Article labeling and information checking method and system in unmanned environment

Abstract

The invention discloses an article labeling and information checking method and system in an unmanned environment, and the method comprises the steps: obtaining labeling parameters and article information of an article labeling system in an operation process, taking the labeling parameters as a training set of a regression model, and taking the article information as an input set of the regression model; constructing a regression model, considering an association relationship between the historical training set and the current training set, and optimizing the regression model through hierarchical processing to output labeled article information; based on the article information, utilizing a machine learning model to obtain historical labeling information, predicting defect probability during future labeling, adjusting labeling parameters, outputting a labeled article image, and performing label identification on the image to complete checking of the article information; the labeling precision and consistency are remarkably improved through the parameter self-adaptive adjustment and defect prediction functions, the defect rate in the labeling process is effectively reduced, the system is made to adapt to various article types and environment requirements, and the labeling efficiency and the information checking accuracy are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing and information verification, and in particular to an object labeling and information verification method and system in an unmanned environment. Background Art

[0002] In recent years, with the rapid development of automation and intelligent technology, unmanned object labeling systems have gradually become an important part of industries such as logistics, manufacturing and warehousing. The research and development of unmanned object labeling systems aims to improve operational efficiency, reduce labor costs, and realize automatic data collection and processing during the labeling process to ensure accurate identification and information verification of items. Current labeling systems generally use basic image recognition technology to verify the information of labeled items, and combine parameters collected by a small number of sensors to monitor the labeling status. However, most unmanned object labeling systems rely on fixed parameters for labeling, lack the ability to flexibly respond to and adjust real-time data, resulting in limited accuracy and consistency of labeling when dealing with different types of items or labels. At the same time, due to the lack of a systematic feedback mechanism, it is difficult for unmanned object labeling systems to perform feedback optimization in each step of the labeling process to ensure labeling quality.

[0003] Specifically, the traditional unmanned item labeling system has the following problems: First, the lack of effective linkage analysis of historical data and real-time data makes it difficult to accurately predict defects in a constantly changing labeling environment, which in turn affects the stability of labeling quality. Second, because the traditional labeling system relies on fixed parameters for labeling, when there are slight environmental changes or differences in items, the labeling quality is easily affected, thereby reducing labeling efficiency. Third, in the information verification link after labeling is completed, the traditional method usually only considers the single verification of the labeling code against the information and ignores the negligence when manually entering the item information, resulting in the later item information verification process being too one-way, and usually the problem of "the labeling code is correct but the information is not correct" will occur.

[0004] Therefore, how to improve the autonomous adjustment and high-precision recognition of unmanned labeling systems has become a key challenge in current technology. Summary of the invention

[0005] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0006] In view of the above existing problems, the present invention is proposed. Therefore, the present invention provides an article labeling and information verification method in an unmanned environment to solve the problems mentioned in the background technology.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, the present invention provides an object labeling and information verification method in an unmanned environment, comprising:

[0009] Obtaining labeling parameters and item information of the item labeling system during operation, using the labeling parameters as a training set of a regression model, and using the item information as an input set of the regression model;

[0010] Constructing a regression model, considering the correlation between the historical training set and the current training set in the regression model, processing the regression model in a hierarchical manner based on the correlation, and outputting the labeled item information;

[0011] According to the labeled item information, a machine learning model is introduced, historical item labeling information is obtained through the machine learning model to predict the probability of defects that will occur when the item is labeled in the future, labeling parameters are adjusted using the defect probability, and an image of the labeled item is output;

[0012] The labeled object image is subjected to label recognition, and the object label and object information verification operation is completed according to the recognition result.

[0013] As a preferred solution of the method for labeling and verifying items in an unmanned environment of the present invention, obtaining labeling parameters of the item labeling system during operation includes:

[0014] The labeling parameters include initial labeling pressure, initial labeling speed and initial labeling angle of the object.

[0015] As a preferred solution of the method for object labeling and information verification in an unmanned environment of the present invention, a regression model is constructed, and the correlation between the historical training set and the current training set in the regression model is considered, including:

[0016] According to the historical training set, a nonlinear influence coefficient of labeling pressure-angle is defined. If the labeling angle of the items in the current training set changes, the nonlinear influence coefficient is calculated to obtain the labeling pressure of the items at the current labeling angle.

[0017] The labeling pressure of the item at the current labeling angle is taken as the standard value. If the initial labeling pressure does not exceed the labeling pressure at the current labeling angle, the initial labeling pressure is updated. Otherwise, the offset of the item at the labeling angle is calculated in combination with the labeling angle of the current item.

[0018] As a preferred solution of the method for labeling and verifying items in an unmanned environment of the present invention, it also includes:

[0019] According to the labeling pressure of the item at the current labeling angle and the offset of the item at the labeling angle, considering the influence of the labeling speed on the current labeling pressure and labeling angle, the combination schemes of labeling pressure and labeling angle at different labeling speeds are calculated.

[0020] As a preferred solution of the method for labeling and verifying items in an unmanned environment of the present invention, the regression model is processed in a hierarchical manner based on the association relationship to output the information of the labeled items, including:

[0021] Based on the combination of labeling pressure and labeling angle at different labeling speeds, the item information is divided into multiple levels;

[0022] In each level, the corresponding combination scheme is assigned according to the material and geometric shape in the item information, and the item labeling quality coefficient of each level is generated at the same time. The items are sorted from high to low according to the quality coefficient, and the labeled item information is output in the order of the sorted quality coefficients.

[0023] As a preferred solution of the method for labeling and verifying items in an unmanned environment of the present invention, a machine learning model is introduced according to the information of the labeled items, and the historical item labeling information is obtained through the machine learning model to predict the probability of defects that will occur when the items are labeled in the future, and the labeling parameters are adjusted using the defect probability, and the labeled item image is output, including:

[0024] Based on the output labeled item information, the defect probability of the current item when labeled is obtained, and at the same time, the labeling information of historical items is obtained through a machine learning model for reverse prediction to obtain the defect probability of the historical items when labeled;

[0025] According to the defect probability of the historical items when labeled and the defect probability of the current items when labeled, the defect probability of the items when labeled in the future is obtained;

[0026] A range interval is defined for the defect probability of the item when it is labeled in the future, and the labeling parameters are adjusted based on the upper and lower limits of each range interval. The adjusted labeling parameters are embedded in the image metadata, and the item image after the labeling parameters are adjusted is output.

[0027] As a preferred solution of the method for labeling and verifying information of items in an unmanned environment of the present invention, the label recognition of the labeled item image includes:

[0028] Identify whether the label information in each item image after labeling is consistent with the label information of the corresponding item in the item label database. If the label information in the item image after labeling or the label information of the corresponding item in the item label database is inconsistent, it will be marked as "item information verification failed" and notify manual review.

[0029] In a second aspect, the present invention provides an article labeling and information verification system in an unmanned environment, which comprises:

[0030] A labeling parameter and item information acquisition module, configured to obtain labeling parameters and item information of the item labeling system during operation, use the labeling parameters as a training set of a regression model, and use the item information as an input set of the regression model;

[0031] A regression model building and hierarchical processing module is configured to build a regression model, consider the association between the historical training set and the current training set in the regression model, process the regression model in a hierarchical processing manner based on the association, and output the labeled item information;

[0032] A defect prediction and labeling parameter optimization module is configured to introduce a machine learning model according to the labeled object information, obtain historical object labeling information through the machine learning model to predict the probability of defects occurring in the future labeling of the object, adjust the labeling parameters using the defect probability, and output the labeled object image;

[0033] The image recognition and information verification module is configured to perform label recognition on the labeled object image and complete the object labeling and object information verification operations according to the recognition result.

[0034] In a third aspect, the present invention provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, wherein: the processor implements any step of the above method when executing the computer program.

[0035] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: the computer program implements any step of the above method when executed by a processor.

[0036] Compared with the prior art, the invention has the following beneficial effects:

[0037] 1. The present invention constructs a regression model and adopts a hierarchical processing method in the regression model, that is, automatically matches the optimal labeling parameter combination scheme according to the type and specification of the item, so that the model can adapt to the characteristics of various items; through the hierarchical output method, the labeling system can label according to the specific needs of each type of item, avoiding the problem of labeling quality fluctuation caused by fixed parameters, and improving the flexibility and stability of the labeling system in processing diversified items;

[0038] 2. The machine learning model combines the real-time information and historical labeling data during the labeling process to predict the probability of defects in future labeling, and adjusts the labeling parameters in real time based on the prediction results, making up for the shortcomings of traditional labeling systems that lack predictive capabilities and are difficult to make preventive adjustments;

[0039] 3. The traditional system lacks an effective feedback mechanism, which makes it impossible for subsequent labeling to draw on previous parameter adjustment experience. The present invention uses a closed-loop feedback design so that each labeling result provides an optimization basis for the next labeling, thereby realizing self-adjustment of the labeling system, allowing the labeling system to continuously optimize the parameter adjustment strategy in each feedback, and ultimately achieving a high-precision labeling process in an unmanned environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0041] Figure 1 This is a general flow chart of a method for labeling and verifying items in an unmanned environment according to an embodiment of the present invention;

[0042] Figure 2 This is a model loss convergence curve diagram of the object labeling and information verification method in an unmanned environment described in an embodiment of the present invention. DETAILED DESCRIPTION

[0043] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0044] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0045] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0046] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0047] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0048] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0049] Example 1

[0050] Reference Figure 1 , which is the first embodiment of the present invention, provides an object labeling and information verification method in an unmanned environment, comprising:

[0051] S1. Obtain labeling parameters and item information of the item labeling system during operation, use the labeling parameters as a training set of the regression model, and use the item information as an input set of the regression model;

[0052] Specifically, the labeling parameters include the initial labeling pressure of the object, the initial labeling speed, and the initial labeling angle;

[0053] Specifically, the item information includes the item ID, item name, item material, and item geometry;

[0054] Furthermore, the labeling parameters (pressure, speed, angle) can be used as a training set, and the object information (geometric structure, material) can be used as an input set to establish a parameter association relationship between the two, so as to achieve customized labeling for different objects;

[0055] It should be explained that the establishment of a parameter association relationship between the two can adjust the labeling parameters according to the differences in information of different items; for example, larger or smoother items require higher labeling pressure, while fragile items require lower labeling pressure; and the parameter association relationship can be used as historical structured data during future model training, thereby providing training support for the data set;

[0056] S2. Build a regression model, consider the correlation between the historical training set and the current training set in the regression model, process the regression model in a hierarchical manner based on the correlation, and output the labeled item information;

[0057] It needs to be explained that since labeling pressure and labeling angle are two important parameters that affect the quality of labeling of items, and the nonlinear relationship between them is rarely mentioned, the nonlinear influence coefficient of labeling pressure-angle can be obtained by modeling the nonlinear relationship between labeling pressure and labeling angle;

[0058] Furthermore, based on the historical training set, the nonlinear relationship between labeling pressure and labeling angle is modeled to define the nonlinear influence coefficient of labeling pressure-angle. If the labeling angle of the item in the current training set changes, the nonlinear influence coefficient is calculated to obtain the labeling pressure of the item at the current labeling angle.

[0059] Specifically, if the labeling angle changes, the labeling pressure needs to be reduced and the reduced labeling pressure needs to be recorded to prevent the label from being deformed or wrinkled due to excessive pressure. If the labeling angle remains unchanged, the previous labeling pressure needs to be restored based on the reduced labeling pressure to ensure that the label can be attached to the surface of the object.

[0060] Furthermore, the labeling pressure of the item at the current labeling angle is taken as the standard value. If the initial labeling pressure does not exceed the labeling pressure at the current labeling angle, the initial labeling pressure is updated. Otherwise, the offset of the item at the labeling angle is calculated in combination with the labeling angle of the current item.

[0061] It should be noted that if the initial pressure does not exceed the labeling pressure at the current labeling angle, it means that the initial labeling pressure is insufficient to complete the labeling operation of the current item, that is, there will be a problem of instability in the labeling operation, so the initial labeling pressure needs to be updated to the pressure required for labeling the current item; if the current initial labeling pressure exceeds the labeling pressure at the current labeling angle, then it is necessary to consider whether the labeling angle of the current item is offset, that is, whether the force point of the item for the labeling pressure changes at the labeling angle, so as to calculate the offset of the item at the labeling angle;

[0062] Furthermore, according to the labeling pressure of the item at the current labeling angle and the offset of the item at the labeling angle, the influence of the labeling speed on the current labeling pressure and labeling angle is considered, and the combination schemes of labeling pressure and labeling angle at different labeling speeds are calculated;

[0063] It should be noted that by combining labeling pressure and labeling angle at different labeling speeds, the optimal labeling parameters of the items can be obtained, which can improve the operating efficiency of the actual labeling system without affecting the labeling quality of the items, and shorten the execution time of each item labeling action, while avoiding the additional load of the labeling system caused by unnecessary high pressure or too low speed;

[0064] Furthermore, based on the combination schemes of labeling pressure and labeling angle at different labeling speeds, the item information is divided into multiple levels;

[0065] It should be noted that there is a unique combination scheme in each level, which includes labeling speed, labeling pressure and labeling angle;

[0066] Furthermore, in each level, a corresponding combination scheme is assigned according to the material and geometric shape in the item information, and a labeling quality coefficient of the items in each level is generated at the same time, and the items are sorted from high to low according to the quality coefficient, and the labeled item information is output according to the sorted quality coefficient order;

[0067] Specifically, the corresponding combination scheme is allocated according to the material and geometric shape in the item information, that is, the labeling speed, labeling pressure and labeling angle corresponding to the item are allocated according to the material and geometric shape of the item; for example, if the current item is a stainless steel thermos cup, then the corresponding item material is hard metal material, and the geometric shape is cylindrical, then the corresponding labeling angle is 6°, the labeling pressure is 50N (moderate pressure, ensuring that the label can cover the cylindrical surface without bubbles), and the labeling speed is 1.2m / s;

[0068] It should be noted that by sorting and outputting the items according to the labeling quality coefficient, valuable items can be inspected first during the labeling quality inspection process;

[0069] S3. Based on the information of the labeled items, a machine learning model is introduced to obtain the historical item labeling information through the machine learning model to predict the probability of defects that will occur when the items are labeled in the future. The defect probability is used to adjust the labeling parameters and output the image of the labeled items.

[0070] Furthermore, based on the output of the labeled item information, the defect probability of the current item when labeled is obtained. At the same time, the labeling information of historical items is obtained through a machine learning model for reverse prediction to obtain the defect probability of historical items when labeled.

[0071] Furthermore, the defect probability of the item when it is labeled in the future is obtained based on the defect probability of the historical item when it is labeled and the defect probability of the current item when it is labeled;

[0072] Furthermore, a range interval is defined for the defect probability of the item when it is labeled in the future, and the labeling parameters are adjusted based on the upper and lower limits of each range interval, and the adjusted labeling parameters are embedded in the image metadata, and the item image after the labeling parameters are adjusted is output;

[0073] Specifically, the range interval is defined as follows:

[0074] Rule 1: The defect probability range is between 0% and 2% (including 0 but excluding 2), the labeling parameters do not need to be adjusted, and the labeling speed remains the same;

[0075] Rule 2: If the defect probability range is between 2% and 5% (including 2 and excluding 5), the labeling pressure needs to be increased or decreased by 5% based on the original labeling pressure, the labeling angle needs to be adjusted by ±1°, and the labeling speed needs to remain the original value;

[0076] Rule 3: If the defect probability range is between 5% and 10% (including 5 and excluding 10), the labeling pressure needs to be increased or decreased by 10% based on the original labeling pressure, the labeling angle needs to be adjusted by ±2°, and the labeling speed needs to be reduced by 10%;

[0077] Rule 4: If the defect probability range is greater than 10%, the labeling pressure needs to be increased or decreased by 15% based on the original labeling pressure, the angle needs to be adjusted by ±4°, and the labeling speed needs to be reduced by 15%;

[0078] It should be noted that the output image of the item after the labeling parameters are adjusted includes: labeling angle, labeling pressure, labeling speed, defect probability of item labeling, and labeling quality coefficient;

[0079] S4, performing label recognition on the labeled object image, and completing the verification operation of the object label and the object information according to the recognition result;

[0080] Further, identify whether the label information in each labeled object image is consistent with the label information of the corresponding object in the object label database. If the label information in the object image after labeling or the label information of the corresponding object in the object label database is inconsistent, it is marked as "object information verification failed" and a manual review is notified;

[0081] Specifically, if the label information in the item image after labeling is inconsistent with the label information of the corresponding item in the database, there are two situations. One situation is that the label information in the item image after labeling is wrong, but the label information of the corresponding item in the item label database exists, which means that the item information after labeling is wrong and the item information needs to be re-verified. The other situation is that the label information in the item image after labeling is correct, but the label information of the corresponding item in the database cannot be found, which means that the item label database has not entered its item information, which means that the item information needs to be supplemented.

[0082] It should be noted that by identifying the tag information, the frequent manual verification of item information is reduced, and efficient and accurate information verification can be achieved even in an unmanned environment.

[0083] Furthermore, this embodiment also provides an item labeling and information verification system in an unmanned environment, comprising:

[0084] A labeling parameter and item information acquisition module, configured to obtain labeling parameters and item information of the item labeling system during operation, use the labeling parameters as a training set of a regression model, and use the item information as an input set of the regression model;

[0085] A regression model building and hierarchical processing module is configured to build a regression model, consider the association between the historical training set and the current training set in the regression model, process the regression model in a hierarchical processing manner based on the association, and output the labeled item information;

[0086] A defect prediction and labeling parameter optimization module is configured to introduce a machine learning model according to the labeled object information, obtain historical object labeling information through the machine learning model to predict the probability of defects occurring in the future labeling of the object, adjust the labeling parameters using the defect probability, and output the labeled object image;

[0087] The image recognition and information verification module is configured to perform label recognition on the labeled object image and complete the object labeling and object information verification operations according to the recognition result.

[0088] This embodiment also provides a computer device, which is applicable to the method for labeling and verifying items in an unmanned environment, and includes:

[0089] Memory and processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the object labeling and information verification method in an unmanned environment as proposed in the above embodiment.

[0090] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0091] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for object labeling and information verification in an unmanned environment proposed in the above embodiment is implemented.

[0092] The storage medium proposed in this embodiment and the data storage method proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be found in the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0093] Example 2

[0094] Reference Figure 2, which is the second embodiment of the present invention, and provides a method for labeling and verifying information of items in an unmanned environment, including: further verifying the beneficial effects of the scheme of the present invention by means of simulation experiments, the simulation experiment is carried out in a simulated unmanned storage environment, the experimental environment temperature is controlled at 25°C, and the relative humidity is about 50% to simulate the real storage environment; three types of typical containers are selected as test objects, namely plastic bottles, metal cans and glass bottles; in order to verify the applicability of the present invention on items of various materials and shapes, containers of different materials and smoothness are specially selected; the adhesion difficulty of the labels of these items is high and the labeling accuracy requirements are high, which can effectively verify the performance of the labeling system; by recording the labeling accuracy, defect probability and information verification accuracy of various items under different labeling conditions, and conducting comparative analysis; in addition, in order to evaluate the improvement effect of the labeling system, the test also records the performance of the existing labeling system under the same conditions; the data in Table 1 are as follows;

[0095] Table 1

[0096]

[0097]

[0098] Firstly, the tabular data show that, among the three test objects, the labeling system based on the content of the present invention significantly improves the labeling accuracy; among them, the labeling accuracy of plastic bottles is improved from 89.1% of the existing system to 96.2%, that of metal cans is improved from 88.4% to 97.3%, and that of glass bottles is improved from 90.0% to 98.7%. The advantage of the improvement lies in the consideration of the nonlinear relationship between labeling pressure and angle, so that it can automatically adjust the labeling parameters according to the surface characteristics and shape of the object, avoiding the problem of wrinkling or displacement of the label due to excessive pressure; in addition, the defect probability is significantly reduced from 6.9% of the existing labeling system to 1.0%; in metal cans, it is reduced from 6.4% to 0.8%; in glass bottles, it is reduced from 5.7% to 0.6%; it shows that the defect probability prediction and real-time parameter optimization functions of the present invention can effectively reduce the undesirable phenomenon of labels in the labeling process, especially accurately predicting the probability of defect occurrence in the machine learning model and dynamically optimizing the parameters, so that the label can be stably and firmly attached to the surface of different objects;

[0099] Secondly, based on the tabular data, the machine learning model is trained through the labeling parameters (pressure, angle, speed) in the scheme of the present invention and the obtained labeling effect (labeling accuracy, defect probability), and the convergence curve of the loss function value with the number of iterations is used to show the training effect of the machine learning model. Figure 2The loss convergence curve shows that after the machine learning model continuously adjusts the labeling parameters, the loss value gradually decreases and eventually tends to converge, indicating that the model is stable and robust in controlling accuracy and defect rate, further verifying the feasibility of the solution of the present invention.

[0100] Those skilled in the art will appreciate that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the application can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program codes. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0101] The present application is described 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 application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or 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, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0102] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0103] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0104] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0105] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for labeling and verifying items in an unmanned environment, characterized in that: include: Obtaining labeling parameters and item information of the item labeling system during operation, using the labeling parameters as a training set of a regression model, and using the item information as an input set of the regression model; Constructing a regression model, considering the correlation between the historical training set and the current training set in the regression model, processing the regression model in a hierarchical manner based on the correlation, and outputting the labeled item information; According to the labeled item information, a machine learning model is introduced, historical item labeling information is obtained through the machine learning model to predict the probability of defects that will occur when the item is labeled in the future, labeling parameters are adjusted using the defect probability, and an image of the labeled item is output; The labeled object image is subjected to label recognition, and the object label and object information verification operation is completed according to the recognition result.

2. The method for labeling and verifying items in an unmanned environment as claimed in claim 1, characterized in that: Obtain the labeling parameters of the item labeling system during operation, including: The labeling parameters include initial labeling pressure, initial labeling speed and initial labeling angle of the object.

3. The method for labeling and verifying items in an unmanned environment as claimed in claim 2, characterized in that: Constructing a regression model, taking into account the association between the historical training set and the current training set in the regression model, including: According to the historical training set, a nonlinear influence coefficient of labeling pressure-angle is defined. If the labeling angle of the items in the current training set changes, the nonlinear influence coefficient is calculated to obtain the labeling pressure of the items at the current labeling angle. The labeling pressure of the item at the current labeling angle is taken as the standard value. If the initial labeling pressure does not exceed the labeling pressure at the current labeling angle, the initial labeling pressure is updated. Otherwise, the offset of the item at the labeling angle is calculated in combination with the labeling angle of the current item.

4. The method for labeling and verifying items in an unmanned environment as claimed in claim 2 or 3, characterized in that: Also includes: According to the labeling pressure of the item at the current labeling angle and the offset of the item at the labeling angle, considering the influence of the labeling speed on the current labeling pressure and labeling angle, the combination schemes of labeling pressure and labeling angle at different labeling speeds are calculated.

5. The method for labeling and verifying items in an unmanned environment as claimed in claim 4, characterized in that: Based on the association relationship, the regression model is processed in a hierarchical manner to output the labeled item information, including: Based on the combination of labeling pressure and labeling angle at different labeling speeds, the item information is divided into multiple levels; In each level, the corresponding combination scheme is assigned according to the material and geometric shape in the item information, and the item labeling quality coefficient of each level is generated at the same time. The items are sorted from high to low according to the quality coefficient, and the labeled item information is output in the order of the sorted quality coefficients.

6. The method for labeling and verifying items in an unmanned environment as claimed in claim 2 or 5, characterized in that: According to the labeled item information, a machine learning model is introduced, historical item labeling information is obtained through the machine learning model to predict the probability of defects that will occur when the item is labeled in the future, labeling parameters are adjusted using the defect probability, and an image of the labeled item is output, including: Based on the output labeled item information, the defect probability of the current item when labeled is obtained, and at the same time, the labeling information of historical items is obtained through a machine learning model for reverse prediction to obtain the defect probability of the historical items when labeled; According to the defect probability of the historical items when labeled and the defect probability of the current items when labeled, the defect probability of the items when labeled in the future is obtained; A range interval is defined for the defect probability of the item when it is labeled in the future, and the labeling parameters are adjusted based on the upper and lower limits of each range interval. The adjusted labeling parameters are embedded in the image metadata, and the item image after the labeling parameters are adjusted is output.

7. The method for labeling and verifying items in an unmanned environment as claimed in claim 6, characterized in that: Performing label recognition on the labeled object image includes: Identify whether the label information in each item image after labeling is consistent with the label information of the corresponding item in the item label database. If the label information in the item image after labeling or the label information of the corresponding item in the item label database is inconsistent, it will be marked as "item information verification failed" and notify manual review.

8. An item labeling and information verification system in an unmanned environment, based on the item labeling and information verification method in an unmanned environment according to any one of claims 1 to 7, characterized in that: include: A labeling parameter and item information acquisition module, configured to obtain labeling parameters and item information of the item labeling system during operation, use the labeling parameters as a training set of a regression model, and use the item information as an input set of the regression model; A regression model building and hierarchical processing module is configured to build a regression model, consider the association between the historical training set and the current training set in the regression model, process the regression model in a hierarchical processing manner based on the association, and output the labeled item information; A defect prediction and labeling parameter optimization module is configured to introduce a machine learning model according to the labeled object information, obtain historical object labeling information through the machine learning model to predict the probability of defects occurring in the future labeling of the object, adjust the labeling parameters using the defect probability, and output the labeled object image; The image recognition and information verification module is configured to perform label recognition on the labeled object image and complete the object labeling and object information verification operations according to the recognition result.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images