, for example, the original hyperparameter 0.005 is adjusted to 0.003. Further, for example, in example 2 of step 702, when the total number of dynamic scales is 1000, the hyperparameter is 0.005, and the actual abnormal information is 5, the anomaly detection accuracy P = 5 / (1000*0.005)*100% = 100%, and the preset accuracy threshold is 60%. At this time, the anomaly detection accuracy P is 100%, which is less than the accuracy threshold of 60%. Therefore, the hyperparameter can be adjusted to be larger according to the preset proportion, for example, the original hyperparameter 0.005 is adjusted to 0.007.

[0145] Therefore, it can be understood that the dynamic scale anomaly risk ranking output by the isolation forest model prompts the equipment engineer of the transfer site to check whether the dynamic scale determined as abnormal by the model actually has an abnormality. If there is an abnormality, the corresponding automatic freight recovery is suspended to avoid the occurrence of false recovery, and the relevant personnel are contacted to repair the equipment. If there is no problem, the dynamic scale continues to work normally. However, regardless of the result, the relevant test results need to be fed back. Here, the accuracy P of the model is calculated according to the feedback results. When the P value is high, the hyperparameter contamination can be appropriately increased to increase the proportion of abnormal scales determined. When the P value is low, the hyperparameter contamination can be appropriately reduced to reduce the proportion of abnormal scales determined. This is because the proportion of abnormal scales in all dynamic scales is also dynamically changing. Therefore, the feedback information from the site inspection can form a closed-loop optimization model.

[0146] In order to better implement the dynamic scale anomaly detection method in the embodiments of the present application, based on the dynamic scale anomaly detection method, a dynamic scale anomaly detection device is also provided in the embodiments of the present application, as shown in Figure 8 The dynamic scale anomaly detection device 800 includes a first acquisition unit 801, a first preprocessing unit 802, a first calculation unit 803, and a first determination unit 804.

[0147] The first acquisition unit 801 is configured to acquire complex weight information of a plurality of dynamic scales in a preset time range, wherein the complex weight information comprises identification information of each dynamic scale in the plurality of dynamic scales, total quantity information of the plurality of dynamic scales, and weight information of a package weighed by each dynamic scale and shipping order information corresponding to the package.

[0148] The first preprocessing unit 802 is configured to preprocess the complex weight information to obtain effective complex weight information of the plurality of dynamic scales.

[0149] The first calculation unit 803 is configured to calculate, based on the effective complex weight information, complex weight deviation information between each dynamic scale and other dynamic scales in the plurality of dynamic scales.

[0150] The first determination unit 804 is configured to determine, based on the complex weight deviation information, a target abnormal dynamic scale from the plurality of dynamic scales.

[0151] In a possible implementation of the present application, the first determination unit 804 specifically comprises:

[0152] The first calculation unit 803 is configured to calculate, based on the complex weight deviation information, an abnormal score value of each dynamic scale.

[0153] The second determination unit is configured to determine, based on the abnormal score value, a target abnormal dynamic scale from the plurality of dynamic scales.

[0154] In the embodiments of the present application, the complex weight information is preprocessed to obtain the effective complex weight information of the plurality of dynamic scales, and the effective complex weight information is specifically used for:

[0155] The preprocessing of the complex weight information to obtain the effective complex weight information of the plurality of dynamic scales comprises:

[0156] The complex weight information outside a preset weight threshold is filtered to obtain first complex weight information.

[0157] Invalid shipping order information and superimposed shipping order information in the shipping order information are identified, wherein the superimposed shipping order information is same shipping order information corresponding to at least two pieces of package logistics information.

[0158] The invalid shipping order information and the superimposed shipping order information are deleted from the first complex weight information to obtain second complex weight information.

[0159] The complex weight information only weighed by one dynamic scale in the second complex weight information is deleted to obtain the effective complex weight information of the plurality of dynamic scales.

[0160] In the embodiments of the present application, the first calculation unit 803 is specifically configured to:

[0161] Based on the shipping order information, the weight information and the identification information, a complex weight deviation ratio between each dynamic scale and other dynamic scales in the plurality of dynamic scales is calculated.

[0162] determine the complex weight deviation feature information based on the complex weight deviation ratio.

[0163] determine the complex weight deviation information between each dynamic scale and other dynamic scales in the plurality of dynamic scales based on the preset time window manner and the complex weight deviation feature information.

[0164] In the embodiments of the present application, the second calculation unit is specifically configured to:

[0165] input the complex weight deviation information into the pre-trained isolation forest model to obtain an abnormal score value of each dynamic scale.

[0166] In the embodiments of the present application, the second determination unit is specifically configured to:

[0167] sort the plurality of dynamic scales in descending order of abnormal risk based on the abnormal score value to obtain dynamic scale descending order sorting information.

[0168] determine the number of target abnormal dynamic scales according to the preset abnormal proportion hyperparameter and the total quantity information.

[0169] determine the target abnormal dynamic scale from the plurality of dynamic scales based on the number of target abnormal dynamic scales, the identification information, and the dynamic scale descending order sorting information.

[0170] In the embodiments of the present application, after determining the target abnormal dynamic scale from the plurality of dynamic scales based on the abnormal score value, the device is further configured to:

[0171] statistically determine actual abnormal information in the target abnormal dynamic scale.

[0172] calculate an abnormal detection accuracy rate based on the actual abnormal information, the total quantity information, and the hyperparameter.

[0173] adjust the hyperparameter based on the abnormal detection accuracy rate and a preset accuracy rate threshold.

[0174] The application provides a dynamic scale abnormality detection device 800, which comprises a first acquisition unit 801, which is used for acquiring complex weight information of a plurality of dynamic scales in a preset time range, wherein the complex weight information comprises identification information of each dynamic scale in the plurality of dynamic scales, total quantity information of the plurality of dynamic scales, and weight information of a package weighed by each dynamic scale and waybill information corresponding to the package; a first preprocessing unit 802, which is used for preprocessing the complex weight information to obtain effective complex weight information of the plurality of dynamic scales; a first calculation unit 803, which is used for calculating complex weight deviation information between each dynamic scale and other dynamic scales in the plurality of dynamic scales based on the effective complex weight information; and a first determination unit 804, which is used for determining a target abnormal dynamic scale from the plurality of dynamic scales based on the complex weight deviation information. Compared with the prior art, the embodiment of the application obtains effective complex weight information of the plurality of dynamic scales by preprocessing the complex weight information of the plurality of dynamic scales in the preset time range, then calculates complex weight deviation information between each dynamic scale and other dynamic scales in the plurality of dynamic scales according to the target complex weight information, and then determines a target abnormal dynamic scale that is more likely to be abnormal from the plurality of dynamic scales according to the complex weight deviation information, thereby avoiding the case that a large area of manual detection is performed on each dynamic scale, saving the cost, and reducing the risk of mis-sorting.

[0175] In addition to the above-mentioned dynamic scale abnormality detection method and device, the embodiment of the application further provides a computer device integrated with any one of the dynamic scale abnormality detection devices provided by the embodiment of the application, and the computer device comprises:

[0176] one or more processors;

[0177] a memory; and

[0178] one or more application programs, wherein the one or more application programs are stored in the memory and are configured to perform the operations of any one of the methods in any one of the dynamic scale abnormality detection method embodiments by the processor.

[0179] The embodiment of the application further provides a computer device integrated with any one of the dynamic scale abnormality detection devices provided by the embodiment of the application. Referring to Figure 9 , Figure 9 which is a structural schematic diagram of one embodiment of the computer device provided by the embodiment of the application.

[0180] As Figure 9 shown, the structure schematic diagram of the dynamic scale abnormality detection device designed by the embodiment of the application is shown, and specifically:

[0181] The dynamic scale anomaly detection device may include components such as a processor 901 with one or more processing cores, a storage unit 902 with one or more computer-readable storage media, a power supply 903, and an input unit 904. Those skilled in the art will understand that... Figure 9 The structure of the dynamic scale anomaly detection device shown does not constitute a limitation on the dynamic scale anomaly detection device. It may include more or fewer components than shown, or combine certain components, or have different component arrangements. Wherein:

[0182] The processor 901 is the control center of the dynamic scale anomaly detection device. It connects to various parts of the device via various interfaces and lines. By running or executing software programs and / or modules stored in the storage unit 902, and by calling data stored in the storage unit 902, it performs various functions and processes data, thereby providing overall monitoring of the dynamic scale anomaly detection device. Optionally, the processor 901 may include one or more processing cores; preferably, the processor 901 may integrate an application processor and a modem processor, wherein the application processor mainly handles the operating system, user interface, and applications, and the modem processor mainly handles wireless communication. It is understood that the modem processor may not be integrated into the processor 901.

[0183] Storage unit 902 can be used to store software programs and modules. Processor 901 executes various functional applications and data processing by running the software programs and modules stored in storage unit 902. Storage unit 902 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, application programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created based on the use of the dynamic scale anomaly detection device, etc. In addition, storage unit 902 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, storage unit 902 may also include a memory controller to provide processor 901 with access to storage unit 902.

[0184] The dynamic scale anomaly detection device also includes a power supply 903 that supplies power to various components. Preferably, the power supply 903 can be logically connected to the processor 901 through a power management system, thereby enabling functions such as charging, discharging, and power consumption management through the power management system. The power supply 903 may also include one or more DC or AC power supplies, recharging systems, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components.

[0185] The dynamic scale abnormality detection apparatus can further include an input unit 904 configured to receive inputted digital or character information and generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function controls.

[0186] Although not shown, the dynamic scale abnormality detection apparatus can further include a display unit, etc., which will not be described herein. In particular embodiments of the present application, the processor 901 in the dynamic scale abnormality detection apparatus loads one or more executable files corresponding to processes of one or more application programs into the storage unit 902 and runs the application programs stored in the storage unit 902 according to the following instructions, thereby implementing various functions, such as:

[0187] Obtaining complex weight information of a plurality of dynamic scales within a preset time range, the complex weight information including identification information of each dynamic scale in the plurality of dynamic scales, total quantity information of the plurality of dynamic scales, and weight information of a package weighed by each dynamic scale and shipping order information corresponding to the package; preprocessing the complex weight information to obtain valid complex weight information of the plurality of dynamic scales; calculating complex weight deviation information between each dynamic scale and other dynamic scales in the plurality of dynamic scales based on the valid complex weight information; and determining a target abnormal dynamic scale from the plurality of dynamic scales based on the complex weight deviation information.

[0188] The present application provides a dynamic scale abnormality detection method, the method comprising: obtaining complex weight information of a plurality of dynamic scales within a preset time range, the complex weight information including identification information of each dynamic scale in the plurality of dynamic scales, total quantity information of the plurality of dynamic scales, and weight information of a package weighed by each dynamic scale and shipping order information corresponding to the package; preprocessing the complex weight information to obtain valid complex weight information of the plurality of dynamic scales; calculating complex weight deviation information between each dynamic scale and other dynamic scales in the plurality of dynamic scales based on the valid complex weight information; and determining a target abnormal dynamic scale from the plurality of dynamic scales based on the complex weight deviation information. Compared with the prior art, the embodiments of the present application preprocess the complex weight information of the plurality of dynamic scales within a preset time range to obtain valid complex weight information of the plurality of dynamic scales, then calculate complex weight deviation information between each dynamic scale and other dynamic scales in the plurality of dynamic scales according to the target complex weight information, and then determine a target abnormal dynamic scale more likely to be abnormal from the plurality of dynamic scales according to the complex weight deviation information, thereby avoiding the case of detecting each dynamic scale by a large number of people, saving costs, and reducing the risk of mis-sorting.

[0189] To this end, an embodiment of the present application provides a computer readable storage medium, which can include a read only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc. The computer readable storage medium stores a plurality of instructions, which can be loaded by a processor to execute steps in any of the dynamic scale anomaly detection methods provided by the embodiments of the present application. For example, the instructions can execute the following steps:

[0190] obtaining complex weight information of a plurality of dynamic scales in a preset time range, the complex weight information including identification information of each dynamic scale in the plurality of dynamic scales, total quantity information of the plurality of dynamic scales, and weight information of a package weighed by each dynamic scale and shipping order information corresponding to the package; preprocessing the complex weight information to obtain effective complex weight information of the plurality of dynamic scales; calculating, based on the effective complex weight information, complex weight deviation information between each dynamic scale and other dynamic scales in the plurality of dynamic scales; and determining, based on the complex weight deviation information, a target abnormal dynamic scale from the plurality of dynamic scales.

[0191] In the above embodiments, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0192] The above describes in detail a dynamic scale anomaly detection method, device and readable storage medium provided by the embodiments of the present application. The principles and implementation manners of the present application are described by applying specific examples. The above embodiment description is only used to help understand the method and its core idea of the present application. Meanwhile, for those skilled in the art, the specific implementation manner and application range of the present application can be changed according to the idea of the present application. In summary, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A dynamic scale anomaly detection method, characterized in that, The method comprises: acquiring complex weight information of a plurality of dynamic scales in a preset time range, the complex weight information comprising identification information of each dynamic scale in the plurality of dynamic scales, total quantity information of the plurality of dynamic scales, and weight information of a package weighed by each dynamic scale and shipping order information corresponding to the package; preprocessing the complex weight information to obtain effective complex weight information of the plurality of dynamic scales; based on the shipping order information, the weight information, and the identification information in the effective complex weight information, calculating a complex weight deviation ratio between each dynamic scale and other dynamic scales in the plurality of dynamic scales; based on the complex weight deviation ratio, determining complex weight deviation feature information; based on a preset time window method and the complex weight deviation feature information, determining complex weight deviation information between each dynamic scale and other dynamic scales in the plurality of dynamic scales; based on the complex weight deviation information, calculating an abnormal score value of each dynamic scale; based on the abnormal score value, determining a target abnormal dynamic scale from the plurality of dynamic scales.

2. The dynamic scale anomaly detection method of claim 1, wherein, The preprocessing of the complex weight information to obtain the effective complex weight information of the plurality of dynamic scales comprises: filtering complex weight information with a weight information greater than a preset weight threshold to obtain first complex weight information; identifying invalid shipping order information and superimposed shipping order information in the shipping order information, wherein the superimposed shipping order information is the same shipping order information corresponding to at least two pieces of package logistics information; deleting the invalid shipping order information and the superimposed shipping order information from the first complex weight information to obtain second complex weight information; deleting complex weight information in the second complex weight information that is weighed by only one dynamic scale to obtain the effective complex weight information of the plurality of dynamic scales.

3. The dynamic scale anomaly detection method of claim 1 or 2, wherein The calculation of the abnormal score value of each dynamic scale based on the complex weight deviation information comprises: inputting the complex weight deviation information into a pre-trained isolation forest model to obtain the abnormal score value of each dynamic scale.

4. The dynamic scale anomaly detection method of claim 1, wherein, The determination of a target abnormal dynamic scale from the plurality of dynamic scales based on the abnormal score value comprises: based on the abnormal score value, sorting the plurality of dynamic scales in descending order of abnormal risk to obtain dynamic scale descending order sorting information; determining a target abnormal dynamic scale quantity based on a pre-set abnormal proportion hyperparameter and the total quantity information; based on the target abnormal dynamic scale quantity, the identification information, and the dynamic scale descending order sorting information, determining a target abnormal dynamic scale from the plurality of dynamic scales.

5. The dynamic scale anomaly detection method of claim 4, wherein, After determining a target abnormal dynamic scale from the plurality of dynamic scales based on the abnormal score value, the method further comprises: counting actual abnormal information in the target abnormal dynamic scale; calculating an abnormal detection accuracy rate based on the actual abnormal information, the total quantity information, and the hyperparameter; based on the abnormal detection accuracy rate and a preset accuracy rate threshold, adjusting the hyperparameter.

6. A dynamic scale abnormality detection device characterized by comprising: The device comprises: The first acquisition unit is configured to acquire complex weight information of a plurality of dynamic scales in a preset time range, wherein the complex weight information comprises identification information of each dynamic scale in the plurality of dynamic scales, total quantity information of the plurality of dynamic scales, and weight information of a package weighed by each dynamic scale and shipping order information corresponding to the package; The first preprocessing unit is configured to preprocess the complex weight information to obtain effective complex weight information of the plurality of dynamic scales; The first calculation unit is configured to calculate, based on the shipping order information, the weight information, and the identification information in the effective complex weight information, a complex weight deviation ratio between each dynamic scale and other dynamic scales in the plurality of dynamic scales; Based on the complex weight deviation ratio, determine complex weight deviation feature information; Based on a preset time windowing manner and the complex weight deviation feature information, determine complex weight deviation information between each dynamic scale and other dynamic scales in the plurality of dynamic scales; The first determination unit is configured to calculate, based on the complex weight deviation information, an abnormal score value of each dynamic scale; Based on the abnormal score value, determine a target abnormal dynamic scale from the plurality of dynamic scales.

7. A computer device, comprising: The computer device comprises: one or more processors; a memory; and one or more application programs, wherein the one or more application programs are stored in the memory and configured to be executed by the processor to implement the dynamic scale anomaly detection method of any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, A computer program is stored thereon, and the computer program is loaded by a processor to execute the steps in the dynamic scale anomaly detection method of any one of claims 1 to 5.

Images