CN117654907A - Automatic eliminating method and system for strip detector - Google Patents
Automatic eliminating method and system for strip detector Download PDFInfo
- Publication number
- CN117654907A CN117654907A CN202311621541.1A CN202311621541A CN117654907A CN 117654907 A CN117654907 A CN 117654907A CN 202311621541 A CN202311621541 A CN 202311621541A CN 117654907 A CN117654907 A CN 117654907A
- Authority
- CN
- China
- Prior art keywords
- detection
- product
- strip
- setting
- defect identification
- 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
- 238000000034 method Methods 0.000 title claims abstract description 32
- 238000001514 detection method Methods 0.000 claims abstract description 301
- 230000007547 defect Effects 0.000 claims abstract description 154
- 230000004913 activation Effects 0.000 claims abstract description 30
- 230000003213 activating effect Effects 0.000 claims abstract description 10
- 238000012360 testing method Methods 0.000 claims description 78
- 238000004458 analytical method Methods 0.000 claims description 21
- 239000000463 material Substances 0.000 claims description 21
- 230000011218 segmentation Effects 0.000 claims description 18
- 230000035945 sensitivity Effects 0.000 claims description 16
- 238000012795 verification Methods 0.000 claims description 15
- 238000012549 training Methods 0.000 claims description 14
- 238000004422 calculation algorithm Methods 0.000 claims description 7
- 238000013527 convolutional neural network Methods 0.000 claims description 5
- 238000007689 inspection Methods 0.000 claims 2
- 238000010200 validation analysis Methods 0.000 claims 1
- 238000004519 manufacturing process Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000013524 data verification Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000001815 facial effect Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000001303 quality assessment method Methods 0.000 description 1
- 238000011895 specific detection Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention discloses an automatic rejection method and a system for a strip detector, which relate to the technical field of automatic rejection detection, wherein the method comprises the following steps: generating an equipment activation instruction and a belt detection setting instruction according to the target detection product; activating a strip detector and a strip detection module; setting the strip detection module to obtain a strip detection module with the set strip detection module; detecting the target detection product based on the strip detection machine and the set strip detection module to obtain a product detection data set; carrying out product defect identification according to the product detection data set to obtain a product detection report; according to the product detection report, the target detection product is automatically removed, so that the problem that different products cannot be effectively removed due to insufficient rigor and insufficient completeness of automatic removal detection work in the prior art is solved, and the automatic removal accuracy of the strip detector is improved.
Description
Technical Field
The invention relates to the technical field of automatic rejection detection, in particular to an automatic rejection method and an automatic rejection system for a strip detector.
Background
With the development of economy and science, strip detectors are increasingly being used in modern production lines for quality detection of products. The strip detector is a machine for detecting the quality of products on a production line. It generally consists of an image acquisition system, a control system and a culling system. The automatic eliminating method is designed for realizing an unmanned production line. By using automatic equipment such as a conveyor belt, the automatic grabbing, moving and classifying operations on the products can be realized, so that the production efficiency and the product quality are improved.
The problem that the automatic rejection detection work in the prior art cannot effectively reject different products due to insufficient rigor and insufficient completeness, so that the automatic rejection accuracy of the strip detector is low finally.
Disclosure of Invention
The application provides an automatic rejection method and an automatic rejection system for a strip detector, which solve the problem that different products cannot be effectively rejected due to insufficient rigor and insufficient completeness of automatic rejection detection work in the prior art, so that the automatic rejection accuracy of the strip detector is improved.
In view of the foregoing, the present application provides an automatic reject method for a strip detector.
In a first aspect, the present application provides an automatic reject method for a strip detector, the method comprising: generating an equipment activation instruction and a belt detection setting instruction according to a target detection product, wherein the belt detection setting instruction comprises preset belt detection setting data corresponding to the target detection product; activating a strip material detector and a strip material detection module according to the equipment activation instruction; setting the strip detection module based on the strip detection setting instruction to obtain a strip detection module with the set strip detection module; detecting the target detection product based on the strip detection machine and the set strip detection module to obtain a product detection data set; based on a pre-constructed product defect analysis channel, carrying out product defect identification according to the product detection data set to obtain a product detection report; and automatically removing the target detection product according to the product detection report.
In a second aspect, the present application provides an automatic reject system for a strip detector, the system comprising: an activation instruction module: generating an equipment activation instruction and a belt detection setting instruction according to a target detection product, wherein the belt detection setting instruction comprises preset belt detection setting data corresponding to the target detection product; and a device activation module: activating a strip material detector and a strip material detection module according to the equipment activation instruction; the instruction setting module: setting the strip detection module based on the strip detection setting instruction to obtain a strip detection module with the set strip detection module; the product detection module: detecting the target detection product based on the strip detection machine and the set strip detection module to obtain a product detection data set; defect identification module: based on a pre-constructed product defect analysis channel, carrying out product defect identification according to the product detection data set to obtain a product detection report; and an automatic rejecting module: and automatically removing the target detection product according to the product detection report.
One or more technical solutions provided in the present application have at least the following technical effects or advantages:
according to the automatic rejection method and the system for the strip detector, the equipment activation instruction and the strip detection setting instruction are generated through the target detection product, then the strip detector and the strip detection module are activated according to the equipment activation instruction, the strip detection module is set, the strip detection module with the set being completed is obtained, the target detection product is further detected based on the strip detector and the strip detection module with the set being completed, a product detection data set is obtained, then the product defect analysis channel is constructed in advance, the product defect identification is carried out according to the product detection data set, a product detection report is obtained, finally the target detection product is automatically rejected according to the product detection report, the problem that different products cannot be effectively rejected due to insufficient rigor and insufficient completeness in automatic rejection detection in the prior art is solved, and the automatic rejection accuracy of the strip detector is improved.
Drawings
FIG. 1 is a schematic flow diagram of an automatic reject method for a strip detector;
fig. 2 is a schematic structural diagram of an automatic reject system for a strip detector according to the present application.
Reference numerals illustrate: the device comprises an activation instruction module 11, a device activation module 12, an instruction setting module 13, a product detection module 14, a defect identification module 15 and an automatic rejection module 16.
Detailed Description
According to the automatic rejection method and system for the strip detector, equipment activating instructions and strip detection setting instructions are generated through target detection products, then the strip detector and the strip detection module are activated according to the equipment activating instructions, the strip detection module is set, the strip detection module with the set being completed is obtained, the target detection products are further detected based on the strip detector and the strip detection module with the set being completed, a product detection data set is obtained, then product defect analysis channels are built in advance, product defect identification is carried out according to the product detection data set, a product detection report is obtained, and finally automatic rejection is carried out on the target detection products according to the product detection report. The problem that different products cannot be effectively removed due to insufficient rigor and insufficient completeness of automatic removal detection in the prior art is solved, and the automatic removal accuracy of the strip detector is improved.
Example 1
As shown in fig. 1, the present application provides an automatic rejecting method and system for a strip detector, the method comprising:
generating an equipment activation instruction and a belt detection setting instruction according to a target detection product, wherein the belt detection setting instruction comprises preset belt detection setting data corresponding to the target detection product;
the target detection product is a product of the belt material detection machine to be used, the target detection product comprises target detection product information, the target detection product comprises basic information such as model, specification and performance of the target detection product, preset belt material detection setting data are determined according to the target detection product information, and the preset belt material detection setting data comprise setting data of a detection area, setting of a detection speed, identification of a material type and the like. According to the information of the target detection product and the preset strip detection setting data, a device activation instruction and a strip detection setting instruction can be generated, wherein the device activation instruction is an instruction for starting the device, configuring device parameters, connecting a network and the like so as to ensure that the device can normally work and execute subsequent detection tasks; the belt detection setting instruction refers to operation instructions such as setting a detection area, adjusting a detection speed, selecting a material type and the like, so that the equipment can accurately detect the material. And generating an equipment activation instruction and a waiting detection setting instruction according to the target detection product, and providing a basis for activating the strip detection machine and the strip detection module according to the equipment activation instruction.
Activating a strip material detector and a strip material detection module according to the equipment activation instruction;
setting the strip detection module based on the strip detection setting instruction to obtain a strip detection module with the set strip detection module;
sending the strip detection setting instruction to a strip inspector;
carrying out identity-authority authentication on the strip inspector to obtain an identity-authority authentication result;
when the identity-permission authentication result passes, setting the strip detection module by the strip inspector according to the strip detection setting instruction to obtain a real-time setting-strip detection module and real-time strip detection setting data corresponding to the real-time setting-strip detection module;
verifying the real-time belt detection setting data to obtain a belt detection setting verification result;
and when the verification result of the strip detection setting is passed, obtaining the strip detection module with the finished setting according to the real-time setting-strip detection module.
Activating the strip material detection machine according to the equipment activation instruction, converting the strip material detection machine from a standby state to a starting state, and configuring the parameters of the strip material detection module according to preset strip material detection setting data to obtain the strip material detection module. The method comprises the steps of sending a strip detection setting instruction to a strip detector, wherein the strip detector is a special person specially setting a strip detection module, authenticating the identity and authority of the strip detector before sending the strip detection setting instruction, authenticating the identity and authority of the strip detector, obtaining an identity-authority authentication result, ensuring that only the detector with necessary authority can receive and execute the instructions, and checking whether the person has enough authority to receive and execute the instructions or not by the aid of technologies such as password verification, fingerprint identification, facial recognition and the like. When the identity-authority authentication result passes, the strip detector can set the strip detection module according to the received instruction. And after the setting is completed, acquiring a belt detection module set in real time and real-time belt detection setting data corresponding to the setting. In order to ensure the correctness and effectiveness of the setting, the real-time tape detection setting data needs to be verified, wherein the verification includes checking the integrity of the data, conforming to preset specifications or standards, and the like, and obtaining a tape detection setting verification result. When the verification result of the strip detection setting is passed, the set strip detection module can be obtained according to the real-time setting-strip detection module. By setting data verification, accuracy and validity can be ensured, security and stability of the system can be increased, and potential problems caused by incorrect setting can be reduced.
Detecting the target detection product based on the strip detection machine and the set strip detection module to obtain a product detection data set;
image acquisition is carried out on the target detection product according to a plurality of detection cameras in the set belt detection module, so that a plurality of image acquisition results are obtained;
image enhancement is carried out on the plurality of image acquisition results based on a Retinex algorithm, so that a plurality of image data sets are obtained;
performing semantic segmentation on the plurality of image data sets based on a semantic segmentation network to obtain a plurality of image semantic segmentation results;
and generating the product detection data set according to the semantic segmentation results of the plurality of images.
When the tape detection module is set, a plurality of detection cameras therein may be used to capture images of the target detection product. And the acquired images are acquired from different angles, different illumination conditions, and the like. Image enhancement is performed on a plurality of image acquisition results based on a Retinex algorithm, wherein the Retinex algorithm is an algorithm commonly used for image enhancement, and can improve brightness, contrast, color and the like of images, and the Retinex algorithm can enhance the acquired plurality of image results, so that interested parts are highlighted and a plurality of image data sets are obtained. The semantic segmentation network is a deep learning technology, can classify images at pixel level, and can perform semantic segmentation on the enhanced image data to obtain a classification result of each pixel. Semantic segmentation is carried out on the plurality of image data sets based on the semantic segmentation network, a plurality of image semantic segmentation results are obtained, and a product detection data set is generated according to the semantic segmentation results. The product detection data set comprises various characteristics and attributes of the product, and a data basis is provided for subsequent product defect identification based on a pre-constructed product defect analysis channel and according to the product detection data set, a product detection report is obtained.
Based on a pre-constructed product defect analysis channel, carrying out product defect identification according to the product detection data set to obtain a product detection report;
and automatically removing the target detection product according to the product detection report.
Setting a multi-dimensional defect identification index based on the target detection product, wherein the multi-dimensional defect identification index comprises a plurality of defect type indexes;
constructing a plurality of defect identification branches based on the multi-dimensional defect identification indexes;
integrating the defect identification branches to generate the product defect analysis channel;
traversing the product detection data set based on the product defect analysis channel to analyze defects of the multi-dimensional defect identification indexes, and obtaining a plurality of defect identification results;
and integrating data according to the defect identification results to generate the product detection report.
Based on the target detection product, a multi-dimensional defect identification index is set, wherein the multi-dimensional defect identification index comprises a plurality of defect type indexes, and the defect type indexes are defects of different types, such as appearance defects, performance defects and the like. Based on the multidimensional defect identification indexes, a plurality of defect identification branches are constructed, and for each defect type index, an independent defect identification branch can be constructed, and a plurality of defect identification branches are integrated to construct a product defect analysis channel. The product defect analysis channel is an analysis model for analyzing and processing each defect identification branch. Based on the product defect analysis channel, traversing the product detection data set to carry out defect analysis of multi-dimensional defect identification indexes, obtaining identification results of each product on different defect type indexes, integrating a plurality of defect identification results, carrying out identification result analysis, generating a product detection report, reporting information such as defect type, severity and the like of the product, and contents such as overall quality assessment and statistical analysis, sending the defect report to rejection execution equipment, and automatically rejecting a target detection product after obtaining the product detection report by the rejection execution equipment. Product detection reports are generated by integrating a plurality of defect identification results, so that more comprehensive and accurate product detection reports are obtained, and a data basis is provided for automatically eliminating target detection products according to the product detection reports.
Further, the method further comprises:
consistency comparison is carried out on the real-time belt detection setting data and the preset belt detection setting data, so that real-time setting consistency is obtained;
judging whether the real-time setting consistency is smaller than a preset consistency;
if the real-time setting consistency is smaller than the preset consistency, the obtained belt detection setting verification result is not passed, and a belt detection setting early warning instruction is generated.
After the sending of the strip detection setting instruction, the identity-authority authentication, the strip detection module setting and the verification of the real-time strip detection setting data are completed, the real-time strip detection setting data can be further subjected to consistency comparison with the preset strip detection setting data so as to obtain the real-time setting consistency. By determining whether the real-time set degree of consistency is less than the preset degree of consistency, it may be determined whether the setting of the strip detection module meets the intended standard or specification. If the real-time setting consistency is smaller than the preset consistency, the belt detection module and the preset setting are indicated to have larger deviation, and the belt detection setting verification result is not passed at the moment, and a belt detection setting early warning instruction is generated, so that the belt detection setting early warning instruction aims to timely find and solve the problem and comprises a guide for resetting or adjusting the belt detection module with inconsistent setting, and notifying and reminding related personnel. Through judging the real-time consistency of setting, the management of setting the belt detection module can be enhanced, and the problem can be reported in time.
Further, the method further comprises:
traversing the multi-dimensional defect identification index and extracting a first defect type index;
performing defect identification record acquisition based on the target detection product and the first defect type index to obtain a first defect identification record library;
obtaining preset dividing weights, and carrying out data division on the first defect identification record library according to the preset dividing weights to obtain a first training data sequence and a first test data sequence;
performing supervised training on the first training data sequence based on a convolutional neural network to obtain a first defect identification network;
testing the first defect identification network according to the first test data sequence to obtain a first defect identification branch meeting a preset convergence condition;
the first defect identification branch is added to the plurality of defect identification branches.
Traversing the multidimensional defect identification index, extracting a first defect type index based on priority, collecting defect identification records by using a target detection product and the first defect type index as references, and carrying out specific detection and recording on the product to obtain detailed information about the first defect type. Obtaining preset dividing weights, and carrying out data division on the first defect identification record library according to the preset dividing weights to obtain input data and supervision data, namely a first training data sequence and a first test data sequence, wherein the input data is the first training data sequence, and the supervision data is the first test data sequence. The first training data sequence is subjected to supervised training based on a convolutional neural network, wherein the convolutional neural network is a feedforward neural network which comprises convolutional calculation and has a depth structure, and input information can be subjected to translation invariant classification mainly according to a hierarchical structure, and training is performed according to the hierarchical structure to obtain a first defect identification network. The first test data sequence is used to test the performance of the first defect identification network. If the performance of the network reaches a preset convergence condition or threshold, the network branch is considered to be qualified, and a first defect identification branch meeting the preset convergence condition is obtained. And adding the first defect identification branch meeting the preset convergence condition into the defect identification branches. When a new product needs to be detected, the branch can be called as an independent defect identification branch. By setting the identification branch for each defect type and independently calling and analyzing, the accuracy and efficiency of defect identification can be improved.
Further, the method further comprises:
performing feature recognition according to the first test data sequence to obtain first test input data and first test expected output data;
testing the first defect identification network according to the first test input data to obtain first test output data;
comparing the first test expected output data with the first test output data to obtain a first test accuracy rate and a first test accuracy rate;
calculating a first test recognition sensitivity based on the first test accuracy and the first test accuracy;
judging whether the first test recognition sensitivity meets the preset convergence condition or not;
and if the first test identification sensitivity meets the preset convergence condition, adding the first defect identification network to the first defect identification branch.
In order to extract the features in the data for subsequent test evaluation, feature recognition is performed on a first test data sequence, wherein the first test input data refers to the tested input data, and the first test expected output data refers to network output results expected according to the features of the test data sequence. And testing the performance of the first defect identification network by using the first test input data, and calculating according to the input data to obtain a corresponding output result, namely the first test output data. And comparing the first test expected output data with the first test output data to evaluate the identification accuracy and the precision of the network. Accuracy refers to the ratio of the number of correctly identified samples to the total number of samples, and accuracy refers to the ratio of the number of correctly identified samples to the number of all identified positive samples; based on the first test accuracy and the first test accuracy, a first test recognition sensitivity is calculated, wherein the recognition sensitivity generally refers to the recognition accuracy of a classifier on a positive sample, the first test accuracy and the first test accuracy are calculated through combination of a confusion matrix, and the recognition sensitivity is calculated according to a sensitivity formula corresponding to the accuracy and the accuracy. Judging whether the first test recognition sensitivity meets a preset convergence condition or not, namely judging that the network branch is qualified if the first test recognition sensitivity meets the preset convergence condition, setting the preset convergence condition based on the requirement of practical application, and adding the first defect recognition network to the first defect recognition branch if the first test recognition sensitivity meets the preset convergence condition. And optimizing the defect identification network by setting preset convergence conditions, so that the accuracy of the defect identification network is improved.
Example two
Based on the same inventive concept as one of the automatic reject methods for a strip detector in the previous embodiments, as shown in fig. 2, the present application provides an automatic reject system for a strip detector, the system comprising:
the activation instruction module 11: the activation instruction module 11 is configured to generate an equipment activation instruction and a strip detection setting instruction according to a target detection product, where the strip detection setting instruction includes predetermined strip detection setting data corresponding to the target detection product;
device activation module 12: the device activation module 12 is configured to activate a tape detector and a tape detection module according to the device activation instruction;
instruction setting module 13: the instruction setting module 13 is configured to set the tape detection module based on the tape detection setting instruction, to obtain a tape detection module with a set tape;
product detection module 14: the product detection module 14 is configured to detect the target detection product based on the tape stock detection machine and the set tape stock detection module, and obtain a product detection data set;
defect recognition module 15: the defect recognition module 15 is configured to perform product defect recognition according to the product detection data set based on a product defect analysis channel constructed in advance, so as to obtain a product detection report;
automatic culling module 16: the automatic rejecting module 16 is configured to automatically reject the target detection product according to the product detection report.
Further, the defect identifying module 15 includes the following steps:
further, the system further comprises:
further, the instruction setting module 13 includes the following execution steps:
sending the strip detection setting instruction to a strip inspector;
carrying out identity-authority authentication on the strip inspector to obtain an identity-authority authentication result;
when the identity-permission authentication result passes, setting the strip detection module by the strip inspector according to the strip detection setting instruction to obtain a real-time setting-strip detection module and real-time strip detection setting data corresponding to the real-time setting-strip detection module;
verifying the real-time belt detection setting data to obtain a belt detection setting verification result;
and when the verification result of the strip detection setting is passed, obtaining the strip detection module with the finished setting according to the real-time setting-strip detection module.
Further, the instruction setting module 13 includes the following execution steps:
consistency comparison is carried out on the real-time belt detection setting data and the preset belt detection setting data, so that real-time setting consistency is obtained;
judging whether the real-time setting consistency is smaller than a preset consistency;
if the real-time setting consistency is smaller than the preset consistency, the obtained belt detection setting verification result is not passed, and a belt detection setting early warning instruction is generated.
Further, the product detection module 14 includes the following execution steps:
image acquisition is carried out on the target detection product according to a plurality of detection cameras in the set belt detection module, so that a plurality of image acquisition results are obtained;
image enhancement is carried out on the plurality of image acquisition results based on a Retinex algorithm, so that a plurality of image data sets are obtained;
performing semantic segmentation on the plurality of image data sets based on a semantic segmentation network to obtain a plurality of image semantic segmentation results;
and generating the product detection data set according to the semantic segmentation results of the plurality of images.
Further, the defect identifying module 15 includes the following steps:
setting a multi-dimensional defect identification index based on the target detection product, wherein the multi-dimensional defect identification index comprises a plurality of defect type indexes;
constructing a plurality of defect identification branches based on the multi-dimensional defect identification indexes;
integrating the defect identification branches to generate the product defect analysis channel;
traversing the product detection data set based on the product defect analysis channel to analyze defects of the multi-dimensional defect identification indexes, and obtaining a plurality of defect identification results;
and integrating data according to the defect identification results to generate the product detection report.
Further, the defect identifying module 15 includes the following steps:
traversing the multi-dimensional defect identification index and extracting a first defect type index;
performing defect identification record acquisition based on the target detection product and the first defect type index to obtain a first defect identification record library;
obtaining preset dividing weights, and carrying out data division on the first defect identification record library according to the preset dividing weights to obtain a first training data sequence and a first test data sequence;
performing supervised training on the first training data sequence based on a convolutional neural network to obtain a first defect identification network;
testing the first defect identification network according to the first test data sequence to obtain a first defect identification branch meeting a preset convergence condition;
the first defect identification branch is added to the plurality of defect identification branches.
Further, the defect identifying module 15 includes the following steps:
performing feature recognition according to the first test data sequence to obtain first test input data and first test expected output data;
testing the first defect identification network according to the first test input data to obtain first test output data;
comparing the first test expected output data with the first test output data to obtain a first test accuracy rate and a first test accuracy rate;
calculating a first test recognition sensitivity based on the first test accuracy and the first test accuracy;
judging whether the first test recognition sensitivity meets the preset convergence condition or not;
and if the first test identification sensitivity meets the preset convergence condition, adding the first defect identification network to the first defect identification branch.
From the foregoing detailed description of an automatic reject method for a strip detector, those skilled in the art will clearly understand that an automatic reject method for a strip detector in this embodiment, for an apparatus disclosed in the embodiments, since the apparatus corresponds to the method disclosed in the embodiments, the description is relatively simple, and the relevant points refer to the method section.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (8)
1. An automatic reject method for a strip detector, the method comprising:
generating an equipment activation instruction and a belt detection setting instruction according to a target detection product, wherein the belt detection setting instruction comprises preset belt detection setting data corresponding to the target detection product;
activating a strip material detector and a strip material detection module according to the equipment activation instruction;
setting the strip detection module based on the strip detection setting instruction to obtain a strip detection module with the set strip detection module;
detecting the target detection product based on the strip detection machine and the set strip detection module to obtain a product detection data set;
based on a pre-constructed product defect analysis channel, carrying out product defect identification according to the product detection data set to obtain a product detection report;
and automatically removing the target detection product according to the product detection report.
2. The method of claim 1, wherein setting the tape detection module based on the tape detection setting instruction to obtain a set tape detection module comprises:
sending the strip detection setting instruction to a strip inspector;
carrying out identity-authority authentication on the strip inspector to obtain an identity-authority authentication result;
when the identity-permission authentication result passes, setting the strip detection module by the strip inspector according to the strip detection setting instruction to obtain a real-time setting-strip detection module and real-time strip detection setting data corresponding to the real-time setting-strip detection module;
verifying the real-time belt detection setting data to obtain a belt detection setting verification result;
and when the verification result of the strip detection setting is passed, obtaining the strip detection module with the finished setting according to the real-time setting-strip detection module.
3. The method of claim 2, wherein validating the real-time tape detection setting data to obtain a tape detection setting validation result comprises:
consistency comparison is carried out on the real-time belt detection setting data and the preset belt detection setting data, so that real-time setting consistency is obtained;
judging whether the real-time setting consistency is smaller than a preset consistency;
if the real-time setting consistency is smaller than the preset consistency, the obtained belt detection setting verification result is not passed, and a belt detection setting early warning instruction is generated.
4. The method of claim 1, wherein obtaining a product detection dataset comprises:
image acquisition is carried out on the target detection product according to a plurality of detection cameras in the set belt detection module, so that a plurality of image acquisition results are obtained;
image enhancement is carried out on the plurality of image acquisition results based on a Retinex algorithm, so that a plurality of image data sets are obtained;
performing semantic segmentation on the plurality of image data sets based on a semantic segmentation network to obtain a plurality of image semantic segmentation results;
and generating the product detection data set according to the semantic segmentation results of the plurality of images.
5. The method of claim 1, wherein performing product defect identification from the product inspection dataset based on a pre-constructed product defect parsing channel to obtain a product inspection report, comprising:
setting a multi-dimensional defect identification index based on the target detection product, wherein the multi-dimensional defect identification index comprises a plurality of defect type indexes;
constructing a plurality of defect identification branches based on the multi-dimensional defect identification indexes;
integrating the defect identification branches to generate the product defect analysis channel;
traversing the product detection data set based on the product defect analysis channel to analyze defects of the multi-dimensional defect identification indexes, and obtaining a plurality of defect identification results;
and integrating data according to the defect identification results to generate the product detection report.
6. The method of claim 5, wherein constructing a plurality of defect identification branches based on the multi-dimensional defect identification index comprises:
traversing the multi-dimensional defect identification index and extracting a first defect type index;
performing defect identification record acquisition based on the target detection product and the first defect type index to obtain a first defect identification record library;
obtaining preset dividing weights, and carrying out data division on the first defect identification record library according to the preset dividing weights to obtain a first training data sequence and a first test data sequence;
performing supervised training on the first training data sequence based on a convolutional neural network to obtain a first defect identification network;
testing the first defect identification network according to the first test data sequence to obtain a first defect identification branch meeting a preset convergence condition;
the first defect identification branch is added to the plurality of defect identification branches.
7. The method of claim 6, wherein testing the first defect recognition network based on the first test data sequence to obtain a first defect recognition branch that satisfies a preset convergence condition comprises:
performing feature recognition according to the first test data sequence to obtain first test input data and first test expected output data;
testing the first defect identification network according to the first test input data to obtain first test output data;
comparing the first test expected output data with the first test output data to obtain a first test accuracy rate and a first test accuracy rate;
calculating a first test recognition sensitivity based on the first test accuracy and the first test accuracy;
judging whether the first test recognition sensitivity meets the preset convergence condition or not;
and if the first test identification sensitivity meets the preset convergence condition, adding the first defect identification network to the first defect identification branch.
8. An automatic reject system for a strip detector, the system comprising:
an activation instruction module: generating an equipment activation instruction and a belt detection setting instruction according to a target detection product, wherein the belt detection setting instruction comprises preset belt detection setting data corresponding to the target detection product;
and a device activation module: activating a strip material detector and a strip material detection module according to the equipment activation instruction;
the instruction setting module: setting the strip detection module based on the strip detection setting instruction to obtain a strip detection module with the set strip detection module;
the product detection module: detecting the target detection product based on the strip detection machine and the set strip detection module to obtain a product detection data set;
defect identification module: based on a pre-constructed product defect analysis channel, carrying out product defect identification according to the product detection data set to obtain a product detection report;
and an automatic rejecting module: and automatically removing the target detection product according to the product detection report.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311621541.1A CN117654907A (en) | 2023-11-29 | 2023-11-29 | Automatic eliminating method and system for strip detector |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311621541.1A CN117654907A (en) | 2023-11-29 | 2023-11-29 | Automatic eliminating method and system for strip detector |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117654907A true CN117654907A (en) | 2024-03-08 |
Family
ID=90072634
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311621541.1A Pending CN117654907A (en) | 2023-11-29 | 2023-11-29 | Automatic eliminating method and system for strip detector |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117654907A (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111667455A (en) * | 2020-04-28 | 2020-09-15 | 广东三三智能科技有限公司 | AI detection method for various defects of brush |
| CN113900935A (en) * | 2021-09-30 | 2022-01-07 | 未鲲(上海)科技服务有限公司 | Automatic defect identification method and device, computer equipment and storage medium |
| CN114750510A (en) * | 2022-04-29 | 2022-07-15 | 道恩周氏(青岛)复合包装材料有限公司 | Defect marking and automatic removing method, system, equipment and medium |
| CN115452737A (en) * | 2022-08-26 | 2022-12-09 | 云南昆船烟草设备有限公司 | Tobacco bale mildew tobacco leaf online detection and elimination method, platform and readable storage medium |
| CN116188475A (en) * | 2023-05-05 | 2023-05-30 | 德中(深圳)激光智能科技有限公司 | Intelligent control method, system and medium for automatic optical detection of appearance defects |
| CN116563641A (en) * | 2023-05-30 | 2023-08-08 | 江西软件职业技术大学 | Surface defect identification method and system based on small target detection |
| WO2023173598A1 (en) * | 2022-03-15 | 2023-09-21 | 中国华能集团清洁能源技术研究院有限公司 | Fan blade defect detection method and system based on improved ssd model |
| CN116883393A (en) * | 2023-09-05 | 2023-10-13 | 青岛理工大学 | Metal surface defect detection method based on anchor frame-free target detection algorithm |
| WO2023216740A1 (en) * | 2022-05-09 | 2023-11-16 | 苏州天准科技股份有限公司 | Crushed material detection and removal apparatus and method, and sorting system |
-
2023
- 2023-11-29 CN CN202311621541.1A patent/CN117654907A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111667455A (en) * | 2020-04-28 | 2020-09-15 | 广东三三智能科技有限公司 | AI detection method for various defects of brush |
| CN113900935A (en) * | 2021-09-30 | 2022-01-07 | 未鲲(上海)科技服务有限公司 | Automatic defect identification method and device, computer equipment and storage medium |
| WO2023173598A1 (en) * | 2022-03-15 | 2023-09-21 | 中国华能集团清洁能源技术研究院有限公司 | Fan blade defect detection method and system based on improved ssd model |
| CN114750510A (en) * | 2022-04-29 | 2022-07-15 | 道恩周氏(青岛)复合包装材料有限公司 | Defect marking and automatic removing method, system, equipment and medium |
| WO2023216740A1 (en) * | 2022-05-09 | 2023-11-16 | 苏州天准科技股份有限公司 | Crushed material detection and removal apparatus and method, and sorting system |
| CN115452737A (en) * | 2022-08-26 | 2022-12-09 | 云南昆船烟草设备有限公司 | Tobacco bale mildew tobacco leaf online detection and elimination method, platform and readable storage medium |
| CN116188475A (en) * | 2023-05-05 | 2023-05-30 | 德中(深圳)激光智能科技有限公司 | Intelligent control method, system and medium for automatic optical detection of appearance defects |
| CN116563641A (en) * | 2023-05-30 | 2023-08-08 | 江西软件职业技术大学 | Surface defect identification method and system based on small target detection |
| CN116883393A (en) * | 2023-09-05 | 2023-10-13 | 青岛理工大学 | Metal surface defect detection method based on anchor frame-free target detection algorithm |
Non-Patent Citations (1)
| Title |
|---|
| 时培成;韦山;: "进口包装机检测剔除装置的创新设计与运用", 液压与气动, no. 01, 20 January 2007 (2007-01-20) * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11029260B2 (en) | Solder paste printing quality inspection system and method | |
| CN107528832B (en) | Baseline construction and unknown abnormal behavior detection method for system logs | |
| CN112036755B (en) | Supervision method and system for quality detection of building engineering | |
| TW201514472A (en) | Method of optical defect detection through image analysis and data mining integrated | |
| CN111783663A (en) | Algorithm evaluation system and detection method for performance detection of human evidence verification equipment | |
| CN115409131B (en) | Production line abnormity detection method based on SPC process control system | |
| CN108711148A (en) | A kind of wheel tyre defect intelligent detecting method based on deep learning | |
| CN110378618A (en) | Quality evaluating method and system based on online pipe tobacco surface defects detection | |
| CN113888480A (en) | MES-based quality tracing method and system | |
| CN112419268A (en) | Method, device, equipment and medium for detecting image defects of power transmission line | |
| CN110738630A (en) | Training method and detection system of recursive deep learning system | |
| CN108763966A (en) | A kind of Tail gas measuring cheating supervisory systems and method | |
| CN115795021A (en) | Big data risk monitoring, recognizing and early warning device and system | |
| WO2023113773A1 (en) | Manufacturing defect detection augmented training sets | |
| CN111078565A (en) | Analysis method of software test result based on HOG feature extraction and SVM multi-classifier | |
| CN112199295A (en) | Deep neural network defect positioning method and system based on frequency spectrum | |
| CN117654907A (en) | Automatic eliminating method and system for strip detector | |
| CN111460198A (en) | Method and device for auditing picture timestamp | |
| CN114565883A (en) | Graphic recognition algorithm for operation faults of equipment | |
| CN111723795B (en) | Abnormal license plate recognition method and device, electronic equipment and storage medium | |
| CN113706178A (en) | Product inspection process supervision method | |
| Beljadid et al. | Application of deep learning for the detection of default in fabric texture | |
| CN113569879A (en) | Training method of abnormal recognition model, abnormal account recognition method and related device | |
| US20240046617A1 (en) | Machine Learning-Based Generation of Rule-Based Classification Recipes for Inspection System | |
| Candra et al. | Computer Vision Implementation in Scratch Inspection and Color Detection on The Car Roof Surface |
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 |