CN115901794A - System and method for detecting bottle opening flaws through strip-shaped light source - Google Patents
System and method for detecting bottle opening flaws through strip-shaped light source Download PDFInfo
- Publication number
- CN115901794A CN115901794A CN202310126946.1A CN202310126946A CN115901794A CN 115901794 A CN115901794 A CN 115901794A CN 202310126946 A CN202310126946 A CN 202310126946A CN 115901794 A CN115901794 A CN 115901794A
- Authority
- CN
- China
- Prior art keywords
- image
- detection
- feature map
- bottle
- enhanced
- 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.)
- Granted
Links
Images
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 application discloses a system and a method for detecting bottle opening flaws by using a bar-shaped light source. Firstly, a detection image of a bottle to be detected and a reference image of a reference bottle, which are acquired by a camera, are respectively subjected to an image overexposure corrector to obtain a corrected detection image and a corrected reference image, then the corrected detection image and the corrected reference image are subjected to a double detection model to obtain a detection characteristic diagram and a reference characteristic diagram, then the detection characteristic diagram and the reference characteristic diagram are subjected to a spatial attention module to obtain an enhanced detection characteristic diagram and an enhanced reference characteristic diagram, then a difference characteristic diagram between the enhanced detection characteristic diagram and the enhanced reference characteristic diagram is calculated, and finally, a classification result for indicating whether the bottle to be detected has a bottle mouth defect is obtained by a classifier after feature value discrimination optimization is carried out on the difference characteristic diagram. By the mode, whether the bottle to be detected has the bottle opening defect can be judged.
Description
Technical Field
The application relates to the technical field of intelligent detection, in particular to a system and a method for detecting bottle opening flaws by using a bar-shaped light source.
Background
A bottle is a container made of a versatile plastic, porcelain or glass for containing liquid substances. Under the pressure of market supply and demand, the improvement of product quality is the guarantee of sales volume, so the production yield and production efficiency need to be improved in the production process of bottles so as to reduce the production cost to the minimum, and the aim of bottle manufacturers is to improve the detection and screening of defective bottles in the mass production stage while improving the production efficiency.
The defects of the bottle comprise pits, holes, burnt particles, black spots, stains, wrinkles, bracing wires, color lines, cracks, bulges, excessive materials of the bottle body, poor cutting, material shortage of the bottle opening, material blockage of the bottle opening, flash and the like.
Therefore, an optimized finish flaw detection scheme is desired.
Disclosure of Invention
The present application is proposed to solve the above-mentioned technical problems. The embodiment of the application provides a system and a method for detecting bottle opening flaws by using a bar-shaped light source. Firstly, a detection image of a bottle to be detected and a reference image of a reference bottle, which are acquired by a camera, are respectively subjected to an image overexposure corrector to obtain a corrected detection image and a corrected reference image, then the corrected detection image and the corrected reference image are subjected to a double detection model to obtain a detection characteristic diagram and a reference characteristic diagram, then the detection characteristic diagram and the reference characteristic diagram are subjected to a spatial attention module to obtain an enhanced detection characteristic diagram and an enhanced reference characteristic diagram, then a difference characteristic diagram between the enhanced detection characteristic diagram and the enhanced reference characteristic diagram is calculated, and finally, a classification result for indicating whether the bottle to be detected has a bottle mouth defect is obtained by a classifier after feature value discrimination optimization is carried out on the difference characteristic diagram. By the method, whether the bottle to be detected has the bottle mouth defect or not can be judged.
According to one aspect of the application, a system for detecting bottle mouth flaws by using a bar-shaped light source is provided, and comprises the following components:
the camera module is used for acquiring a detection image of a bottle to be detected and a reference image of a reference bottle, which are acquired by a camera, wherein the reference bottle is a bottle without a bottle mouth defect;
the overexposure correction module is used for enabling the detection image and the reference image to pass through an image overexposure corrector based on an automatic coder-decoder respectively so as to obtain a corrected detection image and a corrected reference image;
a double detection module, configured to pass the corrected detection image and the corrected reference image through a double detection model including a first image encoder and a second image encoder to obtain a detection feature map and a reference feature map, where the first image encoder and the second image encoder have the same network structure;
the spatial enhancement module is used for enabling the detection feature map and the reference feature map to pass through a spatial attention module to obtain an enhanced detection feature map and an enhanced reference feature map;
a difference module for calculating a difference feature map between the enhanced detection feature map and the enhanced reference feature map;
the characteristic discrimination enhancement module is used for optimizing the characteristic value discrimination of the differential characteristic graph to obtain an optimized differential characteristic graph; and
and the detection result generation module is used for enabling the optimized differential characteristic diagram to pass through a classifier to obtain a classification result, and the classification result is used for indicating whether the bottle to be detected has bottle mouth defects or not.
In the system for detecting defects of a bottle opening by using a bar-shaped light source, the overexposure correction module is further configured to:
respectively carrying out image coding on the detection image and the reference image by using an encoder of the automatic codec-based image overexposure corrector to obtain a detection image characteristic representation and a reference image characteristic representation; and
and respectively performing decoding regression on each image feature representation in the detection image feature representation and each image feature representation in the reference image feature representation by using a decoder of the automatic codec-based image overexposure corrector to obtain the corrected detection image and the corrected reference image.
In the system for detecting bottle mouth flaws by using the strip light source, the first image encoder and the second image encoder are depth residual error network models.
In the above system for detecting defects of a bottle opening by using a bar light source, the spatial enhancement module includes:
a depth convolution unit, configured to perform depth convolution encoding on the detection feature map and the reference feature map respectively using a convolution encoding portion of the spatial attention module to obtain a detection convolution feature map and a reference convolution feature map;
a spatial attention diagram generation unit, configured to input the detection convolution feature map and the reference convolution feature map into a spatial attention portion of the spatial attention module respectively to obtain a detection spatial attention diagram and a reference spatial attention diagram;
a spatial attention feature map generation unit, configured to pass the detection spatial attention map and the reference spatial attention map through a Softmax activation function to obtain a detection spatial attention feature map and a reference spatial attention feature map, respectively; and
and the calculation unit is used for calculating the position-based multiplication of the detection space attention feature map and the detection feature map to obtain the enhanced detection feature map, and calculating the position-based multiplication of the reference space attention feature map and the reference feature map to obtain the enhanced reference feature map.
In the system for detecting bottle mouth flaws by using a bar-shaped light source, the difference module is further configured to: calculating the differential feature map between the enhanced detection feature map and the enhanced reference feature map in the following formula;
wherein the formula is:
wherein is present>
Represents the enhanced detection feature map, based on the feature map, and>
represents the enhanced reference feature map>
Represents the differential characteristic map, and>
indicating a difference by position.
In the system for detecting bottle mouth flaws by using the strip light source, the feature discrimination enhancement module is further configured to: carrying out characteristic value discrimination optimization on the differential characteristic diagram by using the following formula to obtain the optimized differential characteristic diagram;
wherein the formula is:
wherein the content of the first and second substances,
represents the differential feature map, is based on the characteristic map, and is based on the characteristic map>
Represents the optimized differential feature map>
And &>
Is a predetermined hyper-parameter, is>
Represents a subtraction by position of the feature map, the division representing each position of the feature map divided by a response value, and->
Representing the convolution operation through a single convolutional layer.
In the above system for detecting defects of a bottle mouth with a bar light source, the detection result generating module includes:
the characteristic diagram unfolding unit is used for unfolding the optimized differential characteristic diagram into classification characteristic vectors according to row vectors or column vectors;
a full-concatenation encoding unit, configured to perform full-concatenation encoding on the classification feature vector using a plurality of full-concatenation layers of the classifier to obtain an encoded classification feature vector; and
and the classification unit is used for enabling the encoding classification characteristic vector to pass through a Softmax classification function of the classifier so as to obtain the classification result.
According to another aspect of the application, a method for detecting bottle mouth flaws by using a strip-shaped light source is provided, and comprises the following steps:
acquiring a detection image of a bottle to be detected and a reference image of a reference bottle, which are acquired by a camera, wherein the reference bottle is a bottle without a bottle mouth defect;
respectively passing the detection image and the reference image through an image overexposure corrector based on an automatic coder-decoder to obtain a corrected detection image and a corrected reference image;
passing the corrected detection image and the corrected reference image through a dual detection model comprising a first image encoder and a second image encoder to obtain a detection feature map and a reference feature map, wherein the first image encoder and the second image encoder have the same network structure;
passing the detection feature map and the reference feature map through a spatial attention module to obtain an enhanced detection feature map and an enhanced reference feature map;
calculating a difference feature map between the enhanced detection feature map and the enhanced reference feature map;
carrying out characteristic value discrimination optimization on the differential characteristic diagram to obtain an optimized differential characteristic diagram; and
and passing the optimized differential characteristic diagram through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the bottle mouth of the bottle to be detected has a defect.
In the method for detecting bottle mouth flaws by using a strip light source, the step of obtaining a corrected detection image and a corrected reference image by respectively passing the detection image and the reference image through an image overexposure corrector based on an automatic codec comprises the following steps:
respectively carrying out image coding on the detection image and the reference image by using an encoder of the automatic codec-based image overexposure corrector to obtain a detection image characteristic representation and a reference image characteristic representation; and
and respectively performing decoding regression on each image feature representation in the detection image feature representation and each image feature representation in the reference image feature representation by using a decoder of the automatic codec-based image overexposure corrector to obtain the corrected detection image and the corrected reference image.
In the method for detecting bottle mouth flaws by using the strip light source, the first image encoder and the second image encoder are depth residual error network models.
Compared with the prior art, the system and the method for detecting the bottle opening flaws by the strip-shaped light source are provided. The method comprises the steps of firstly, respectively enabling a detection image of a bottle to be detected and a reference image of a reference bottle, which are acquired by a camera, to pass through an image overexposure corrector to obtain a corrected detection image and a corrected reference image, then enabling the corrected detection image and the corrected reference image to pass through a dual detection model to obtain a detection characteristic diagram and a reference characteristic diagram, then enabling the detection characteristic diagram and the reference characteristic diagram to pass through a spatial attention module to obtain an enhanced detection characteristic diagram and an enhanced reference characteristic diagram, then calculating a difference characteristic diagram between the enhanced detection characteristic diagram and the enhanced reference characteristic diagram, and finally, optimizing the characteristic value discrimination of the difference characteristic diagram and then enabling the difference characteristic diagram to pass through a classifier to obtain a classification result for representing whether the bottle to be detected has bottle mouth defects. By the mode, whether the bottle to be detected has the bottle mouth defect or not can be judged, so that the problems of low efficiency, omission and the like caused by adopting a manual visual detection method are solved.
Drawings
The above and other objects, features and advantages of the present application will become more apparent by describing in more detail embodiments of the present application with reference to the attached drawings. The accompanying drawings are included to provide a further understanding of the embodiments of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the principles of the application. In the drawings, like reference numbers generally represent like parts or steps.
Fig. 1 is an application scene diagram of a system for detecting bottle mouth flaws by using a bar-shaped light source according to an embodiment of the application.
FIG. 2 is a block diagram illustrating a system for detecting defects on a bottle opening by using a bar light source according to an embodiment of the present disclosure.
FIG. 3 is a block diagram illustrating the spatial enhancement module in the system for detecting defects on a bottle opening with a bar light source according to the embodiment of the present application.
Fig. 4 is a schematic block diagram of the detection result generation module in the system for detecting a bottle mouth flaw by using a bar light source according to the embodiment of the application.
FIG. 5 is a flowchart of a method for detecting defects of a bottle opening by using a bar light source according to an embodiment of the present application.
FIG. 6 is a schematic diagram of a system architecture of a method for detecting a bottle opening defect by using a bar light source according to an embodiment of the present disclosure.
Detailed Description
Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be understood that the described embodiments are only some embodiments of the present application and not all embodiments of the present application, and that the present application is not limited by the example embodiments described herein.
Overview of a scene
In recent years, deep learning and neural networks have been widely used in the fields of computer vision, natural language processing, text signal processing, and the like. In addition, deep learning and neural networks also exhibit a level close to or even exceeding that of humans in the fields of image classification, object detection, semantic segmentation, text translation, and the like.
The deep learning and the development of the neural network provide a new solution for detecting the bottle mouth flaws. In particular, if there is a flaw or defect in the bottle mouth, it will appear in the bottle mouth image, and the machine vision based on the deep neural network will have stronger resolving power than the naked eye, so it will have natural advantages in the detection of the flaw in the bottle mouth. However, the types of bottle mouth flaws are quite numerous and there is uncertainty in the presentation of the images, and therefore, there is a relatively high technical difficulty if the bottle mouth flaw detection is performed in a positive manner.
In order to solve the technical problems, in the technical scheme of the application, bottle opening defect detection is performed through comparison between a detection image of a bottle to be detected and a reference image of a reference bottle and comparison characteristics, wherein the reference bottle is a bottle without bottle opening defects. It can be understood that if the bottle to be detected has a bottle mouth defect, the bottle mouth defect is different from the reference image at the image presenting end, so that whether the bottle to be detected has the bottle mouth defect can be judged through the characteristic expression difference of the bottle mouth defect and the reference image.
Specifically, in the technical scheme of the application, firstly, a detection image of a bottle to be detected and a reference image of a reference bottle are acquired by a camera, wherein the reference bottle is a bottle without a bottle mouth defect. In the technical scheme of this application, in order to improve image quality, the flaw on the outstanding bottle adopts bar light source to treat and detects the bottle and carry out the light filling, and bar light source is used for supplementing the required light source when the body detects of shooing. In particular, the stripes of the bar-shaped light source irradiating the flaws on the bottle can be bent, so that the flaws on the bottle can be more easily highlighted, and the detection is convenient. For example, the principle or structure of the bar light source can refer to the content of a PET body detection system and method based on different light sources disclosed in application No. 201911078082.0. However, since the presence of the stripe light source also causes the problem of image overexposure, image overexposure correction is performed first before feature extraction is performed. Specifically, the detection image and the reference image are respectively passed through an automatic codec-based image overexposure corrector to obtain a corrected detection image and a corrected reference image.
In the technical solution of the present application, the automatic codec includes an image encoder and an image decoder, wherein the image encoder is configured to extract effective component features in the detected image and the reference image, and the image decoder is configured to perform image decoding on the effective component features in the detected image and the reference image, so as to achieve the technical purpose of image overexposure correction. In a specific embodiment of the present application, the image encoder includes one or more convolution layers, and the image decoder includes one or more deconvolution layers, that is, in the embodiment of the present application, the automatic codec-based image overexposure modifier implements the automatic codec by performing convolution-deconvolution operation to implement the image overexposure correction in such a way.
Then, the corrected detection image and the corrected reference image are passed through a dual detection model including a first image encoder and a second image encoder to obtain a detection feature map and a reference feature map, wherein the first image encoder and the second image encoder have the same network structure. That is, feature extraction based on a convolution kernel is performed on the corrected detection image and the corrected reference image respectively using a dual detection model including a first image encoder and a second image encoder having the same network structure to obtain a detection feature map and a reference feature map. In one particular example of the present application, the first image encoder and the second image encoder are depth residual network models.
In particular, in the technical solution of the present application, considering that the contribution degrees of feature values of respective positions on the spatial dimensions of the detection feature map and the reference feature map are different for the final classification judgment, in order to highlight the spatial feature identifiability, the detection feature map and the reference feature map are passed through a spatial attention module to obtain an enhanced detection feature map and an enhanced reference feature map.
After obtaining the enhanced detection feature map and the enhanced reference feature map, further calculating a difference feature map between the enhanced detection feature map and the enhanced reference feature map, so as to represent a high-dimensional feature expression of a difference between the detection image of the bottle to be detected and the reference image of the reference bottle. And then, the differential feature map is passed through a classifier to obtain a classification result for indicating whether the bottle to be detected has the bottle mouth defect, that is, the classifier is used to determine the class probability label to which the differential feature map belongs, wherein the class probability label comprises the bottle to be detected having the bottle mouth defect (first label) and the bottle to be detected having no bottle mouth defect (second label).
In particular, in the technical solution of the present application, when the differential feature map between the enhanced detection feature map and the enhanced reference feature map is calculated, since the enhanced detection feature map and the enhanced reference feature map are each obtained by enhancing a feature space distribution through a spatial attention module, feature values of partial spatial positions in the differential feature map have a more significant importance with respect to other feature values, and if the feature values can be effectively distinguished during classification, it is obvious that a training speed of a classifier and an accuracy of a classification result can be improved.
Therefore, the applicant of the present application describes the differential signature, for example, as an interaction enhancement based on distinguishable physical stimuli, and the representation is:
where the sum is a predetermined hyper-parameter, and represents a position-wise addition and subtraction of the feature map, the division represents each position of the feature map divided by a response value, and represents a convolution operation by a single convolution layer.
Here, the discriminative physical excitation-based interaction enhancement is used to promote interaction between a feature space and a solution space of a classification problem in a back propagation process through gradient descent, and extracts and imitates a feasible feature (actionable feature) in a physical excitation-like manner, thereby obtaining a physical expression of the feasible feature with gradient discriminativity using a general-purpose low-dimensional conductive physical excitation manner, and thus, strengthening active parts in the differential feature map in a training process to promote training speed of a classifier and accuracy of a classification result of a trained classification feature.
Based on this, this application provides a system for bar light source detects bottleneck flaw, it includes: the camera module is used for acquiring a detection image of a bottle to be detected and a reference image of a reference bottle, which are acquired by a camera, wherein the reference bottle is a bottle without a bottle mouth defect; the overexposure correction module is used for enabling the detection image and the reference image to pass through an image overexposure corrector based on an automatic coder-decoder respectively so as to obtain a corrected detection image and a corrected reference image; a double detection module, configured to pass the corrected detection image and the corrected reference image through a double detection model including a first image encoder and a second image encoder to obtain a detection feature map and a reference feature map, where the first image encoder and the second image encoder have the same network structure; the spatial enhancement module is used for enabling the detection feature map and the reference feature map to pass through a spatial attention module to obtain an enhanced detection feature map and an enhanced reference feature map; a difference module for calculating a difference feature map between the enhanced detection feature map and the enhanced reference feature map; the characteristic discrimination enhancement module is used for optimizing the characteristic value discrimination of the differential characteristic graph to obtain an optimized differential characteristic graph; and the detection result generation module is used for enabling the optimized differential characteristic diagram to pass through a classifier to obtain a classification result, and the classification result is used for indicating whether the bottle to be detected has a bottle mouth defect.
Fig. 1 is an application scene diagram of a system for detecting bottle mouth flaws by using a bar-shaped light source according to an embodiment of the application. As shown in fig. 1, in this application scenario, first, a detection image (e.g., D1 as illustrated in fig. 1) of a bottle to be detected (e.g., F1 as illustrated in fig. 1) acquired by a camera (e.g., C as illustrated in fig. 1) and a reference image (e.g., D2 as illustrated in fig. 1) of a reference bottle (e.g., F2 as illustrated in fig. 1) are acquired, wherein the reference bottle is a bottle without a bottle mouth defect, and then, the detection image and the reference image are input into a server (e.g., S as illustrated in fig. 1) deployed with an algorithm for detecting a bottle mouth defect by using a bar light source, wherein the server can process the detection image and the reference image by using the algorithm for detecting a bottle mouth defect by using the bar light source to obtain a classification result for indicating whether the bottle to be detected has a bottle mouth defect.
Having described the general principles of the present application, various non-limiting embodiments of the present application will now be described with reference to the accompanying drawings.
Exemplary System
FIG. 2 is a block diagram illustrating a system for detecting defects on a bottle opening by using a bar light source according to an embodiment of the present disclosure. As shown in fig. 2, a system 100 for detecting defects of a bottle opening by using a bar-shaped light source according to an embodiment of the present application includes: the camera module 110 is used for acquiring a detection image of a bottle to be detected and a reference image of a reference bottle, which are acquired by a camera, wherein the reference bottle is a bottle without a bottle mouth defect; an overexposure correction module 120, configured to pass the detection image and the reference image through an automatic codec-based image overexposure modifier to obtain a corrected detection image and a corrected reference image, respectively; a double detection module 130, configured to pass the corrected detection image and the corrected reference image through a double detection model including a first image encoder and a second image encoder to obtain a detection feature map and a reference feature map, where the first image encoder and the second image encoder have the same network structure; a spatial enhancement module 140, configured to pass the detection feature map and the reference feature map through a spatial attention module to obtain an enhanced detection feature map and an enhanced reference feature map; a difference module 150, configured to calculate a difference feature map between the enhanced detection feature map and the enhanced reference feature map; a feature differentiation enhancement module 160, configured to perform feature value differentiation optimization on the difference feature map to obtain an optimized difference feature map; and a detection result generation module 170, configured to pass the optimized difference feature map through a classifier to obtain a classification result, where the classification result is used to indicate whether the bottle to be detected has a bottle mouth defect.
More specifically, in the embodiment of the present application, the camera module 110 is configured to acquire a detection image of a bottle to be detected and a reference image of a reference bottle, where the reference bottle is a bottle without a bottle mouth defect. If the bottle mouth has flaws or defects, the flaws or defects can be displayed in the bottle mouth image, and the machine vision based on the deep neural network can have stronger resolving power than the naked eye, so that the flaw detection of the bottle mouth has natural advantages. However, the types of finish defects are numerous and there is uncertainty in the presentation in the image, and therefore, there is a relatively high technical difficulty if finish defect detection is performed in a positive manner. In the embodiment of the application, the bottle opening defect detection is carried out through the comparison between the detection image of the bottle to be detected and the reference image of the reference bottle and through the comparison characteristics, wherein the reference bottle is a bottle without the bottle opening defect. It can be understood that if the bottle to be detected has a bottle mouth defect, the bottle mouth defect is different from the reference image at the image presenting end, so that whether the bottle to be detected has the bottle mouth defect can be judged through the characteristic expression difference of the bottle mouth defect and the reference image.
In the technical scheme of this application, in order to improve image quality, can be in the inside of waiting to detect the bottle sets up the bar light source, and the bar light source is used for supplementing the body and shoots required light source when detecting. However, since the presence of the stripe light source also causes the problem of image overexposure, image overexposure correction is performed first before feature extraction is performed.
More specifically, in the embodiment of the present application, the overexposure correction module 120 is configured to pass the detection image and the reference image through an automatic codec-based image overexposure corrector respectively to obtain a corrected detection image and a corrected reference image. In the technical solution of the present application, the automatic codec includes an image encoder and an image decoder, wherein the image encoder is configured to extract effective component features in the detected image and the reference image, and the image decoder is configured to perform image decoding on the effective component features in the detected image and the reference image, so as to achieve the technical purpose of image overexposure correction. In a specific embodiment of the present application, the image encoder includes one or more convolution layers, and the image decoder includes one or more deconvolution layers, that is, in the embodiment of the present application, the automatic codec-based image overexposure modifier implements automatic codec by convolution-deconvolution operation, and thus implements image overexposure correction.
Accordingly, in a specific example, the overexposure correction module 120 is further configured to: respectively carrying out image coding on the detection image and the reference image by using an encoder of the automatic codec-based image overexposure corrector to obtain a detection image characteristic representation and a reference image characteristic representation; and decoding and regressing each image feature representation in the detection image feature representation and each image feature representation in the reference image feature representation respectively by using a decoder of the automatic codec-based image overexposure corrector to obtain the corrected detection image and the corrected reference image.
More specifically, in the embodiment of the present application, the dual detection module 130 is configured to pass the corrected detection image and the corrected reference image through a dual detection model including a first image encoder and a second image encoder to obtain a detection feature map and a reference feature map, where the first image encoder and the second image encoder have the same network structure. That is, feature extraction based on a convolution kernel is performed on the corrected detection image and the corrected reference image respectively using a dual detection model including a first image encoder and a second image encoder having the same network structure to obtain a detection feature map and a reference feature map.
Accordingly, in one particular example, the first image encoder and the second image encoder are depth residual network models.
More specifically, in the embodiment of the present application, the spatial enhancement module 140 is configured to pass the detection feature map and the reference feature map through a spatial attention module to obtain an enhanced detection feature map and an enhanced reference feature map. In particular, in the technical solution of the present application, in consideration of the fact that the contribution degrees of the feature values of the positions in the spatial dimensions of the detection feature map and the reference feature map are different for the final classification judgment, in order to highlight the spatial feature identifiability, the detection feature map and the reference feature map are passed through a spatial attention module to obtain an enhanced detection feature map and an enhanced reference feature map.
Accordingly, in one specific example, as shown in fig. 3, the spatial enhancement module 140 includes: a depth convolution unit 141, configured to perform depth convolution encoding on the detection feature map and the reference feature map respectively using a convolution encoding portion of the spatial attention module to obtain a detection convolution feature map and a reference convolution feature map; a spatial attention diagram generating unit 142, configured to input the detection convolution feature diagram and the reference convolution feature diagram into a spatial attention portion of the spatial attention module respectively to obtain a detection spatial attention diagram and a reference spatial attention diagram; a spatial attention feature map generation unit 143 configured to pass the detection spatial attention map and the reference spatial attention map through a Softmax activation function to obtain a detection spatial attention feature map and a reference spatial attention feature map, respectively; and a calculating unit 144, configured to calculate a point-by-point multiplication of the detection spatial attention feature map and the detection feature map to obtain the enhanced detection feature map, and calculate a point-by-point multiplication of the reference spatial attention feature map and the reference feature map to obtain the enhanced reference feature map.
More specifically, in the embodiment of the present application, the difference module 150 is configured to calculate a difference feature map between the enhanced detection feature map and the enhanced reference feature map. And calculating a difference feature map between the enhanced detection feature map and the enhanced reference feature map so as to represent a high-dimensional feature expression of the difference between the detection image of the bottle to be detected and the reference image of the reference bottle.
Accordingly, in a specific example, the difference module 150 is further configured to: calculating the differential feature map between the enhanced detection feature map and the enhanced reference feature map in the following formula; wherein the formula is: wherein the enhanced detection feature map is represented, the enhanced reference feature map is represented, the difference feature map is represented, and the difference by position is represented.
More specifically, in the embodiment of the present application, the feature differentiation enhancement module 160 is configured to perform feature value differentiation optimization on the differential feature map to obtain an optimized differential feature map.
In particular, in the technical solution of the present application, when the differential feature map between the enhanced detection feature map and the enhanced reference feature map is calculated, since the enhanced detection feature map and the enhanced reference feature map are each obtained by enhancing a feature space distribution through a spatial attention module, feature values of partial spatial positions in the differential feature map have a more significant importance with respect to other feature values, and if the feature values can be effectively distinguished during classification, it is obvious that a training speed of a classifier and an accuracy of a classification result can be improved. The applicant of the present application thus has recorded, for example, said differential characteristic map
Interaction reinforcement based on distinguishable physical stimuli is performed.
Accordingly, in a specific example, the feature differentiation enhancing module 160 is further configured to: carrying out characteristic value discrimination optimization on the differential characteristic diagram by using the following formula to obtain the optimized differential characteristic diagram; wherein the formula is:
wherein F represents the differential characteristic diagram,
represents the optimized difference characteristic map, a and b are predetermined hyper-parameters, ->
Represents a subtraction by position of the characteristic map, based on the position>
Divide each location of the representation by a response value and +>
Representing the convolution operation through a single convolutional layer.
Here, the discriminative physical excitation-based interaction reinforcement is used to promote the interaction between the feature space and the solution space of the classification problem in the back propagation process through gradient descent, which extracts and imitates the feasible feature (actionable feature) in a manner similar to the physical excitation, thereby obtaining the physical expression of the feasible feature with gradient discriminative using the general-purpose low-dimensional conductive physical excitation manner, and thus reinforcing the active part in the differential feature map F in the training process to promote the training speed of the classifier and the accuracy of the classification result of the trained classification feature.
More specifically, in the embodiment of the present application, the detection result generating module 170 is configured to pass the optimized differential feature map through a classifier to obtain a classification result, where the classification result is used to indicate whether a bottle mouth defect exists in a bottle to be detected. That is, the classifier is used to determine class probability labels to which the differential feature map belongs, wherein the class probability labels include bottle mouth defects (first labels) of the bottles to be detected, and bottle mouth defects (second labels) of the bottles to be detected.
Accordingly, in a specific example, as shown in fig. 4, the detection result generating module 170 includes: a feature map expanding unit 171, configured to expand the optimized differential feature map into classification feature vectors according to row vectors or column vectors; a full-concatenation encoding unit 172, configured to perform full-concatenation encoding on the classification feature vector using a plurality of full-concatenation layers of the classifier to obtain an encoded classification feature vector; and a classification unit 173 for passing the encoded classification feature vector through a Softmax classification function of the classifier to obtain the classification result.
In summary, the system 100 for detecting defects of bottle openings by using a strip-shaped light source according to the embodiment of the present application is illustrated, and first a detection image of a bottle to be detected and a reference image of a reference bottle, which are acquired by a camera, are respectively processed by an image overexposure corrector to obtain a corrected detection image and a corrected reference image, then the corrected detection image and the corrected reference image are processed by a dual detection model to obtain a detection feature map and a reference feature map, then the detection feature map and the reference feature map are processed by a spatial attention module to obtain an enhanced detection feature map and an enhanced reference feature map, then a difference feature map between the enhanced detection feature map and the enhanced reference feature map is calculated, and finally, after feature value discrimination optimization is performed on the difference feature map, a classifier is used to obtain a classification result indicating whether the bottle to be detected has defects or not. By the mode, whether the bottle to be detected has the bottle opening defect can be judged.
As described above, the system 100 for detecting bottle mouth flaws by using a bar light source according to the embodiment of the present application can be implemented in various terminal devices, such as a server with an algorithm for detecting bottle mouth flaws by using a bar light source. In one example, the system 100 for detecting defects in a bottle opening with a bar light source can be integrated into a terminal device as a software module and/or a hardware module. For example, the system 100 for detecting bottle mouth defects by the bar light source may be a software module in the operating system of the terminal device, or may be an application program developed for the terminal device; of course, the system 100 for detecting bottle mouth defects by using the bar light source can also be one of a plurality of hardware modules of the terminal equipment.
Alternatively, in another example, the system 100 for detecting bottle mouth flaws by the bar light source and the terminal device may also be separate devices, and the system 100 for detecting bottle mouth flaws by the bar light source may be connected to the terminal device through a wired and/or wireless network and transmit interactive information according to an agreed data format.
Exemplary method
FIG. 5 is a flowchart of a method for detecting defects on a bottle opening by using a bar light source according to an embodiment of the present application. As shown in fig. 5, a method for detecting a bottle mouth defect by using a strip light source according to an embodiment of the present application includes: s110, acquiring a detection image of a bottle to be detected and a reference image of a reference bottle, which are acquired by a camera, wherein the reference bottle is a bottle without a bottle mouth defect; s120, respectively enabling the detection image and the reference image to pass through an image overexposure corrector based on an automatic coder-decoder to obtain a corrected detection image and a corrected reference image; s130, passing the corrected detection image and the corrected reference image through a dual detection model comprising a first image encoder and a second image encoder to obtain a detection characteristic diagram and a reference characteristic diagram, wherein the first image encoder and the second image encoder have the same network structure; s140, passing the detection feature map and the reference feature map through a spatial attention module to obtain an enhanced detection feature map and an enhanced reference feature map; s150, calculating a difference feature map between the enhanced detection feature map and the enhanced reference feature map; s160, carrying out characteristic value discrimination optimization on the differential characteristic diagram to obtain an optimized differential characteristic diagram; and S170, passing the optimized differential characteristic diagram through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the bottle mouth of the bottle to be detected has a bottle mouth defect.
FIG. 6 is a schematic diagram of a system architecture of a method for detecting a bottle opening defect by using a bar light source according to an embodiment of the present disclosure. As shown in fig. 6, in the system architecture of the method for detecting a bottle mouth defect by using a bar light source, first, a detection image of a bottle to be detected and a reference image of a reference bottle, which are acquired by a camera, are acquired, wherein the reference bottle is a bottle without a bottle mouth defect; then, the detection image and the reference image respectively pass through an image overexposure corrector based on an automatic coder and a decoder to obtain a corrected detection image and a corrected reference image; then, the corrected detection image and the corrected reference image pass through a double detection model comprising a first image encoder and a second image encoder to obtain a detection characteristic diagram and a reference characteristic diagram, wherein the first image encoder and the second image encoder have the same network structure; then, the detection feature map and the reference feature map are processed by a spatial attention module to obtain an enhanced detection feature map and an enhanced reference feature map; then, calculating a difference characteristic diagram between the enhanced detection characteristic diagram and the enhanced reference characteristic diagram; then, carrying out characteristic value discrimination optimization on the differential characteristic diagram to obtain an optimized differential characteristic diagram; and finally, passing the optimized differential characteristic diagram through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the bottle to be detected has a bottle mouth defect.
In a specific example, in the method for detecting a bottle mouth defect by using a bar light source, the passing the detection image and the reference image through an automatic codec-based image overexposure corrector to obtain a corrected detection image and a corrected reference image respectively includes: respectively carrying out image coding on the detection image and the reference image by using an encoder of the automatic codec-based image overexposure corrector to obtain a detection image characteristic representation and a reference image characteristic representation; and decoding and regressing each image feature representation in the detection image feature representation and each image feature representation in the reference image feature representation respectively by using a decoder of the automatic codec-based image overexposure corrector to obtain the corrected detection image and the corrected reference image.
In a specific example, in the method for detecting bottle mouth flaws by using bar light sources, the first image encoder and the second image encoder are depth residual error network models.
In a specific example, in the method for detecting defects of a bottle mouth by using a bar light source, the passing the detection feature map and the reference feature map through a spatial attention module to obtain an enhanced detection feature map and an enhanced reference feature map includes: depth convolution coding is carried out on the detection feature map and the reference feature map respectively by using a convolution coding part of the spatial attention module so as to obtain a detection convolution feature map and a reference convolution feature map; inputting the detection convolution feature map and the reference convolution feature map into a spatial attention portion of the spatial attention module respectively to obtain a detection spatial attention map and a reference spatial attention map; respectively activating functions of the detection space attention diagram and the reference space attention diagram through Softmax to obtain a detection space attention feature map and a reference space attention feature map; and calculating the position-based multiplication of the detection space attention feature map and the detection feature map to obtain the enhanced detection feature map, and calculating the position-based multiplication of the reference space attention feature map and the reference feature map to obtain the enhanced reference feature map.
In a specific example, in the method for detecting a bottle mouth defect by using a bar-shaped light source, the calculating a difference characteristic diagram between the enhanced detection characteristic diagram and the enhanced reference characteristic diagram includes: calculating the differential feature map between the enhanced detection feature map and the enhanced reference feature map in the following formula; wherein the formula is:
in which>
Represents the enhanced detection feature map, based on the feature map, and>
represents the enhanced reference feature map>
Represents the differential characteristic map, and>
indicating a difference by position.
In a specific example, in the method for detecting a bottle mouth defect by using a bar light source, performing feature value discrimination optimization on the differential feature map to obtain an optimized differential feature map includes: carrying out characteristic value discrimination optimization on the differential characteristic diagram by using the following formula to obtain the optimized differential characteristic diagram; wherein the formula is:
wherein the content of the first and second substances,
represents the differential feature map, is based on the characteristic map, and is based on the characteristic map>
Represents the optimized difference feature map>
And &>
Is a predetermined hyper-parameter which is,
by position subtraction, division, representing feature mapsIs divided by the response value and->
Representing the convolution operation through a single convolutional layer.
In a specific example, in the method for detecting a bottle mouth defect by using a bar light source, the passing the optimized differential feature map through a classifier to obtain a classification result, where the classification result is used to indicate whether a bottle mouth defect exists in a bottle to be detected, includes: expanding the optimized differential feature map into classified feature vectors according to row vectors or column vectors; performing full-join encoding on the classification feature vectors using a plurality of full-join layers of the classifier to obtain encoded classification feature vectors; and passing the encoding classification feature vector through a Softmax classification function of the classifier to obtain the classification result.
Here, it can be understood by those skilled in the art that the detailed operations of the steps of the method for detecting a defective bottle mouth by using the bar light source described above have been described in detail in the description of the system for detecting a defective bottle mouth by using the bar light source in fig. 1 to 4, and thus, a repeated description thereof will be omitted.
The foregoing describes the general principles of the present application in conjunction with specific embodiments, however, it is noted that the advantages, effects, etc. mentioned in the present application are merely examples and are not limiting, and they should not be considered essential to the various embodiments of the present application. Furthermore, the foregoing disclosure of specific details is for the purpose of illustration and description and is not intended to be limiting, since the foregoing disclosure is not intended to be exhaustive or to limit the disclosure to the precise details disclosed.
The block diagrams of devices, apparatuses, systems referred to in this application are only given as illustrative examples and are not intended to require or imply that the connections, arrangements, configurations, etc. must be made in the manner shown in the block diagrams. These devices, apparatuses, devices, systems may be connected, arranged, configured in any manner, as will be appreciated by those skilled in the art. Words such as "including," "comprising," "having," and the like are open-ended words that mean "including, but not limited to," and are used interchangeably therewith. As used herein, the words "or" and "refer to, and are used interchangeably with, the word" and/or, "unless the context clearly dictates otherwise. The word "such as" is used herein to mean, and is used interchangeably with, the phrase "such as but not limited to".
It should also be noted that the components or steps of the apparatus, devices and methods of the present application may be disassembled and/or reassembled. These decompositions and/or recombinations are to be considered as equivalents of the present application.
The previous description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
The foregoing description has been presented for purposes of illustration and description. Furthermore, the description is not intended to limit embodiments of the application to the form disclosed herein. While a number of example aspects and embodiments have been discussed above, those of skill in the art will recognize certain variations, modifications, alterations, additions and sub-combinations thereof.
Claims (9)
1. A system for detecting bottle opening flaws by using a bar-shaped light source is characterized by comprising:
the camera module is used for acquiring a detection image of a bottle to be detected and a reference image of a reference bottle, which are acquired by a camera, wherein the reference bottle is a bottle without a bottle mouth defect;
the overexposure correction module is used for enabling the detection image and the reference image to respectively pass through an image overexposure corrector based on an automatic coder-decoder so as to obtain a corrected detection image and a corrected reference image;
a double detection module, configured to pass the corrected detection image and the corrected reference image through a double detection model including a first image encoder and a second image encoder to obtain a detection feature map and a reference feature map, where the first image encoder and the second image encoder have the same network structure;
the spatial enhancement module is used for enabling the detection feature map and the reference feature map to pass through a spatial attention module to obtain an enhanced detection feature map and an enhanced reference feature map;
a difference module for calculating a difference feature map between the enhanced detection feature map and the enhanced reference feature map;
the characteristic discrimination enhancement module is used for optimizing the characteristic value discrimination of the differential characteristic graph to obtain an optimized differential characteristic graph; and
the detection result generation module is used for enabling the optimized differential characteristic diagram to pass through a classifier to obtain a classification result, and the classification result is used for indicating whether the bottle to be detected has a bottle mouth defect;
the feature discrimination enhancement module is further configured to: carrying out characteristic value discrimination optimization on the differential characteristic diagram by using the following formula to obtain the optimized differential characteristic diagram;
wherein the formula is:
wherein the content of the first and second substances,
represents the differential characteristic map, and>
represents the optimized difference feature map>
And &>
Is a predetermined hyper-parameter which is,
represents a subtraction by position of the feature map, the division representing each position of the feature map divided by a response value, and->
Representing the convolution operation through a single convolutional layer.
2. The system for detecting defects of bottle openings by using a bar light source according to claim 1, wherein the overexposure correction module is further configured to:
use of the sameRespectively carrying out image coding on the detection image and the reference image by an encoder of an image overexposure corrector based on an automatic codec so as to obtain a detection image characteristic representation and a reference image characteristic representation; and
and respectively performing decoding regression on each image feature representation in the detection image feature representation and each image feature representation in the reference image feature representation by using a decoder of the automatic codec-based image overexposure corrector to obtain the corrected detection image and the corrected reference image.
3. The system for detecting bottle mouth flaws with a bar light source as claimed in claim 2, wherein the first image encoder and the second image encoder are depth residual error network models.
4. The system for detecting bottle mouth flaws by using a bar light source according to claim 3, wherein the space enhancement module comprises:
a depth convolution unit, configured to perform depth convolution encoding on the detection feature map and the reference feature map respectively using a convolution encoding portion of the spatial attention module to obtain a detection convolution feature map and a reference convolution feature map;
a spatial attention diagram generation unit, configured to input the detection convolution feature map and the reference convolution feature map into a spatial attention portion of the spatial attention module respectively to obtain a detection spatial attention diagram and a reference spatial attention diagram;
a spatial attention feature map generation unit, configured to pass the detection spatial attention map and the reference spatial attention map through a Softmax activation function to obtain a detection spatial attention feature map and a reference spatial attention feature map, respectively; and
and the calculation unit is used for calculating the position-based multiplication of the detection space attention feature map and the detection feature map to obtain the enhanced detection feature map, and calculating the position-based multiplication of the reference space attention feature map and the reference feature map to obtain the enhanced reference feature map.
5. The system for detecting defects of bottle openings by using a bar light source as claimed in claim 4, wherein the difference module is further used for: calculating the differential feature map between the enhanced detection feature map and the enhanced reference feature map in the following formula;
wherein the formula is:
wherein is present>
Representing a map of said enhanced detection features, device for selecting or keeping>
Represents the enhanced reference feature map>
Represents the differential characteristic map, and>
indicating a difference by position.
6. The system for detecting bottle mouth flaws by using the bar-type light source according to claim 5, wherein the detection result generation module comprises:
the characteristic map expanding unit is used for expanding the optimized differential characteristic map into classified characteristic vectors according to row vectors or column vectors;
a full-concatenation encoding unit, configured to perform full-concatenation encoding on the classification feature vector using a plurality of full-concatenation layers of the classifier to obtain an encoded classification feature vector; and
and the classification unit is used for enabling the coded classification feature vector to pass through a Softmax classification function of the classifier so as to obtain the classification result.
7. A method for detecting bottle mouth flaws by using a strip-shaped light source is characterized by comprising the following steps:
acquiring a detection image of a bottle to be detected and a reference image of a reference bottle, which are acquired by a camera, wherein the reference bottle is a bottle without a bottle mouth defect;
respectively passing the detection image and the reference image through an image overexposure corrector based on an automatic coder-decoder to obtain a corrected detection image and a corrected reference image;
enabling the corrected detection image and the corrected reference image to pass through a double detection model comprising a first image encoder and a second image encoder to obtain a detection characteristic diagram and a reference characteristic diagram, wherein the first image encoder and the second image encoder have the same network structure;
passing the detection feature map and the reference feature map through a spatial attention module to obtain an enhanced detection feature map and an enhanced reference feature map;
calculating a difference feature map between the enhanced detection feature map and the enhanced reference feature map;
carrying out characteristic value discrimination optimization on the differential characteristic diagram to obtain an optimized differential characteristic diagram; and
the optimized differential characteristic diagram is processed by a classifier to obtain a classification result, and the classification result is used for indicating whether the bottle to be detected has a bottle mouth defect;
carrying out characteristic value discrimination optimization on the differential characteristic diagram by using the following formula to obtain the optimized differential characteristic diagram;
wherein the formula is:
wherein the content of the first and second substances,
represents the differential feature map, is based on the characteristic map, and is based on the characteristic map>
Represents the optimized difference feature map>
And &>
The predetermined hyper-parameter is a function of,
representing a subtraction by position of the feature map, division representing each position of the feature map divided by a response value, and +>
Representing the convolution operation through a single convolutional layer.
8. The method for detecting bottle mouth flaws by using a strip light source as claimed in claim 7, wherein the step of passing the detection image and the reference image through an automatic codec-based image overexposure corrector to obtain a corrected detection image and a corrected reference image respectively comprises the following steps:
respectively carrying out image coding on the detection image and the reference image by using an encoder of the automatic codec-based image overexposure corrector to obtain a detection image characteristic representation and a reference image characteristic representation; and
and respectively performing decoding regression on each image feature representation in the detection image feature representation and each image feature representation in the reference image feature representation by using a decoder of the automatic codec-based image overexposure corrector to obtain the corrected detection image and the corrected reference image.
9. The method of claim 8, wherein the first image encoder and the second image encoder are depth residual network models.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310126946.1A CN115901794B (en) | 2023-02-17 | 2023-02-17 | System and method for detecting bottleneck flaws by using strip-shaped light source |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310126946.1A CN115901794B (en) | 2023-02-17 | 2023-02-17 | System and method for detecting bottleneck flaws by using strip-shaped light source |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115901794A true CN115901794A (en) | 2023-04-04 |
| CN115901794B CN115901794B (en) | 2023-06-23 |
Family
ID=86489926
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310126946.1A Active CN115901794B (en) | 2023-02-17 | 2023-02-17 | System and method for detecting bottleneck flaws by using strip-shaped light source |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115901794B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116309446A (en) * | 2023-03-14 | 2023-06-23 | 浙江固驰电子有限公司 | Method and system for manufacturing power module for industrial control field |
| CN116580029A (en) * | 2023-07-12 | 2023-08-11 | 浙江海威汽车零件有限公司 | Quality inspection control system and method for aluminum alloy casting finished product |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112581392A (en) * | 2020-12-15 | 2021-03-30 | 中山大学 | Image exposure correction method, system and storage medium based on bidirectional illumination estimation and fusion restoration |
| CN112767342A (en) * | 2021-01-14 | 2021-05-07 | 成都缇娣莉科技有限公司 | Intelligent gas detection method based on double-branch inference mechanism |
| CN115239515A (en) * | 2022-07-28 | 2022-10-25 | 德玛克(长兴)精密机械有限公司 | Precise intelligent processing and manufacturing system for mechanical parts and manufacturing method thereof |
| CN115375691A (en) * | 2022-10-26 | 2022-11-22 | 济宁九德半导体科技有限公司 | Image-based semiconductor diffusion paper source defect detection system and method thereof |
| CN115526865A (en) * | 2022-09-30 | 2022-12-27 | 深圳市创瑞鑫科技有限公司 | Insulation testing method and system for heat dissipation module of notebook computer |
| CN115561243A (en) * | 2022-09-30 | 2023-01-03 | 东莞市言科新能源有限公司 | Pole piece quality monitoring system and method in lithium battery preparation |
-
2023
- 2023-02-17 CN CN202310126946.1A patent/CN115901794B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112581392A (en) * | 2020-12-15 | 2021-03-30 | 中山大学 | Image exposure correction method, system and storage medium based on bidirectional illumination estimation and fusion restoration |
| CN112767342A (en) * | 2021-01-14 | 2021-05-07 | 成都缇娣莉科技有限公司 | Intelligent gas detection method based on double-branch inference mechanism |
| CN115239515A (en) * | 2022-07-28 | 2022-10-25 | 德玛克(长兴)精密机械有限公司 | Precise intelligent processing and manufacturing system for mechanical parts and manufacturing method thereof |
| CN115526865A (en) * | 2022-09-30 | 2022-12-27 | 深圳市创瑞鑫科技有限公司 | Insulation testing method and system for heat dissipation module of notebook computer |
| CN115561243A (en) * | 2022-09-30 | 2023-01-03 | 东莞市言科新能源有限公司 | Pole piece quality monitoring system and method in lithium battery preparation |
| CN115375691A (en) * | 2022-10-26 | 2022-11-22 | 济宁九德半导体科技有限公司 | Image-based semiconductor diffusion paper source defect detection system and method thereof |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116309446A (en) * | 2023-03-14 | 2023-06-23 | 浙江固驰电子有限公司 | Method and system for manufacturing power module for industrial control field |
| CN116580029A (en) * | 2023-07-12 | 2023-08-11 | 浙江海威汽车零件有限公司 | Quality inspection control system and method for aluminum alloy casting finished product |
| CN116580029B (en) * | 2023-07-12 | 2023-10-13 | 浙江海威汽车零件有限公司 | Quality inspection control system and method for aluminum alloy casting finished product |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115901794B (en) | 2023-06-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115901794A (en) | System and method for detecting bottle opening flaws through strip-shaped light source | |
| Liu et al. | Adaptive early-learning correction for segmentation from noisy annotations | |
| Wang et al. | Tire defect detection using fully convolutional network | |
| US20210174149A1 (en) | Feature fusion and dense connection-based method for infrared plane object detection | |
| CN116030048B (en) | Lamp inspection machine and method thereof | |
| CN111079563A (en) | Traffic signal lamp identification method and device, electronic equipment and storage medium | |
| JP2011214903A (en) | Appearance inspection apparatus, and apparatus, method and program for generating appearance inspection discriminator | |
| CN111784673A (en) | Defect detection model training and defect detection method, device and storage medium | |
| Zipfel et al. | Anomaly detection for industrial quality assurance: A comparative evaluation of unsupervised deep learning models | |
| CN114037674B (en) | Industrial defect image segmentation detection method and device based on semantic context | |
| Park et al. | MarsNet: multi-label classification network for images of various sizes | |
| CN112906794A (en) | Target detection method, device, storage medium and terminal | |
| CN117036271A (en) | Production line quality monitoring method and system thereof | |
| CN117011274A (en) | Automatic glass bottle detection system and method thereof | |
| CN109034121B (en) | Commodity identification processing method, device, equipment and computer storage medium | |
| CN113468946A (en) | Semantically consistent enhanced training data for traffic light detection | |
| CN111210417B (en) | Cloth defect detection method based on convolutional neural network | |
| Feng et al. | Learning an invariant and equivariant network for weakly supervised object detection | |
| Shit et al. | An encoder‐decoder based CNN architecture using end to end dehaze and detection network for proper image visualization and detection | |
| CN117173154A (en) | Online image detection system and method for glass bottle | |
| CN116994049A (en) | Full-automatic flat knitting machine and method thereof | |
| CN117274689A (en) | Detection method and system for detecting defects of packaging box | |
| CN115908978A (en) | Defect sample simulation method, system, computer and readable storage medium | |
| CN116205881A (en) | Digital jet printing image defect detection method based on lightweight semantic segmentation | |
| CN113269137B (en) | Non-matching face recognition method combining PCANet and shielding positioning |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |