CN112102263A - Defect detection model generation system, method and device and computer equipment - Google Patents
Defect detection model generation system, method and device and computer equipment Download PDFInfo
- Publication number
- CN112102263A CN112102263A CN202010892996.7A CN202010892996A CN112102263A CN 112102263 A CN112102263 A CN 112102263A CN 202010892996 A CN202010892996 A CN 202010892996A CN 112102263 A CN112102263 A CN 112102263A
- Authority
- CN
- China
- Prior art keywords
- defect
- defect detection
- detection model
- training
- module
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Medical Informatics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
The application relates to a defect detection model generation system, a method, a device, a computer device and a storage medium, wherein the system comprises: the defect data storage module receives a first defect form image; acquiring annotation information corresponding to the first defect form image; obtaining a second defect form image according to the first defect form image and the corresponding labeling information, and sending the second defect form image to a defect detection model training module; the defect detection model training module determines an algorithm mirror image and a candidate training server cluster according to the parameter configuration information; establishing a corresponding defect detection model according to the algorithm mirror image; and scheduling the candidate training server cluster, and training the defect detection model based on the second defect form image to obtain the trained defect detection model. The system establishes a corresponding model according to the algorithm mirror image, and can meet the model requirements for detecting different defect types; by selecting the candidate training server cluster, manual participation is not needed, and the generation efficiency of the defect detection model is improved.
Description
Technical Field
The present application relates to the field of artificial intelligence technologies, and in particular, to a system, a method, an apparatus, a computer device, and a storage medium for generating a defect detection model.
Background
In recent years, 3C products emerge endlessly, and the products need to be supported by a great variety of parts; the parts have the characteristics of small volume and irregular appearance, and the fine defects of the parts are difficult to accurately detect manually in a complex production environment.
With the popularization of artificial intelligence technology, deep learning technology is also introduced into the 3C industry; through the defect detection model, the detection work of quality defects of 3C products can be replaced by manpower. The traditional machine learning method cannot be well adapted to quick product change, a large amount of time cost is consumed for manually training and debugging the model, the type of the model which can be trained is single, the overall timeliness is poor, and the diversified requirements of customers cannot be met in time. Therefore, the existing defect detection model is low in generation efficiency.
Disclosure of Invention
In view of the above, it is necessary to provide a defect detection model generation system, method, apparatus, computer device and storage medium capable of improving the generation efficiency of the defect detection model.
A defect detection model generation system, the system comprising: the defect detection system comprises a defect data storage module and a defect detection model training module; the defect data storage module is in communication connection with the defect detection training module;
the defect data storage module is used for receiving a first defect form image uploaded by the terminal; acquiring annotation information corresponding to the first defect form image; obtaining a second defect form image carrying the marking information according to the first defect form image and the corresponding marking information; sending the second defect form image to the defect detection model training module;
the defect detection model training module is used for acquiring parameter configuration information carried in a defect detection model training request sent by a terminal; determining an algorithm mirror image and a candidate training server cluster according to the parameter configuration information; establishing a corresponding defect detection model according to the algorithm mirror image; and scheduling the candidate training server cluster, and training the defect detection model based on the second defect form image to obtain a trained defect detection model.
In one embodiment, the defect detection model generation system further comprises an account change module; the account changing module comprises an account creating module, an account identifying module and an authority adjusting module; the account creation module is used for responding to an account creation request sent by the terminal and generating corresponding account information; the account information carries an account identifier; the account identification module is used for identifying account information corresponding to the account identification; the authority adjusting module is used for adjusting the authority parameters corresponding to the account information.
In one embodiment, the defect data storage module further comprises a defect data acquisition module; the defect data acquisition module is used for responding to a defect form image uploading request sent by the terminal; acquiring the account information carried in the defect form image uploading request, and determining an authority parameter corresponding to the account information; if the authority parameter is smaller than a preset authority threshold value, refusing to execute the defect form image uploading request; and if the permission parameter is greater than or equal to the preset permission threshold, acquiring the first defect form image and the marking information corresponding to the first defect form image according to the defect form image uploading request.
In one embodiment, the defect detection model training module further comprises a cluster scheduling module; the cluster scheduling module is used for acquiring the running states of a plurality of training server clusters; determining a load state parameter of the training server cluster according to the running state of the training server cluster; and taking the training server cluster with the load state parameter smaller than a preset load threshold value as the candidate training server cluster.
In one embodiment, the defect detection model training module is further configured to perform a defect identification test on the trained defect detection model to obtain a defect identification rate; and repeatedly training the defect detection model according to the defect recognition rate until the defect recognition rate obtained according to the defect detection model reaches a preset recognition rate threshold value, and taking the defect detection model as the trained defect detection model.
In one embodiment, the defect detection model training module is further configured to send the trained defect detection model to a terminal corresponding to the defect detection model training request; and the terminal is used for calling the trained defect detection model, detecting whether the image to be identified contains defect information or not, and obtaining a defect detection result.
A method of generating a defect detection model, the method comprising:
acquiring parameter configuration information carried in a defect detection model training request sent by a terminal, and determining an algorithm mirror image and a candidate training server cluster according to the parameter configuration information;
establishing a corresponding defect detection model according to the algorithm mirror image;
and scheduling the candidate training server cluster, and training the defect detection model based on a defect detection training image and the labeling information carried by the defect detection training image to obtain the trained defect detection model.
A defect detection model generation apparatus, the apparatus comprising:
the parameter acquisition module is used for acquiring parameter configuration information carried in a defect detection model training request sent by a terminal and determining an algorithm mirror image and a candidate training server cluster according to the parameter configuration information;
the model establishing module is used for establishing a corresponding defect detection model according to the algorithm mirror image;
and the model training module is used for scheduling the candidate training server cluster, training the defect detection model based on the defect detection training image and the labeling information carried by the defect detection training image, and obtaining the trained defect detection model.
A computer device comprising a memory and a processor, the memory storing a computer program, the processor implementing the following steps when executing the computer program:
acquiring parameter configuration information carried in a defect detection model training request sent by a terminal, and determining an algorithm mirror image and a candidate training server cluster according to the parameter configuration information;
establishing a corresponding defect detection model according to the algorithm mirror image;
and scheduling the candidate training server cluster, and training the defect detection model based on a defect detection training image and the labeling information carried by the defect detection training image to obtain the trained defect detection model.
A computer-readable storage medium, on which a computer program is stored which, when executed by a processor, carries out the steps of:
acquiring parameter configuration information carried in a defect detection model training request sent by a terminal, and determining an algorithm mirror image and a candidate training server cluster according to the parameter configuration information;
establishing a corresponding defect detection model according to the algorithm mirror image;
and scheduling the candidate training server cluster, and training the defect detection model based on a defect detection training image and the labeling information carried by the defect detection training image to obtain the trained defect detection model.
The system, the method, the device, the computer equipment and the storage medium for generating the defect detection model comprise: the defect detection system comprises a defect data storage module and a defect detection model training module; the defect data storage module is in communication connection with the defect detection training module; the defect data storage module is used for receiving a first defect form image uploaded by the terminal; acquiring annotation information corresponding to the first defect form image; obtaining a second defect form image carrying annotation information according to the first defect form image and the corresponding annotation information; sending the second defect form image to a defect detection model training module; the defect detection model training module is used for acquiring parameter configuration information carried in a defect detection model training request sent by the terminal; determining an algorithm mirror image and a candidate training server cluster according to the parameter configuration information; establishing a corresponding defect detection model according to the algorithm mirror image; and scheduling the candidate training server cluster, and training the defect detection model based on the second defect form image to obtain the trained defect detection model. The system determines a proper algorithm mirror image according to parameter configuration information carried in a defect detection model training request, and establishes a corresponding model according to the algorithm mirror image, so that the model requirements for detecting different defect types can be met; by selecting the candidate training server cluster, the training of a plurality of defect detection models with different types is realized without manual participation, and the generation efficiency of the defect detection models is improved.
Drawings
FIG. 1 is a diagram of an exemplary application of a system for generating a defect inspection model;
FIG. 2 is a schematic diagram of a page of an item detail page in one embodiment;
FIG. 3 is a schematic diagram of a page for account creation in one embodiment;
FIG. 4 is a system diagram of a system for generating a defect inspection model in accordance with one embodiment;
FIG. 5 is a schematic flow chart diagram illustrating a method for generating a defect detection model in one embodiment;
FIG. 6 is a block diagram showing the structure of a defect detection model generating apparatus according to an embodiment;
FIG. 7 is a diagram illustrating an internal structure of a computer device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
The defect detection model generation system provided by the application can be applied to the application environment shown in fig. 1. The terminal 11 communicates with the defect detection model generation system 12 via a network. The defect detection model generation system 12 includes: a defect data storage module 121 and a defect detection model training module 122; the defect data storage module 121 is in communication connection with the defect detection training module 122; the defect data storage module 121 is configured to receive the first defect form image uploaded by the terminal 11; the defect data storage module 121 acquires annotation information corresponding to the first defect form image; the defect data storage module 121 obtains a second defect form image carrying the annotation information according to the first defect form image and the corresponding annotation information; the defect data storage module 121 sends the second defect form image to the defect detection model training module 122; the defect detection model training module 122 is configured to obtain parameter configuration information carried in a defect detection model training request sent by the terminal 11; the defect detection model training module 122 determines an algorithm image and a candidate training server cluster 1222 from the algorithm image library 1221 according to the parameter configuration information; the defect detection model training module 122 establishes a corresponding defect detection model according to the algorithm mirror image; the defect detection model training module 122 schedules the candidate training server cluster 1222 to train the defect detection model based on the second defect morphology image, resulting in a trained defect detection model. The terminal 11 may be, but not limited to, various personal computers, notebook computers, smart phones, tablet computers, and portable wearable devices, and the defect detection model generation system 12 and each internal module may be implemented by an independent server or a server cluster formed by a plurality of servers. The defect detection model generation system 12 may support a plurality of deployment modes, such as standalone deployment, cluster deployment, online deployment, and the like, to maximize computational efficiency.
The defect detection model generation system can provide a web-like management and control interface for a user, the user can log in the defect detection model generation system through a terminal to perform corresponding operation, and each module in the defect detection model generation system can provide corresponding data support according to the operation of the user, so that the training and generation processes of the defect detection model are realized.
In one embodiment, the defect data storage module is used for receiving a first defect form image uploaded by the terminal; acquiring annotation information corresponding to the first defect form image; obtaining a second defect form image carrying annotation information according to the first defect form image and the corresponding annotation information; and sending the second defect form image to a defect detection model training module.
The defect data storage module is equivalent to a data foreground and can be interactively butted with a terminal, so that a user can add, delete, modify, replace and the like data; other modules may also obtain data needed for model training from the defect data storage module. The first defect type image is image information including defect type characteristics, for example, the 3C product surface image is the first defect type image, and the scratch appearing in the product surface image is the defect type characteristic. The annotation information is annotation information corresponding to the first defect form image and the defect form feature, and for example, the image is annotated as "3C component a" according to the content of the first defect form image, and the scratch region in the image is annotated as "scratch type a". The second defect form image is an image which is formed by combining the first defect form image and the corresponding marking information by the defect data storage module and contains the marking information.
Specifically, a user logs in a defect detection model generation system through a terminal and uploads image information stored in the terminal; when the image is uploaded or after the image is uploaded, a user can mark the uploaded image through the defect data storage module; the defect data storage module takes image information uploaded by a user through a terminal as a first defect form image, takes the marked content as marking information corresponding to the first defect form image, correspondingly combines the first defect form image and the marking information to obtain a second defect form image carrying the marking information, and sends the second defect form image as training data of a defect detection model to the defect detection model training module. It should be noted that the first defect form image uploaded by the terminal and the obtaining of the annotation information corresponding to the first defect form image may be performed discontinuously; for example, a first defect form image uploaded by a user is obtained first, original accumulation of image data is completed, and when the user needs to perform a defect detection model, the corresponding first defect form image is displayed so that the user performs annotation operation, so that annotation information corresponding to the first defect form image is obtained. Certainly, the user can also directly upload pre-labeled image data, and the defect data storage module can identify whether the uploaded data carries corresponding labeling information or whether the labeling information is complete, so as to make corresponding identification.
In the embodiment, the second defect form image carrying the annotation information is obtained by acquiring the first defect form image and the corresponding annotation information, so that the data set accumulation required by the defect detection model generation is completed, the model requirements for detecting different defect types can be met, and the defect detection model generation efficiency is improved.
In one embodiment, the defect detection model training module is configured to obtain parameter configuration information carried in a defect detection model training request sent by a terminal; determining an algorithm mirror image and a candidate training server cluster according to the parameter configuration information; establishing a corresponding defect detection model according to the algorithm mirror image; and scheduling the candidate training server cluster, and training the defect detection model based on the second defect form image to obtain the trained defect detection model.
The defect detection model training module is an execution module for specifically training and generating a defect detection model. The parameter configuration information is specific setting of a defect detection model training task and can comprise data enhancement configuration, training parameter configuration and the like; the data enhancement configuration may include a number of options for pre-processing the training data, such as image blurring, image cropping, image noise processing, image sharpening, image color adjustment, image angle adjustment, etc.; the training parameter configuration can comprise aspects of model building size selection, learning times, hardware configuration, algorithm and the like, for example, a large model is selected to be built for training, the large model has the characteristics of high precision, large memory, low learning speed and poor real-time performance, and correspondingly, the small model has the characteristics of high speed, small memory, high reasoning speed, high real-time performance and low precision; the hardware configuration can select hardware parameters such as the number supported by the GPU card.
Specifically, the defect detection model training module acquires a defect detection model training request sent by a terminal, acquires parameter configuration information from the defect detection model training request, and determines a suitable algorithm mirror image and a candidate training server cluster according to the parameter configuration information. And then establishing a corresponding defect detection model according to the algorithm mirror image, and scheduling the candidate training server cluster to perform a training process on the defect detection model based on the second defect form image until the training of the defect detection model is finished.
In the embodiment, a proper algorithm mirror image is determined through parameter configuration information carried in a defect detection model training request, and a corresponding model is established according to the algorithm mirror image, so that the model requirements for detecting different defect types can be met; by selecting the candidate training server cluster, the training of a plurality of defect detection models with different types is realized without manual participation, and the generation efficiency of the defect detection models is improved.
The defect detection model generation system includes: the defect detection system comprises a defect data storage module and a defect detection model training module; the defect data storage module is in communication connection with the defect detection training module; the defect data storage module is used for receiving a first defect form image uploaded by the terminal; acquiring annotation information corresponding to the first defect form image; obtaining a second defect form image carrying annotation information according to the first defect form image and the corresponding annotation information; sending the second defect form image to a defect detection model training module; the defect detection model training module is used for acquiring parameter configuration information carried in a defect detection model training request sent by the terminal; determining an algorithm mirror image and a candidate training server cluster according to the parameter configuration information; establishing a corresponding defect detection model according to the algorithm mirror image; and scheduling the candidate training server cluster, and training the defect detection model based on the second defect form image to obtain the trained defect detection model. The system determines a proper algorithm mirror image according to parameter configuration information carried in a defect detection model training request, and establishes a corresponding model according to the algorithm mirror image, so that the model requirements for detecting different defect types can be met; by selecting the candidate training server cluster, the training of a plurality of defect detection models with different types is realized without manual participation, and the generation efficiency of the defect detection models is improved.
In one embodiment, the defect data storage module comprises one or more of a defect data acquisition module, a defect data labeling module, a visualization module and a project module; the modules cooperate with each other to realize the steps or the method executed by the defect data storage module.
The defect data acquisition module can acquire defect data sent by the terminal according to a defect form image uploading request sent by the terminal, wherein the defect form image uploading request comprises information such as a defect data list, the size of the defect data, the type of the defect data, whether to label and the like; the defect data acquisition module establishes communication connection with the terminal after verifying a defect form image uploading request sent by the terminal, and acquires defect data corresponding to the defect form image uploading request; the defect form image uploading request can also directly carry corresponding defect data, and the defect data acquisition module verifies the defect form image uploading request sent by the terminal and then directly acquires the corresponding defect data from the defect form image uploading request.
The defect data labeling module can enable the terminal to label the defect data, and the labeling information includes but is not limited to a defect type label, a custom defect label, and the like. The defect data labeling module can sequentially display each picture in a preset interface of the defect detection model generation system to a terminal user, the user respectively checks each picture, and corresponding defect labeling information is added to each first defect form image through the terminal. The defect data labeling module can also call a pre-trained general defect detection model, identify partial defect features in the first defect form image and add corresponding labeling information, and then a user confirms the defect data and the corresponding labeling information through the terminal to finish labeling of the first defect form image. In addition, the defect data labeling module can also identify an area containing defects and add a frame to the area, so that a user can directly input labeling information in the frame range.
The project module is used for checking and managing the whole process of generating the defect detection model, and the functions of the project module comprise project creation, project modification, project deletion and the like. The newly-built project requires a user to log in a defect detection model generation system through a terminal to enter newly-built project parameters to preset basic attributes of the project, the parameters include but are not limited to project name, project operation authority, task type and other information, a project module can generate a corresponding project according to the newly-built project parameters, allocate an identifier (project ID) to the project, and display the project in a project detail page provided by the defect detection model generation system, as shown in a project detail page schematic diagram of fig. 2, the display content includes information such as the project ID, the project name, a project creator (i.e., account information corresponding to the project), project state, defect detection model number, task type and the like. Subsequent steps of defect data acquisition, defect detection model training and the like can be carried out by taking the item as a unit.
The visualization module can generate a visualized chart according to the type, the size, the labeling condition and the like of the stored defect data. For example, a number change line graph is generated monthly according to the number of the second defect form image, and a pie type graph containing different colors is generated according to the specific type of the defect form. Further, the data visualization module can also summarize the defect data to form a visualization report of the defect data.
In one embodiment, the defect detection model training module includes one or more of a resource monitoring module, a cluster scheduling module, an algorithm mirroring module, a training module, and a data management module.
The resource monitoring module is used for acquiring resource utilization states, cluster states and the like of the training servers and the training server clusters, generating corresponding running states and mastering the health degree of the training servers and the training server clusters. And the cluster scheduling module is used for calling the training server cluster according to the running state of the training server cluster and executing the training task of the corresponding defect detection model. The algorithm mirror image module can manage and control the version of the algorithm mirror image stored in the algorithm mirror image library, and comprises the steps of importing, exporting, updating iteration, deleting, modifying and the like of the algorithm mirror image; the algorithm mirror image module determines an applicable algorithm mirror image according to the training requirements of the defect detection model and establishes a corresponding model; the algorithm mirror comprises a general algorithm and a plurality of industry customized algorithms, and different selections can be carried out according to specific requirements. And the training module trains the defect detection model according to the second defect form image until the defect detection model meets the use requirement. The resource monitoring module can also acquire progress information of the training module for training the defect detection model, and the progress information is displayed as a project state on a project detail page in forms of a progress bar and the like, so that a user can know the progress of the training of the defect detection model.
In one embodiment, the defect detection model generation system further comprises an account change module; the account changing module comprises an account creating module, an account identifying module and a permission adjusting module; the account creation module is used for responding to an account creation request sent by the terminal and generating corresponding account information; the account information carries an account identifier; the account identification module is used for identifying account information corresponding to the account identifier; and the authority adjusting module is used for adjusting the authority parameters corresponding to the account information.
The account changing module is a module for configuring an account for a system user. The account creating module can create an account for a new user and also can delete and modify account information; the classification and the hierarchical management of the user accounts can be realized in a user group mode, and the batch management of the accounts can also be realized by configuring the user group through the authority parameters.
Specifically, the account creation module verifies an account creation request sent by the terminal, and generates corresponding account information according to the account creation request after the verification is passed; as shown in the account creation diagram of fig. 3, the account information may include a user name, a user group, a password, a permission, a creation time, and the like; meanwhile, an identifier can be allocated to the account information to serve as an account identifier; and each module in the defect detection model generation system can identify and read corresponding account information according to the account identification through the account identification module. The account identification module is also capable of determining a terminal corresponding to the account information, or account information corresponding to the terminal identification. The permission adjusting module is used for adjusting permission parameters corresponding to the account information, and the permission parameter adjusting object can be the permission parameters corresponding to the account information of a single user or the permission parameters corresponding to the user group. Different permission parameters are different corresponding to the operable modules, and after a user logs in the defect detection model through account information to generate, a displayed interface can be correspondingly changed according to the difference of the operable modules.
According to the embodiment, relatively comprehensive account management is realized through the account creating module, the account identifying module and the authority adjusting module, the efficiency of a user for operating the defect detection model generating system is improved, and meanwhile, the safety and the stability of the defect detection model generating system are also improved through the authority supervision of the user.
In one embodiment, the defect data acquisition module is used for responding to a defect form image uploading request sent by a terminal; acquiring account information carried in a defect form image uploading request, and determining authority parameters corresponding to the account information; if the authority parameter is smaller than the preset authority threshold value, refusing to execute the defect form image uploading request; and if the authority parameter is greater than or equal to the preset authority threshold, acquiring a first defect form image and annotation information corresponding to the first defect form image according to the defect form image uploading request.
Specifically, multiple modules in the defect detection model generation system may perform restriction on a preset permission threshold on permission parameters corresponding to account information, so as to maintain transmission security of data of each module. Taking a defect data acquisition module as an example, when a defect form image uploading request sent by a terminal is received, acquiring account information carried in the request, and determining corresponding authority parameters; for example, the preset permission threshold of the defect data acquisition module is 5, the permission parameter corresponding to the account information is 0, the permission parameter is smaller than the preset permission threshold, and the account information does not have the qualification for uploading the defect form image, so that the defect form image uploading request sent by the terminal can be rejected. If the authority parameter corresponding to the account information of the other terminal is 10, and the authority parameter is greater than the preset authority threshold, the account information has the qualification of uploading the defect form image, so that the first defect form image sent by the terminal can be acquired.
According to the embodiment, the limitation on the modules which can be operated by account information of different levels is made through the authority parameters corresponding to the account information and the preset authority threshold, and the safety of the defect detection model generation system is improved. The defect data acquisition module also avoids random uploading of defect form images by setting a preset authority threshold value, and improves the availability of the whole data of the system.
In one embodiment, the cluster scheduling module is configured to obtain operating states of a plurality of training server clusters; determining a load state parameter of the training server cluster according to the running state of the training server cluster; and taking the training server cluster with the load state parameter smaller than the preset load threshold value as a candidate training server cluster.
Specifically, the candidate training server cluster is better in overall operation performance due to the fact that the load state parameter is smaller than the preset load threshold, and is preferentially selected during training. The cluster scheduling module realizes load balance among a plurality of training server clusters by determining the load state parameters of the training server clusters, and improves the overall operation efficiency of the defect detection model generation system.
In one embodiment, the training module is further configured to perform a defect identification test on the trained defect detection model to obtain a defect identification rate; and repeatedly training the defect detection model according to the defect recognition rate until the defect recognition rate obtained according to the defect detection model reaches a preset recognition rate threshold value, and taking the defect detection model as the trained defect detection model.
Specifically, the training module can determine the training progress of the defect detection model according to the obtained defect recognition rate by performing defect recognition test on the trained defect detection model; and when the defect recognition rate does not reach the preset recognition rate threshold value, continuing training, and if the defect recognition rate does not reach the preset recognition rate threshold value after a plurality of times of long-time defect recognition tests, packaging information such as parameters generated in the training process, generating early warning information and sending the early warning information to a person corresponding to the terminal for reason investigation. When the defect recognition rate reaches the preset recognition rate threshold after multiple tests, the training of the defect detection model can be considered to be completed.
In the embodiment, the defect identification test is carried out on the trained defect detection model, the training progress of the defect detection model is mastered through the defect identification rate, and the detection accuracy of the trained defect detection model is ensured.
In one embodiment, the training module is further configured to send the trained defect detection model to a terminal corresponding to the defect detection model training request; and the terminal is used for calling the trained defect detection model, detecting whether the image to be identified contains defect information or not, and obtaining a defect detection result.
Specifically, after the defect detection model training is completed, the defect detection model training module can send the trained defect detection model to the corresponding terminal. And the terminal can finish the work of defect detection through the trained defect detection model. The defect form image uploading request can include information of a demand side of the defect detection model, and the defect detection model training module can determine a target terminal of the defect detection model according to the information of the demand side. And then, or the defect detection model training module stores the trained defect detection model to the defect detection model generating system, and the demand terminal can log in the defect detection model generating system to obtain the defect detection model. According to the embodiment, the efficiency of acquiring the defect detection model by the terminal is improved by sending the trained defect detection model to the corresponding terminal.
In an embodiment, in order to more clearly illustrate the technical solution provided by the embodiment of the present application, an architecture of the system will be described below with reference to fig. 4, which includes the following specific contents:
the defect detection model generation system can be divided into a foreground part and a background part, and the foreground part and the background part can be communicated through http/tcp; the foreground part comprises a defect data acquisition module, a defect data labeling module and a data visualization module, and a user can upload a defect form image required by training a defect detection model to a defect detection model generation system through the defect data acquisition module and label the defect form image through the defect data labeling module; and then, a visualization report containing a chart can be exported by using a data visualization module, and the defect form image and the labeling condition are determined. The background part mainly comprises an account changing module and a defect detection model training module. The account management module comprises an account creating module, a permission adjusting module and an account identification module. The system can modify the account information of the user logging in the system through the account changing module, such as account creation, account deletion and the like, the authority parameter is set through the authority adjusting module to limit the operable module range of the user, the overall security of the system is ensured, the account identifying module is used for confirming the identity of the user, and meanwhile, the account information of the user can be provided for other modules.
The defect detection model training module comprises a resource monitoring module, a training module, an algorithm mirror module, a data management module and a cluster scheduling module. The system determines the running state of each training server cluster through a resource monitoring module to obtain a cluster state and sends the cluster state to a training module, the training module determines a training server cluster to be called according to the cluster state, and the cluster scheduling module carries out scheduling operation; after a training server cluster capable of performing model training is scheduled, a defect detection model to be trained is established according to a corresponding mirror image is pulled from the algorithm mirror image module, and then a defect form image of a foreground part is acquired from the data management module for training to obtain the trained defect detection model.
As shown in fig. 5, a defect inspection model generation method is provided, which is described by taking the defect inspection model generation system in fig. 1 as an example, and includes the following steps:
and step 53, scheduling the candidate training server cluster, and training the defect detection model based on the defect detection training image and the labeling information carried by the defect detection training image to obtain the trained defect detection model.
In the defect detection model generation method, the server determines a corresponding algorithm mirror image and a candidate training server cluster according to parameter configuration information carried in a defect detection model training request; and establishing a defect detection model to be trained according to the algorithm mirror image, acquiring a defect detection training image carrying the labeling information, and training the defect detection model to obtain a trained defect detection model. The method can quickly meet the generation requirements of various different defect detection models; and dispatching the candidate training server cluster, training a defect detection model to be trained based on the defect detection training image carrying the labeling information, realizing the training of a plurality of types of defect detection models without manual participation, and improving the generation efficiency of the defect detection model.
It should be understood that, although the steps in the flowchart of fig. 5 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least a portion of the steps in fig. 5 may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed in turn or alternately with other steps or at least a portion of the other steps or stages.
In one embodiment, as shown in fig. 6, there is provided a defect detection model generation apparatus, including: a parameter acquisition module 61, a model building module 62 and a model training module 63, wherein:
the parameter obtaining module 61 is configured to obtain parameter configuration information carried in a defect detection model training request sent by a terminal, and determine an algorithm mirror image and a candidate training server cluster according to the parameter configuration information;
the model establishing module 62 is used for establishing a corresponding defect detection model according to the algorithm mirror image;
and the model training module 63 is configured to schedule the candidate training server cluster, train the defect detection model based on the defect detection training image and the label information carried by the defect detection training image, and obtain the trained defect detection model.
For the specific definition of the defect detection model generation apparatus, reference may be made to the above definition of the defect detection model generation method, which is not described herein again. The modules in the defect detection model generation device can be wholly or partially realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, which may be a server, the internal structure of which may be as shown in fig. 7. The computer device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used for storing defect detection model generation data. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a defect detection model generation method.
Those skilled in the art will appreciate that the architecture shown in fig. 7 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, a computer device is provided, comprising a memory and a processor, the memory having a computer program stored therein, the processor implementing the following steps when executing the computer program:
acquiring parameter configuration information carried in a defect detection model training request sent by a terminal, and determining an algorithm mirror image and a candidate training server cluster according to the parameter configuration information;
establishing a corresponding defect detection model according to the algorithm mirror image;
and scheduling the candidate training server cluster, and training the defect detection model based on the defect detection training image and the labeling information carried by the defect detection training image to obtain the trained defect detection model.
In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of:
acquiring parameter configuration information carried in a defect detection model training request sent by a terminal, and determining an algorithm mirror image and a candidate training server cluster according to the parameter configuration information;
establishing a corresponding defect detection model according to the algorithm mirror image;
and scheduling the candidate training server cluster, and training the defect detection model based on the defect detection training image and the labeling information carried by the defect detection training image to obtain the trained defect detection model.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above examples only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A system for generating a defect inspection model, the system comprising: the defect detection system comprises a defect data storage module and a defect detection model training module; the defect data storage module is in communication connection with the defect detection training module;
the defect data storage module is used for receiving a first defect form image uploaded by the terminal; acquiring annotation information corresponding to the first defect form image; obtaining a second defect form image carrying the marking information according to the first defect form image and the corresponding marking information; sending the second defect form image to the defect detection model training module;
the defect detection model training module is used for acquiring parameter configuration information carried in a defect detection model training request sent by a terminal; determining an algorithm mirror image and a candidate training server cluster according to the parameter configuration information; establishing a corresponding defect detection model according to the algorithm mirror image; and scheduling the candidate training server cluster, and training the defect detection model based on the second defect form image to obtain a trained defect detection model.
2. The system of claim 1, wherein the defect detection model generation system further comprises an account change module; the account changing module comprises an account creating module, an account identifying module and an authority adjusting module;
the account creation module is used for responding to an account creation request sent by the terminal and generating corresponding account information; the account information carries an account identifier;
the account identification module is used for identifying account information corresponding to the account identification;
the authority adjusting module is used for adjusting the authority parameters corresponding to the account information.
3. The system of claim 2, wherein the defect data storage module further comprises a defect data acquisition module;
the defect data acquisition module is used for responding to a defect form image uploading request sent by the terminal; acquiring the account information carried in the defect form image uploading request, and determining an authority parameter corresponding to the account information; if the authority parameter is smaller than a preset authority threshold value, refusing to execute the defect form image uploading request; and if the permission parameter is greater than or equal to the preset permission threshold, acquiring the first defect form image and the marking information corresponding to the first defect form image according to the defect form image uploading request.
4. The system of claim 1, wherein the defect detection model training module further comprises a cluster scheduling module;
the cluster scheduling module is used for acquiring the running states of a plurality of training server clusters; determining a load state parameter of the training server cluster according to the running state of the training server cluster; and taking the training server cluster with the load state parameter smaller than a preset load threshold value as the candidate training server cluster.
5. The system of claim 1, wherein the defect detection model training module is further configured to perform a defect recognition test on the trained defect detection model to obtain a defect recognition rate; and repeatedly training the defect detection model according to the defect recognition rate until the defect recognition rate obtained according to the defect detection model reaches a preset recognition rate threshold value, and taking the defect detection model as the trained defect detection model.
6. The system according to any one of claims 1 to 5, wherein the defect detection model training module is further configured to send the trained defect detection model to a terminal corresponding to the defect detection model training request; and the terminal is used for calling the trained defect detection model, detecting whether the image to be identified contains defect information or not, and obtaining a defect detection result.
7. A method for generating a defect detection model, the method comprising:
acquiring parameter configuration information carried in a defect detection model training request sent by a terminal, and determining an algorithm mirror image and a candidate training server cluster according to the parameter configuration information;
establishing a corresponding defect detection model according to the algorithm mirror image;
and scheduling the candidate training server cluster, and training the defect detection model based on a defect detection training image and the labeling information carried by the defect detection training image to obtain the trained defect detection model.
8. A defect inspection model generation apparatus, the apparatus comprising:
the parameter acquisition module is used for acquiring parameter configuration information carried in a defect detection model training request sent by a terminal and determining an algorithm mirror image and a candidate training server cluster according to the parameter configuration information;
the model establishing module is used for establishing a corresponding defect detection model according to the algorithm mirror image;
and the model training module is used for scheduling the candidate training server cluster, training the defect detection model based on the defect detection training image and the labeling information carried by the defect detection training image, and obtaining the trained defect detection model.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor realizes the steps of the method of claim 7 when executing the computer program.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method as claimed in claim 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010892996.7A CN112102263A (en) | 2020-08-31 | 2020-08-31 | Defect detection model generation system, method and device and computer equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010892996.7A CN112102263A (en) | 2020-08-31 | 2020-08-31 | Defect detection model generation system, method and device and computer equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112102263A true CN112102263A (en) | 2020-12-18 |
Family
ID=73758302
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010892996.7A Pending CN112102263A (en) | 2020-08-31 | 2020-08-31 | Defect detection model generation system, method and device and computer equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112102263A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112711603A (en) * | 2020-12-30 | 2021-04-27 | 广东粤云工业互联网创新科技有限公司 | Workpiece detection method and system based on cloud and computer-readable storage medium |
| CN112712505A (en) * | 2020-12-30 | 2021-04-27 | 广东粤云工业互联网创新科技有限公司 | Workpiece detection method and system based on cloud and computer-readable storage medium |
| CN113298760A (en) * | 2021-05-07 | 2021-08-24 | 厦门宏泰智能制造有限公司 | Quality detection method, device and system based on deep learning |
| CN113487548A (en) * | 2021-06-29 | 2021-10-08 | 嵊州市浙江工业大学创新研究院 | Product defect detection method based on YOLO-v3 |
| CN113743447A (en) * | 2021-07-15 | 2021-12-03 | 上海朋熙半导体有限公司 | Semiconductor defect identification method and device, computer equipment and storage medium |
| CN114782445A (en) * | 2022-06-22 | 2022-07-22 | 深圳思谋信息科技有限公司 | Object defect detection method and device, computer equipment and storage medium |
| CN117763355A (en) * | 2023-12-21 | 2024-03-26 | 正泰集团研发中心(上海)有限公司 | Iterative training system and method for industrial quality inspection model and industrial quality inspection system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108734293A (en) * | 2017-04-13 | 2018-11-02 | 北京京东尚科信息技术有限公司 | Task management system, method and apparatus |
| CN109840591A (en) * | 2017-11-29 | 2019-06-04 | 华为技术有限公司 | Model training systems, method and storage medium |
| CN110554047A (en) * | 2019-09-06 | 2019-12-10 | 腾讯科技（深圳）有限公司 | method, device, system and equipment for processing product defect detection data |
| CN111190718A (en) * | 2020-01-07 | 2020-05-22 | 第四范式（北京）技术有限公司 | Method, device and system for realizing task scheduling |
| CN111242317A (en) * | 2020-01-09 | 2020-06-05 | 深圳供电局有限公司 | Method and device for managing application, computer equipment and storage medium |
| CN111435449A (en) * | 2018-12-26 | 2020-07-21 | 深圳市优必选科技有限公司 | Model self-training method and device, computer equipment and storage medium |
-
2020
- 2020-08-31 CN CN202010892996.7A patent/CN112102263A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108734293A (en) * | 2017-04-13 | 2018-11-02 | 北京京东尚科信息技术有限公司 | Task management system, method and apparatus |
| CN109840591A (en) * | 2017-11-29 | 2019-06-04 | 华为技术有限公司 | Model training systems, method and storage medium |
| CN111435449A (en) * | 2018-12-26 | 2020-07-21 | 深圳市优必选科技有限公司 | Model self-training method and device, computer equipment and storage medium |
| CN110554047A (en) * | 2019-09-06 | 2019-12-10 | 腾讯科技（深圳）有限公司 | method, device, system and equipment for processing product defect detection data |
| CN111190718A (en) * | 2020-01-07 | 2020-05-22 | 第四范式（北京）技术有限公司 | Method, device and system for realizing task scheduling |
| CN111242317A (en) * | 2020-01-09 | 2020-06-05 | 深圳供电局有限公司 | Method and device for managing application, computer equipment and storage medium |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112711603A (en) * | 2020-12-30 | 2021-04-27 | 广东粤云工业互联网创新科技有限公司 | Workpiece detection method and system based on cloud and computer-readable storage medium |
| CN112712505A (en) * | 2020-12-30 | 2021-04-27 | 广东粤云工业互联网创新科技有限公司 | Workpiece detection method and system based on cloud and computer-readable storage medium |
| CN113298760A (en) * | 2021-05-07 | 2021-08-24 | 厦门宏泰智能制造有限公司 | Quality detection method, device and system based on deep learning |
| CN113487548A (en) * | 2021-06-29 | 2021-10-08 | 嵊州市浙江工业大学创新研究院 | Product defect detection method based on YOLO-v3 |
| CN113743447A (en) * | 2021-07-15 | 2021-12-03 | 上海朋熙半导体有限公司 | Semiconductor defect identification method and device, computer equipment and storage medium |
| CN113743447B (en) * | 2021-07-15 | 2024-05-17 | 上海朋熙半导体有限公司 | Semiconductor flaw identification method, device, computer equipment and storage medium |
| CN114782445A (en) * | 2022-06-22 | 2022-07-22 | 深圳思谋信息科技有限公司 | Object defect detection method and device, computer equipment and storage medium |
| CN114782445B (en) * | 2022-06-22 | 2022-10-11 | 深圳思谋信息科技有限公司 | Object defect detection method and device, computer equipment and storage medium |
| CN117763355A (en) * | 2023-12-21 | 2024-03-26 | 正泰集团研发中心(上海)有限公司 | Iterative training system and method for industrial quality inspection model and industrial quality inspection system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112102263A (en) | Defect detection model generation system, method and device and computer equipment | |
| CN112085078A (en) | Image classification model generation system, method and device and computer equipment | |
| CN109558864A (en) | Face critical point detection method, apparatus and storage medium | |
| CN111476871B (en) | Method and device for generating video | |
| US10614347B2 (en) | Identifying parameter image adjustments using image variation and sequential processing | |
| CN111340237A (en) | Data processing and model operation method, device and computer equipment | |
| CN111401722B (en) | Intelligent decision method and intelligent decision system | |
| CN109949286A (en) | Method and apparatus for output information | |
| CN112799850A (en) | Model training method, model prediction method, and model control system | |
| Mass et al. | Alternative tests for the selection of model variables | |
| CN112884092A (en) | AI model generation method, electronic device, and storage medium | |
| CN112085019A (en) | Character recognition model generation system, method and device and computer equipment | |
| CN110990053A (en) | Method for creating and using machine learning scheme template and device | |
| CN110245679A (en) | Image clustering method, device, electronic equipment and computer readable storage medium | |
| CN108255707A (en) | Development roles creation method, device, equipment and the storage medium of test case | |
| CN112351337B (en) | Video quality inspection method and device, computer equipment and storage medium | |
| CN110969600A (en) | Product defect detection method and device, electronic equipment and storage medium | |
| CN108647235A (en) | A kind of data analysing method, equipment and medium based on data warehouse | |
| US11809969B2 (en) | Dynamically integrating interactive machine learning multi-models | |
| US9508006B2 (en) | System and method for identifying trees | |
| CN110909768B (en) | Method and device for acquiring marked data | |
| CN113821296B (en) | Visual interface generation method, electronic equipment and storage medium | |
| US20230169149A1 (en) | Image recognition training data quality enhancement | |
| US11941519B2 (en) | Machine learning training platform | |
| CN108921138A (en) | Method and apparatus for generating information |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20201218 |
|
| RJ01 | Rejection of invention patent application after publication |