CN112836471B - Batch labeling interface automation method and system - Google Patents

Abstract

The invention provides a batch labeling interface automation method and system, which are based on a jmeter as a development tool and comprise the following steps: verifying login; after logging in, constructing a tagging task list taskid; and executing the labeling work and the auditing work according to the labeling task list. The invention solves the problem that the interface between the simulation client and the server of the project is automated in the testing stage, and realizes the condition that a tester cannot mark a large number of picture defects in the industrial MEB testing process.

Description

Batch labeling interface automation method and system

Technical Field

The invention relates to the field of intelligent batch labeling, in particular to a batch labeling interface automation method and system, and especially relates to an industrial MEB (Manufacturing Executive Brain) feature engineering center batch labeling interface automation method and system.

Background

At present, quality detection links of a large number of enterprises in the industrial field are realized by manual quality inspection, namely, defect identification is mainly performed manually. However, manual quality inspection is difficult to standardize, and has high missing inspection and false inspection rates and high labor cost.

In addition, a small number of enterprises use traditional machine vision techniques for quality inspection. For example, patent document CN109584250a discloses a robust method for automatically dividing and labeling visual regions, in which a fourier polar logarithm coordinate correlation method is used to obtain a rotation angle and a scaling ratio between an image to be matched and a reference image to obtain a primary matching image, and then a secondary matching method is used to further match each region to obtain a final matching image. However, the traditional machine vision has high dependency on the object regularity and cannot realize modeling sharing.

Disclosure of Invention

Aiming at the defects in the prior art, the invention aims to provide a batch labeling interface automation method and system. In particular to a batch labeling interface automation method and system for quality detection.

The invention provides a batch labeling interface automation method based on a jmeter as a development tool, which comprises the following steps:

step SA: verifying login;

step SB: after logging in, constructing a tagging task list taskid;

step SC: and executing the labeling work and the auditing work according to the labeling task list.

Preferably, the step SA includes:

step SA1: initializing project management information;

step SA2: managing login information according to items in the login request information, and introducing a jar packet to perform base64 signature verification; if the signature verification fails, logging out; if the signature verification is successful, triggering the step SA3 to execute;

step SA3: obtaining a sample picture;

step SA4: storing the sample picture by mysql;

step SA5: screening the sample picture;

step SA6: obtaining a screened picture, and recording the picture as a picture to be marked;

step SA7: creating an annotation task according to the picture to be annotated;

step SA8: and generating an annotation task list according to the annotation task.

Preferably, the step SB includes:

step SB1: labeling the picture to be labeled according to the operation of a labeling operator aiming at the labeling task list; wherein, the annotator is the name of the virtual role which is automatically annotated by the computer;

step SB2: the method comprises the steps that a label of a label operator to-be-labeled picture is stored, and a labeled picture is obtained;

step SB3: creating an audit task according to the marked pictures;

step SB4: the method comprises the following steps of (1) aligning an interface request, namely whether the labeling results of auditors in an audit task are aligned or not is auditors, wherein the auditors are names of virtual roles for automatically executing audit by a computer;

step SB5: receiving an auditing result of an auditor; if the result of the verification is passed, the marked picture is considered to be marked successfully; and if the examination is not passed, the annotator modifies the annotation and then carries out examination again until the examination is passed.

Preferably, the first and second electrodes are formed of a metal,

the project management information comprises public variables, the public variables defined by a user are set through a setUp thread group for interface calling, and initialization data are transmitted to the public variables;

switching login items according to the item management information received from the login requester; acquiring an access token _ token verification request for item switching, referring to a base64 encrypted jar packet of three parties for signature verification and setting parameters after base64 encryption as a global state; setting a Transaction Controller to manage a use case set of each interface;

setting a PreProcessor BeanShell PreProcessor, and presetting base64 encryption and signature verification for requesting calling; respectively setting a preprocessor admin for a annotator;

according to a request of a marker for acquiring a marking task list, returning a directory identifier directoryId, a picture identification number imageId and a file name fileName of a picture to be marked to create a marking task, and returning a marking task list taskId to perform a marking workflow; parameterizing a picture identification number imageId through a CSV Data Set Config element;

setting index to count and store the request times of the labeling interface, and using count function as parameter to control the cycle times;

through an HTTP Header Manager, the Header information of the image marking, the mark storage and the submission of the interface checking request is stored, and the format of the interface response data is specified;

after the interface integrity request is edited, introducing a Loop Controller circulation Controller to control and store the labeling interface;

setting a Listener Listener; and using the View Results Tree to check a result Tree and check the log information requested by the interface.

Preferably, the script is developed through the bearer shell, the jar packet is introduced to realize encryption verification of the interface, the pictures to be annotated are encrypted through the jar packet, and a user is allowed to realize import of a batch of pictures to be annotated.

The invention provides a batch labeling interface automation system based on a jmeter as a development tool, which comprises:

and a module MA: verifying login;

a module MB: after logging in, constructing a tagging task list taskid;

and a module MC: and executing the labeling work and the auditing work according to the labeling task list.

Preferably, the module MA comprises:

module MA1: initializing project management information;

and a module MA2: managing login information according to items in the login request information, and introducing a jar packet to perform base64 signature verification; if the signature verification fails, logging out; if the signature verification is successful, triggering the module MA3 to execute;

module MA3: acquiring a sample picture;

module MA4: storing the sample picture by mysql;

module MA5: screening the sample picture;

module MA6: obtaining a screened picture, and recording the picture as a picture to be marked;

module MA7: creating an annotation task according to the picture to be annotated;

module MA8: and generating an annotation task list according to the annotation task.

Preferably, said module MB comprises:

the module MB1: labeling the picture to be labeled according to the operation of a labeling operator aiming at the labeling task list; wherein, the annotator is the name of the virtual role which is automatically annotated by the computer;

the module MB2: saving the label of the picture to be labeled by a label maker to obtain a labeled picture;

the module MB3: creating an audit task according to the marked pictures;

the module MB4: the method comprises the following steps of (1) aligning an interface request, namely whether the labeling results of auditors in an audit task are aligned or not is auditors, wherein the auditors are names of virtual roles for automatically executing audit by a computer;

the module MB5: receiving an auditing result of an auditor; if the result of the verification is passed, the marked picture is considered to be marked successfully; and if the examination is not passed, the annotation is modified by the annotator and then is examined again until the examination is passed.

Preferably, the first and second liquid crystal display panels are,

the project management information comprises public variables, the public variables defined by a user are set through a setUp thread group for interface calling, and initialization data are transmitted to the public variables;

switching login items according to the item management information received from the login requester; acquiring an access token _ token verification request for item switching, referring to a base64 encrypted jar packet of three parties for signature verification and setting parameters after base64 encryption as a global state; setting a Transaction Controller to manage a use case set of each interface;

setting a PreProcessor BeanShell PreProcessor, and presetting base64 encryption and signature verification for requesting calling; respectively setting a preprocessor admin for a annotator;

according to a request of a marker for acquiring a marking task list, returning a directory identifier directoryId, a picture identification number imageId and a file name fileName of a picture to be marked to create a marking task, and returning a marking task list taskId to perform a marking workflow; parameterizing a picture identification number imageId through a CSV Data Set Config element;

setting index to count and store the request times of the labeling interface, and using count function as parameter to control the cycle times;

through an HTTP Header Manager, the Header information of the image marking, the mark storage and the submission of the interface checking request is stored, and the format of the interface response data is specified;

after the interface integrity request is edited, introducing a Loop Controller circulation Controller to control and store the labeling interface;

setting a Listener monitor; and using the View Results Tree to check a result Tree and check the log information requested by the interface.

Preferably, the script is developed through the benshell, the jar packet is introduced to realize encryption verification of the interface, the pictures to be marked are encrypted through the jar packet, and a user is allowed to realize importing of batch pictures to be marked.

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

the invention uses AI technology to carry out quality inspection, makes up the defects of the traditional machine vision technology and manual quality inspection, realizes the standardization and the modeling of the quality inspection, saves the investment of the quality inspection (comprising labor cost, time cost and the like) for enterprises, improves the accuracy and the consistency of the quality inspection, and can provide a customized test scene for a test team.

According to the invention, java is used as a basic development language, a meter (interface and pressure testing tool) is used as a basic development tool, a series of processes of examining and checking pictures (obtaining directoryId and imageId), creating a labeling task, constructing a labeling task list (obtaining taskId), allowing a annotator to label, allowing the annotator to submit an audit, allowing the auditor to align a request interface, allowing the auditor to examine and pass and the like are simulated under a real environment, and great convenience is provided for a manual batch labeling test of a simulated product.

According to the invention, the simulation client sends a request to the server, and the single thread and the multithreading send the sorted sample picture or the picture data to be marked to the back-end server serving as the server, so that the automation of an interface between the simulation client and the server of a project in a test stage is solved, and the condition that a marker cannot mark a large number of picture defects during an industrial MEB test is realized.

Drawings

Other features, objects and advantages of the invention will become more apparent upon reading of the detailed description of non-limiting embodiments with reference to the following drawings:

FIG. 1 is a schematic flow chart of the method of the present invention.

Detailed Description

The present invention will be described in detail with reference to specific examples. The following examples will assist those skilled in the art in further understanding the invention, but are not intended to limit the invention in any way. It should be noted that it would be obvious to those skilled in the art that various changes and modifications can be made without departing from the spirit of the invention. All falling within the scope of the present invention.

The jmeter tool cannot directly simulate batch marking test of pictures, the self-contained function of the jmeter tool cannot realize the defects in the batch marking of the pictures, marking is difficult to complete in a short time, in order to save time and cost, a script needs to be developed secondarily, a jar packet is introduced to realize encryption verification (signature verification) of an interface, the script is developed by means of the benshell, and the jar packet is introduced. Therefore, the invention relies on some characteristics of the jmeter tool to carry out secondary development, increases jar package introduction, completes script development by the benshell, completes batch marking, and solves the difficult point that marking personnel and testing personnel are difficult to manually mark defect tasks in a large scale in a short time.

The invention provides a batch labeling interface automation method based on a jmeter as a development tool, which comprises the following steps:

step SA: verifying login;

step SB: after logging in, constructing a tagging task list taskid;

step SC: and executing the labeling work and the auditing work according to the labeling task list.

Preferably, the step SA includes:

step SA1: initializing project management information;

step SA2: managing login information according to items in the login request information, and introducing a jar packet to perform base64 signature verification; if the signature verification fails, logging out; if the signature verification is successful, triggering the step SA3 to execute;

step SA3: obtaining a sample picture;

step SA4: storing the sample picture by mysql;

step SA5: screening the sample picture;

step SA6: obtaining a screened picture, and recording the picture as a picture to be marked; in a preferred example, the standard for screening pictures can be to screen out pictures which are not in compliance, such as the definition of the pictures is lower than a definition threshold value, and sample pictures can also be used as pictures to be labeled; the picture can be screened according to whether the ID of the picture meets the set condition.

Step SA7: creating an annotation task according to the picture to be annotated;

step SA8: and generating an annotation task list according to the annotation task.

Preferably, the step SB includes:

step SB1: labeling the picture to be labeled according to the operation of a labeling operator aiming at the labeling task list; the annotator is the name of the virtual role which is automatically annotated by the computer;

step SB2: saving the label of the picture to be labeled by a label maker to obtain a labeled picture;

step SB3: creating an audit task according to the marked pictures;

step SB4: the method comprises the following steps of (1) aligning an interface request, namely whether the labeling results of auditors in an audit task are aligned or not is auditors, wherein the auditors are names of virtual roles for automatically executing audit by a computer; in a preferred embodiment, the defect list marked by the marker a is compared with the defect list marked by the marker B one by one, and if the comparison result exceeds a set ratio, for example 75%, the following conditions are met, then the marking results of the marker a and the marker B are considered to be aligned:

1) The types of defects are consistent;

2) The description of the defects is consistent. Therefore, the method can assist the auditor to audit the labeling results of a plurality of annotators,

step SB5: receiving an auditing result of an auditor; if the result of the verification is passed, the marked picture is considered to be marked successfully; and if the examination is not passed, the annotator modifies the annotation and then carries out examination again until the examination is passed.

Preferably, the first and second electrodes are formed of a metal,

the project management information comprises public variables, the public variables defined by a user are set through the setUp thread group for interface calling, and initialization data are transmitted to the public variables;

switching login items according to the item management information received from the login requester; acquiring an access token _ token verification request for item switching, referring to a base64 encrypted jar packet of three parties for signature verification and setting parameters after base64 encryption as a global state; setting a Transaction Controller to manage a use case set of each interface;

setting a PreProcessor BeanShell PreProcessor, and presetting base64 encryption and signature verification for requesting calling; respectively setting a preprocessor admin for a label operator;

according to a request of a marker for acquiring a marking task list, returning a directory identifier directoryId, a picture identification number imageId and a file name fileName of a picture to be marked to create a marking task, and returning a marking task list taskId to perform a marking workflow; parameterizing a picture identification number imageId through a CSV Data Set Config element;

setting index to count and store the request times of the labeling interface, and using count function as parameter to control the cycle times;

through an HTTP Header Manager, the Header information of the image marking, the mark storage and the submission of the interface checking request is stored, and the format of the interface response data is specified;

after the interface integrity request is edited, introducing a Loop Controller circulation Controller to control and store the labeling interface;

setting a Listener Listener; and using the View Results Tree to check a result Tree and check the log information requested by the interface.

Preferably, the script is developed through the benshell, the jar packet is introduced to realize encryption verification of the interface, the pictures to be marked are encrypted through the jar packet, and a user is allowed to realize importing of batch pictures to be marked.

The present invention will be described more specifically with reference to preferred examples.

The Project Management (PM) information comprises public variables, the public variables (user-defined variables) are set through the setUp thread group, the acquired data and basic data information such as initialization related enterprise names, enterprise ids, request addresses, port numbers, login user name passwords, project ids, annotator name passwords and auditor name passwords are transmitted to the public variables, and some task names, scheme names, data set names and the like are defined by users.

Switching login items according to the item management information received from the login requester; and obtaining an access token access _ token of item switching for later request use, referring to a three-party base64 encrypted jar package for signature verification (bearer shell sample), and setting the parameters of the base64 encrypted as the global.

Adding thread group-characteristics, setting a Transaction Controller for Transaction control to manage the case sets of each interface.

And a BeanShell PreProcessor is arranged for presetting base64 encryption and signature verification, so that the later request calling is facilitated. An admin preprocessor, a annotator A and a annotator B preprocessors are respectively arranged.

And according to the request of the annotator for acquiring the annotation task list, returning the directory identifier directoryId, the picture identification number imageId and the file name fileName of the picture to be annotated to create the annotation task, and returning the annotation task list taskId to perform the annotation workflow.

The picture identification number imageId is parameterized by a CSV Data Set Config element, so that mass labeling of preparation Data is facilitated, for example, labeling of 1-1000 pictures with 10 defects per picture is facilitated.

And an index is set, so that statistics and storage of the number of times of the request of the labeling interface are facilitated, and the count is used as a parameter for cycle number control.

Marking pictures, storing the marks and storing the Header information submitting an audit interface request through an HTTP Header Manager, and specifying the format of interface response data: content-Type = application/json; charset = UTF-8.

And after the interface complete request is edited, introducing a Loop Controller circulating Controller to control and store the labeling interface.

Setting a Listener Listener; and using the View Results Tree to check a result Tree, checking detailed log information of the interface request, and observing an execution result.

In more preferred embodiments, the present invention is a testing tool, comprising:

1) User Defined Variables User defines variable setting for other interfaces to call;

2) Adding an interface request header, and calling ip and port in User Defined Variables;

3) Login-item switch interface setup request;

4) Login-item switch interface assertion;

5) Obtaining access _ token;

6) Setting a bearer shell sampler, calling a three-party jar packet, and setting the global;

7) A characteristic submodule:

8) Adding admin, a annotator A and a BeanShell PreProcessor of an annotator B;

9) A task auditing interface request;

10 Obtaining directoryId, imageId and fileName;

11 Create an annotation task interface request;

12 Annotate task list interface requests;

13 Obtain taskId;

14 Annotator a-annotate flow (transaction controller) -login-item switch-set base64 encryption

15 Annotator a-start an annotate interface request;

16 Loop Controller:

17 Counter-for counting the number of annotations;

18 CSV data file setup-parameterization file for mass labeling tasks;

19 Annotator a-save annotation interface request;

20 Annotator a-submit audit interface requests;

21 Annotator B-annotate flow (transaction controller) -login-item switch-set base64 encryption:

22 Annotator B) -start annotating interface requests;

23 Loop Controller:

24 Counter-for counting the number of annotations;

25 CSV data file setup-parameterization file for mass labeling tasks;

26 Annotator B-save annotation interface request;

27 Annotator B-submit audit interface requests;

28 An align interface request;

29 ) an auditor reject interface request;

30 Auditor-audit pass:

31 Annotator A, B saves the annotation interface request;

32 Annotators A, B submit audit interface requests;

33 Auditor-audit pass interface request;

34 Listener listening.

The invention also provides a batch labeling interface automatic system, and a user can use the batch labeling interface automatic system as a tool to realize the purpose of batch labeling of pictures.

The invention provides a batch annotation interface automation system based on a meter as a development tool, which comprises:

and a module MA: verifying login;

a module MB: after logging in, constructing a tagging task list taskid;

and a module MC: and executing the labeling work and the auditing work according to the labeling task list.

Preferably, the module MA comprises:

module MA1: initializing project management information;

and a module MA2: managing login information according to items in the login request information, and introducing a jar packet to perform base64 signature verification; if the signature verification fails, logging out; if the signature verification is successful, triggering the module MA3 to execute;

module MA3: obtaining a sample picture;

module MA4: storing the sample picture by mysql;

module MA5: screening the sample picture;

module MA6: obtaining a screened picture, and recording the picture as a picture to be marked; in a preferred embodiment, the standard for screening the pictures can be to screen out the pictures which are not in compliance, such as the definition of the pictures is lower than a definition threshold value, and the sample pictures can also be used as the pictures to be labeled; the picture can be screened according to whether the ID of the picture meets the set condition.

Module MA7: creating an annotation task according to the picture to be annotated;

module MA8: and generating an annotation task list according to the annotation task.

Preferably, said module MB comprises:

the module MB1: labeling the picture to be labeled according to the operation of a labeling operator aiming at the labeling task list; wherein, the annotator is the name of the virtual role which is automatically annotated by the computer;

the module MB2: saving the label of the picture to be labeled by a label maker to obtain a labeled picture;

the module MB3: creating an audit task according to the marked pictures;

the module MB4: the method comprises the following steps of (1) aligning an interface request, namely whether the labeling results of auditors in an audit task are aligned or not is auditors, wherein the auditors are names of virtual roles for automatically executing audit by a computer; in a preferred embodiment, the defect list marked by the marker a is compared with the defect list marked by the marker B one by one, and if the comparison result exceeds a set ratio, for example 75%, the following conditions are met, then the marking results of the marker a and the marker B are considered to be aligned:

1) The kind of defect is consistent;

2) The description of the defects is consistent. Therefore, the method can assist the auditor to audit the annotation results of a plurality of annotators,

the module MB5: receiving an auditing result of an auditor; if the result of the verification is passed, the marked picture is considered to be marked successfully; and if the examination is not passed, the annotator modifies the annotation and then carries out examination again until the examination is passed.

Preferably, the first and second electrodes are formed of a metal,

the project management information comprises public variables, the public variables defined by a user are set through a setUp thread group for interface calling, and initialization data are transmitted to the public variables;

switching login items according to the item management information received from the login requester; acquiring an access token _ token verification request for item switching, referring to a base64 encrypted jar packet of three parties for signature verification and setting parameters after base64 encryption as a global state; setting a Transaction Controller to manage a use case set of each interface;

setting a PreProcessor BeanShell PreProcessor, and presetting base64 encryption and signature verification for requesting calling; respectively setting a preprocessor admin for a annotator;

according to a request of a marker for acquiring a marking task list, returning a directory identifier directoryId, a picture identification number imageId and a file name fileName of a picture to be marked to create a marking task, and returning a marking task list taskId to perform a marking workflow; parameterizing a picture identification number imageId through a CSV Data Set Config element;

setting index to count and store the request times of the labeling interface, and using count function as parameter to control the cycle times;

through an HTTP Header Manager, the Header information of the image marking, the mark storage and the submission of the interface checking request is stored, and the format of the interface response data is specified;

after the interface integrity request is edited, introducing a Loop Controller circulation Controller to control and store the labeling interface;

setting a Listener Listener; and using the View Results Tree to check a result Tree and check the log information requested by the interface.

Preferably, the script is developed through the benshell, the jar packet is introduced to realize encryption verification of the interface, the pictures to be marked are encrypted through the jar packet, and a user is allowed to realize importing of batch pictures to be marked.

It is known to those skilled in the art that, in addition to implementing the system, apparatus and its various modules provided by the present invention in pure computer readable program code, the system, apparatus and its various modules provided by the present invention can be implemented in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers and the like by completely programming the method steps. Therefore, the system, the device and the modules thereof provided by the present invention can be considered as a hardware component, and the modules included in the system, the device and the modules thereof for implementing various programs can also be considered as structures in the hardware component; modules for performing various functions may also be considered to be both software programs for performing the methods and structures within hardware components.

The foregoing description of specific embodiments of the present invention has been presented. It is to be understood that the present invention is not limited to the specific embodiments described above, and that various changes or modifications may be made by one skilled in the art within the scope of the appended claims without departing from the spirit of the invention. The embodiments and features of the embodiments of the present application may be combined with each other arbitrarily without conflict.

Claims (2)

1. A batch labeling interface automation method is characterized in that a jmeter is used as a development tool, and comprises the following steps:

step SA: verifying login;

step SB: after logging in, constructing a tagging task list taskid;

step SC: according to the labeling task list, performing labeling work and auditing work;

the step SA includes:

step SA1: initializing project management information;

step SA2: managing login information according to items in the login request information, and introducing a jar packet to perform base64 signature verification; if the signature verification fails, logging out; if the signature verification is successful, triggering the step SB to be executed;

the step SB includes:

step SB1: obtaining a sample picture;

step SB2: storing the sample picture by mysql;

step SB3: screening the sample picture;

step SB4: obtaining a screened picture, and recording the picture as a picture to be marked;

step SB5: creating an annotation task according to the picture to be annotated;

step SB6: generating an annotation task list according to the annotation task;

the step SC comprises the following steps:

step SC1: labeling the picture to be labeled according to the operation of a labeling operator aiming at the labeling task list; wherein, the annotator is the name of the virtual role which is automatically annotated by the computer;

step SC2: saving the label of the picture to be labeled by a label maker to obtain a labeled picture;

step SC3: creating an audit task according to the marked pictures;

step SC4: an alignment interface request; namely, an auditor audits whether the annotation results among the annotators are aligned in an audit task, wherein the auditor is the name of the virtual role which is automatically audited by the computer;

step SC5: receiving an auditing result of an auditor; if the result of the verification is passed, the marked picture is considered to be marked successfully; if the examination is not passed, the annotator modifies the annotation and then carries out examination again until the examination is passed;

the project management information comprises public variables, the public variables defined by a user are set through the setUp thread group for interface calling, and initialization data are transmitted to the public variables;

switching login items according to the item management information received from the login requester; acquiring an access token _ token verification request for item switching, referring to a base64 encrypted jar packet of three parties for signature verification and setting parameters after base64 encryption as a global state; setting a Transaction Controller to manage a use case set of each interface;

setting a PreProcessor BeanShell PreProcessor, and presetting base64 encryption and signature verification for requesting calling; respectively setting a preprocessor admin for a annotator;

according to a request of a marker for acquiring a marking task list, returning a directory identifier directoryId, a picture identification number imageId and a file name fileName of a picture to be marked to create a marking task, and returning a marking task list taskId to perform a marking workflow; parameterizing a picture identification number imageId through a CSV Data Set Config element;

setting index to count and store the request times of the labeling interface, and using count function as parameter to control the cycle times;

through an HTTP Header Manager, the Header information of the image marking, the mark storage and the submission of the interface checking request is stored, and the format of the interface response data is specified;

after the interface integrity request is edited, introducing a Loop Controller circulation Controller to control and store the labeling interface;

setting a Listener monitor; using a View Results Tree to check a fruit Tree and check log information requested by an interface;

the script is developed through the benshell, the jar packet is introduced to realize encryption verification of the interface, the pictures to be marked are encrypted through the jar packet, and a user is allowed to import batch pictures to be marked.

2. A batch annotation interface automation system, based on a meter as a development tool, comprising:

and a module MA: verifying login;

a module MB: after logging in, constructing a tagging task list taskid;

and a module MC: according to the labeling task list, performing labeling work and auditing work;

the module MA comprises:

module MA1: initializing project management information;

and a module MA2: managing login information according to items in the login request information, and introducing a jar packet to perform base64 signature verification; if the signature verification fails, logging out; if the signature verification is successful, triggering the module MB to execute;

the module MB comprises:

the module MB1: obtaining a sample picture;

the module MB2: storing the sample picture by mysql;

the module MB3: screening the sample picture;

the module MB4: obtaining a screened picture, and recording the picture as a picture to be marked;

the module MB5: creating an annotation task according to the picture to be annotated;

the module MB6: generating an annotation task list according to the annotation task;

the module MC comprises:

module MC1: labeling the picture to be labeled according to the operation of a labeling operator aiming at the labeling task list; wherein, the annotator is the name of the virtual role which is automatically annotated by the computer;

and a module MC2: saving the label of the picture to be labeled by a label maker to obtain a labeled picture;

and a module MC3: creating an audit task according to the marked pictures;

and a module MC4: the method comprises the following steps of (1) aligning an interface request, namely whether the labeling results of auditors in an audit task are aligned or not is auditors, wherein the auditors are names of virtual roles for automatically executing audit by a computer;

and a module MC5: receiving an auditing result of an auditor; if the result of the verification is passed, the marked picture is considered to be marked successfully; if the examination is not passed, the annotator modifies the annotation and then carries out examination again until the examination is passed;

the project management information comprises public variables, the public variables defined by a user are set through the setUp thread group for interface calling, and initialization data are transmitted to the public variables;

switching login items according to the item management information received from the login requester; acquiring an access token _ token verification request for item switching, referring to a base64 encrypted jar packet of three parties for signature verification, and setting parameters after base64 encryption as the whole; setting a Transaction Controller to manage a use case set of each interface;

setting a PreProcessor BeanShell PreProcessor, and presetting base64 encryption and signature verification for requesting calling; respectively setting a preprocessor admin for a label operator;

according to a request of a marker for acquiring a marking task list, returning a directory identifier directoryId, a picture identification number imageId and a file name fileName of a picture to be marked to create a marking task, and returning a marking task list taskId to perform a marking workflow; parameterizing a picture identification number imageId through a CSV Data Set Config element;

setting index to count and store the request times of the labeling interface, and using count function as parameter to control the cycle times;

through an HTTP Header Manager, the Header information of the image marking, the mark storage and the submission of the interface checking request is stored, and the format of the interface response data is specified;

after the interface integrity request is edited, introducing a Loop Controller circulation Controller to control and store the labeling interface;

setting a Listener Listener; using a View Results Tree to check a fruit Tree and check log information requested by an interface;

the script is developed through the benshell, the jar packet is introduced to realize encryption verification of the interface, the pictures to be marked are encrypted through the jar packet, and a user is allowed to import batch pictures to be marked.

Images