CN113344862B - Defect detection method, device, electronic equipment and storage medium - Google Patents
Defect detection method, device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN113344862B CN113344862B CN202110554117.4A CN202110554117A CN113344862B CN 113344862 B CN113344862 B CN 113344862B CN 202110554117 A CN202110554117 A CN 202110554117A CN 113344862 B CN113344862 B CN 113344862B
- Authority
- CN
- China
- Prior art keywords
- feature
- scaling
- pyramid network
- feature pyramid
- channel
- 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.)
- Active
Links
- 230000007547 defect Effects 0.000 title claims abstract description 112
- 238000001514 detection method Methods 0.000 title claims abstract description 55
- 238000012549 training Methods 0.000 claims abstract description 27
- 238000000034 method Methods 0.000 claims abstract description 20
- 230000004927 fusion Effects 0.000 claims description 13
- 238000012545 processing Methods 0.000 claims description 13
- 238000004590 computer program Methods 0.000 claims description 10
- 238000002372 labelling Methods 0.000 claims description 10
- 238000003384 imaging method Methods 0.000 claims description 9
- 238000007781 pre-processing Methods 0.000 claims description 4
- 230000002950 deficient Effects 0.000 claims 1
- 238000013473 artificial intelligence Methods 0.000 abstract description 9
- 238000013135 deep learning Methods 0.000 abstract description 5
- 238000004891 communication Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000003491 array Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformation in the plane of the image
- G06T3/40—Scaling the whole image or part thereof
- G06T3/4046—Scaling the whole image or part thereof using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
Abstract
The disclosure provides a defect detection method, a defect detection device, electronic equipment and a storage medium, relates to the technical field of computers, and particularly relates to the technical field of artificial intelligence such as computer vision and deep learning. The specific implementation scheme is as follows: the method comprises the steps of obtaining a first feature pyramid network corresponding to a product image and a second feature pyramid network corresponding to a template image, respectively scaling each feature layer in the first feature pyramid network and each feature layer in the second feature pyramid network based on first scaling parameters and second scaling parameters generated through training to obtain the first scaling feature pyramid network and the second scaling feature pyramid network, respectively fusing each feature layer in the first scaling feature pyramid network with a corresponding feature layer in the second scaling feature pyramid network, and finally identifying each feature layer in the fused feature pyramid network to determine whether defects exist in products in the product image. Thereby, the accuracy of defect detection is improved.
Description
Technical Field
The disclosure relates to the technical field of computers, in particular to the technical field of artificial intelligence such as computer vision and deep learning, and specifically relates to a defect detection method, a defect detection device, electronic equipment and a storage medium.
Background
As artificial intelligence technology continues to develop and improve, it has played an extremely important role in various fields related to human daily life, for example, artificial intelligence has made significant progress in the application scenario of defect detection. Therefore, improving the accuracy of product defect detection is becoming a research direction of hot spots.
Disclosure of Invention
The present disclosure provides a defect detection method, a defect detection device, an electronic device, and a storage medium.
According to a first aspect of the present disclosure, there is provided a defect detection method including:
acquiring a product image to be detected and a template image;
inputting the product image and the template image into a twin network to determine a first feature pyramid network corresponding to the product image and a second feature pyramid network corresponding to the template image;
scaling each feature layer in the first feature pyramid network and each feature layer in the second feature pyramid network based on a first scaling parameter and a second scaling parameter generated by training respectively to obtain a first scaled feature pyramid network and a second scaled feature pyramid network;
Respectively fusing each feature layer in the first scaled feature pyramid network with a corresponding feature layer in the second scaled feature pyramid network to generate a fused feature pyramid network;
and respectively identifying each feature layer in the fused feature pyramid network to determine whether the product in the product image has defects.
According to a second aspect of the present disclosure, there is provided a defect detection apparatus including:
the first acquisition module is used for acquiring a product image to be detected and a template image;
the first determining module is used for inputting the product image and the template image into a twin network to determine a first characteristic pyramid network corresponding to the product image and a second characteristic pyramid network corresponding to the template image;
the second obtaining module is used for respectively carrying out scaling processing on each feature layer in the first feature pyramid network and each feature layer in the second feature pyramid network based on the first scaling parameter and the second scaling parameter generated by training so as to obtain the first scaling feature pyramid network and the second scaling feature pyramid network;
the first generation module is used for respectively fusing each feature layer in the first scaling feature pyramid network with the corresponding feature layer in the second scaling feature pyramid network to generate a fused feature pyramid network;
And the second determining module is used for respectively identifying each characteristic layer in the fused characteristic pyramid network so as to determine whether the product in the product image has defects.
According to a third aspect of the present disclosure, there is provided an electronic device comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the defect detection method according to the first aspect.
According to a fourth aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions for causing the computer to perform the defect detection method according to the first aspect.
According to a fifth aspect of the present disclosure, there is provided a computer program product comprising computer instructions which, when executed by a processor, implement the steps of the defect detection method as described in the first aspect.
The defect detection method, the defect detection device, the electronic equipment and the storage medium have the following beneficial effects:
Firstly, inputting an acquired product image to be detected and a template image into a twin network to determine a first feature pyramid network corresponding to the product image and a second feature pyramid network corresponding to the template image, then respectively scaling each feature layer in the first feature pyramid network and each feature layer in the second feature pyramid network based on a first scaling parameter and a second scaling parameter generated by training, respectively fusing each feature layer in the first scaling feature pyramid network and a corresponding feature layer in the second scaling feature pyramid network to generate a fused feature pyramid network, and finally respectively identifying each feature layer in the fused feature pyramid network to determine whether a product in the product image has defects. Therefore, each feature layer in the first feature pyramid network corresponding to the product image and the second feature pyramid network corresponding to the template image is subjected to scaling treatment respectively, and then the feature layers subjected to scaling treatment are subjected to feature fusion, so that the determined defect type and position can be more accurate, and the defect detection accuracy is improved.
It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the disclosure, nor is it intended to be used to limit the scope of the disclosure. Other features of the present disclosure will become apparent from the following specification.
Drawings
The drawings are for a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein:
FIG. 1 is a flow chart of a defect detection method according to an embodiment of the present disclosure;
FIG. 2 is a flow chart of a defect detection method according to another embodiment of the present disclosure;
FIG. 3 is a flow chart of a defect detection method according to another embodiment of the present disclosure;
FIG. 4 is a flow chart of a defect detection method according to another embodiment of the present disclosure;
FIG. 5 is a schematic diagram of a defect detection apparatus according to an embodiment of the present disclosure;
fig. 6 is a block diagram of an electronic device for implementing a defect detection method of an embodiment of the present disclosure.
Detailed Description
Exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding, and should be considered as merely exemplary. Accordingly, one of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.
The embodiment of the disclosure relates to the technical field of artificial intelligence such as computer vision, deep learning and the like.
Artificial intelligence (Artificial Intelligence), english is abbreviated AI. It is a new technical science for researching, developing theory, method, technology and application system for simulating, extending and expanding human intelligence.
Deep learning is the inherent regularity and presentation hierarchy of learning sample data, and the information obtained during such learning is helpful in interpreting data such as text, images and sounds. The final goal of deep learning is to enable a machine to analyze learning capabilities like a person, and to recognize text, images, and sound data.
Computer vision refers to machine vision such as identifying, tracking and measuring targets by using a camera and a computer instead of human eyes, and further performing graphic processing, so that the computer processing becomes an image which is more suitable for human eyes to observe or transmit to an instrument for detection.
Fig. 1 is a flow chart of a defect detection method according to an embodiment of the disclosure.
It should be noted that, the main execution body of the defect detection method in this embodiment is a defect detection device, and the device may be implemented in a software and/or hardware manner, and the device may be configured in an electronic device, where the electronic device may include, but is not limited to, a terminal, a server, and the like.
As shown in fig. 1, the defect detection method includes:
s101: and acquiring a product image to be detected and a template image.
The product image to be detected is an image of a product which is acquired by adopting image acquisition equipment and needs to be subjected to defect detection. The template image is an image of a product without defects, which is acquired by adopting image acquisition equipment, and can be used as a template to detect whether other products have defects.
The image capturing device includes a camera, a video camera, a scanner, and other devices with photographing function (such as a mobile phone, a tablet computer, etc.), which is not limited in this disclosure.
S102: inputting the product image and the template image into a twin network to determine a first feature pyramid network corresponding to the product image and a second feature pyramid network corresponding to the template image.
The twin network consists of two sub-networks, and the product image and the template image can be respectively input into different sub-networks for feature extraction so as to obtain a plurality of product feature layers with different scales corresponding to the product image and a plurality of template feature layers with different scales corresponding to the template image. And determining a first characteristic pyramid network corresponding to the product image and a second characteristic pyramid network corresponding to the template image.
It should be noted that, the two sub-networks in the twin network may be convolutional neural networks or cyclic neural networks, and the network structures of the two sub-networks may be the same or different, which is not limited in this disclosure.
The first feature pyramid network may be a feature pyramid network constructed by extracting n product feature layers from a plurality of product feature layers; the second feature pyramid network may be a feature pyramid network constructed to extract n template feature layers from a plurality of template feature layers.
It should be noted that the foregoing examples are only illustrative, and should not be taken as limiting the first feature pyramid network and the second feature pyramid network in the embodiments of the present disclosure.
In the embodiment of the disclosure, the product image may be preprocessed before being input into the twin network, so that an imaging angle of the product in the preprocessed product image is the same as an imaging angle of the product in the template image. Wherein the preprocessing may include rotating, translating, scaling, etc. the product image, which is not limited by the present disclosure.
S103: and scaling each feature layer in the first feature pyramid network and each feature layer in the second feature pyramid network based on the first scaling parameter and the second scaling parameter generated by training respectively to acquire the first scaling feature pyramid network and the second scaling feature pyramid network.
The first scaling parameters and the second scaling parameters are parameters which are generated in advance in a training mode and used for scaling the feature layers in the feature pyramid network.
The first scaling parameters are used for scaling each feature layer in the first feature pyramid network so as to obtain the first scaled feature pyramid network; and the second scaling parameter is used for scaling each feature layer in the second feature pyramid network to obtain the second scaled feature pyramid network.
S104: and respectively fusing each feature layer in the first scaled feature pyramid network with a corresponding feature layer in the second scaled feature pyramid network to generate a fused feature pyramid network.
Alternatively, the features in each feature layer in the first scaled feature pyramid network may be added to the features in the corresponding feature layer in the second scaled feature pyramid network to generate a fused feature pyramid network.
Optionally, the features in each feature layer in the first scaled feature pyramid network may also be spliced with the features in the corresponding feature layer in the second scaled feature pyramid network to generate a fused feature pyramid network, and the like, which is not limited in this disclosure.
It should be noted that the foregoing examples are merely illustrative, and are not meant to be limiting of the fused feature pyramid network in the embodiments of the present disclosure.
In the embodiment of the disclosure, each feature layer in the first feature pyramid network corresponding to the product image and the second feature pyramid network corresponding to the template image is respectively subjected to scaling treatment, and then the feature layers subjected to scaling treatment are subjected to feature fusion, so that the defect type and the defect position determined based on the fused feature pyramid network can be more accurate.
S105: and respectively identifying each feature layer in the fused feature pyramid network to determine whether the product in the product image has defects.
Specifically, each feature layer in the fused feature pyramid network is identified, an identification result of each feature layer is obtained, and whether a product in the product image has a defect can be determined by analyzing the identification result of each feature layer.
In this embodiment, an acquired product image to be detected and a template image are input into a twin network to determine a first feature pyramid network corresponding to the product image and a second feature pyramid network corresponding to the template image, then scaling each feature layer in the first feature pyramid network and each feature layer in the second feature pyramid network based on a first scaling parameter and a second scaling parameter generated by training, respectively fusing each feature layer in the first scaling feature pyramid network and a corresponding feature layer in the second scaling feature pyramid network to generate a fused feature pyramid network, and finally identifying each feature layer in the fused feature pyramid network to determine whether a defect exists in a product in the product image. Therefore, each feature layer in the first feature pyramid network corresponding to the product image and the second feature pyramid network corresponding to the template image is subjected to scaling treatment respectively, and then the feature layers subjected to scaling treatment are subjected to feature fusion, so that the determined defect type and position can be more accurate, and the defect detection accuracy is improved.
Fig. 2 is a schematic flow chart of a defect detection method according to another embodiment of the present disclosure, as shown in fig. 2, before inputting a product image and a template image into a twin network, the method may further include the following steps:
s201: and acquiring a training data set, wherein the training data set comprises a template image and a plurality of marked product images.
Wherein the training data set may be acquired using an image acquisition device.
The marked product image is an image obtained by manually marking the defects in the product image acquired by the image acquisition equipment.
S202: and inputting the marked product image and the template image into an initial twin network to determine a third feature pyramid network corresponding to the marked product image and a fourth feature pyramid network corresponding to the template image.
Specifically, before the annotated product image and the template image are input into the initial twinning network, the annotated product image may be preprocessed, so that the imaging angle of the product in the preprocessed annotated product image is the same as the imaging angle of the product in the template image.
The initial twin network is an untrained network, and the initial twin network needs to be trained by adopting a training data set so as to generate a twin network of the feature pyramid network which can be used for accurately extracting the product image and the template image.
S203: and respectively carrying out scaling treatment on each feature layer in the third feature pyramid network and each feature layer in the fourth feature pyramid network based on a preset third scaling parameter and a preset fourth scaling parameter so as to obtain the third scaling feature pyramid network and the fourth scaling feature pyramid network.
The third scaling parameter is a preset parameter for scaling each feature layer in the third feature pyramid network.
The fourth scaling parameter is a preset parameter for scaling each feature layer in the fourth feature pyramid network.
Wherein the scaling process may be to multiply the features in each feature layer with corresponding scaling parameters, which is not limited by the present disclosure.
For example, multiplying each feature layer in the third feature pyramid network by a third scaling parameter obtains the third scaled feature pyramid network. Alternatively, multiplying each feature layer in the fourth feature pyramid network by a fourth scaling parameter to obtain the fourth scaled feature pyramid network.
S204: and respectively fusing each feature layer in the third zooming feature pyramid network with the corresponding feature layer in the fourth zooming feature pyramid network to generate a fused zooming feature pyramid network.
S205: and respectively identifying each feature layer in the fused scaled feature pyramid network to determine a corresponding defect prediction result in the marked product image.
For example, each feature layer in the fused scaled feature pyramid network is identified respectively to determine a corresponding identification result of each feature layer, and a corresponding defect prediction result in the marked product image is determined according to each identification result and the weight value of each feature layer.
S206: and respectively correcting the third scaling parameter, the fourth scaling parameter and the initial twin network according to the difference between the defect prediction result and the defect labeling result corresponding to the labeled product image until the difference between the corresponding defect prediction result and the labeling result in the labeled product image is smaller than a threshold value so as to determine the twin network, the first scaling parameter and the second scaling parameter.
For example, under the condition that the difference between the defect prediction result and the defect labeling result corresponding to the labeled product image is large, respectively adjusting the preset third scaling parameter, the preset fourth scaling parameter and the internal parameters in the initial twin network until the difference between the corresponding defect prediction result and the labeling result in the labeled product image is smaller than a threshold value according to the fused scaling feature pyramid network generated through the twin network and the scaling processing, and finishing training. And then, detecting whether the product in the product image to be detected has defects or not by utilizing the trained twin network, the first scaling parameters and the second scaling parameters.
In this embodiment, training is performed on the initial twin network, the preset third scaling parameter and the preset fourth scaling parameter according to the training data set, so as to determine the twin network, the first scaling parameter and the second scaling parameter. Therefore, the product image and the template image are input into the trained twin network, the first scaling feature pyramid network and the second scaling feature pyramid network are obtained based on the first scaling parameter and the second scaling parameter, and then feature fusion is carried out on the feature layer of each of the first scaling feature pyramid network and the second scaling feature pyramid network, so that the determined defect type and the determined position can be more accurate, and the defect detection accuracy is improved.
In one possible implementation of the present disclosure, the feature pyramid network output by the twin network includes at least one channel, and then each channel feature in the first feature pyramid network may be scaled based on the same scaling parameter, or each channel feature in the first feature pyramid network may be scaled based on a different scaling parameter. The process of scaling the feature pyramid network in this disclosure is described in detail below in conjunction with fig. 3 and 4.
Fig. 3 is a flow chart of a defect detection method according to another embodiment of the present disclosure, as shown in fig. 3, where the defect detection method provided by the present disclosure includes:
and 301, acquiring a product image to be detected and a template image.
302, inputting the product image and the template image into a twin network to determine a first feature pyramid network corresponding to the product image and a second feature pyramid network corresponding to the template image, wherein each feature layer in the first feature pyramid network and the second feature pyramid network respectively comprises a plurality of channels.
The specific implementation manners of the above step 301 and step 302 may refer to the detailed descriptions of other embodiments of the disclosure, which are not repeated herein.
S303: and based on the first scaling parameters, respectively scaling the features of each first channel in each feature layer in the first feature pyramid network to obtain the scaling features of each first channel.
Wherein each feature layer in the first feature pyramid network comprises a plurality of channels, each channel in the plurality of channels being collectively referred to as a first channel.
The scaling characteristic of each first channel is obtained by scaling the characteristic of each first channel by adopting the same first scaling parameter.
For example, the feature of each first channel is multiplied by a first scaling parameter to obtain a corresponding scaling feature.
It should be noted that the above examples are only illustrative, and should not be taken as limiting the scaling characteristics of each first channel in the embodiments of the present disclosure.
S304: and based on the second scaling parameters, respectively scaling the features of each second channel in each feature layer in the second feature pyramid network to obtain the scaling features of each second channel.
Wherein each feature layer in the second feature pyramid network comprises a plurality of channels, each channel in the plurality of channels being collectively referred to as a second channel.
The scaling characteristic of each second channel is obtained by scaling the characteristic of each second channel by adopting the same second scaling parameter.
For example, the feature of each second channel is multiplied by a second scaling parameter to obtain a corresponding scaling feature.
It should be noted that the above examples are only illustrative and should not be taken as limiting the scaling characteristics of each second channel in the embodiments of the present disclosure.
S305: the scaled features of each first channel are fused with the scaled features of the corresponding second channel to generate fused features for each channel in each feature layer.
For example, adding the scaled features of each first channel to the scaled features of the corresponding second channel generates a fused feature for each channel in each feature layer.
It should be noted that the above examples are illustrative only and should not be taken as limiting the fusion characteristics of each channel in the embodiments of the present disclosure.
In this embodiment, based on the first scaling parameter, scaling the features of each first channel in each feature layer in the first feature pyramid network to obtain scaled features of each first channel; and based on the second scaling parameters, respectively scaling the features of each second channel in each feature layer in the second feature pyramid network to obtain the scaling features of each second channel, and finally fusing the scaling features of each first channel with the scaling features of the corresponding second channel to generate the fused features of each channel in each feature layer. Therefore, after different channels are scaled by adopting different scaling parameters, the scaling feature pyramid network is fused, so that whether the product in the product image has defects or not is determined according to the fusion features, the accuracy of the determined defect type and position can be further improved, and the accuracy of defect detection is improved.
Fig. 4 is a flowchart of a defect detection method according to another embodiment of the present disclosure, where, as shown in fig. 4, each of the first feature pyramid network and the second feature pyramid network includes at least one channel, the first scaling parameter includes a first sub-scaling parameter corresponding to each channel in each feature layer in the first feature pyramid network, and the second scaling parameter includes a second sub-scaling parameter corresponding to each channel in each feature layer in the second feature pyramid network, then the defect detection method provided in the present disclosure includes:
s401, acquiring a product image to be detected and a template image.
S402, inputting the product image and the template image into a twin network to determine a first feature pyramid network corresponding to the product image and a second feature pyramid network corresponding to the template image, wherein each feature layer in the first feature pyramid network and the second feature pyramid network respectively comprises a plurality of channels.
The specific implementation manners of the steps 401 and 402 may refer to the detailed descriptions of other embodiments of the disclosure, and are not repeated herein.
S403: and scaling the features of the corresponding first channels based on the first sub-scaling parameters corresponding to each first channel in each feature layer in each first feature pyramid network respectively so as to acquire the scaling features of each first channel.
It should be noted that, each first channel in the first feature pyramid network corresponds to one first scaling parameter, and the first scaling parameters corresponding to different first channels may be the same or different.
S404: and scaling the features of each corresponding second channel based on second sub-scaling parameters corresponding to each second channel in each feature layer in each second feature pyramid network respectively so as to acquire the scaling features of each second channel.
It should be noted that, each second channel in the second feature pyramid network corresponds to a second scaling parameter, and the second scaling parameters corresponding to different second channels may be the same or different.
S405: fusing the scaling characteristics of each first channel with the scaling characteristics of the corresponding second channel to generate the fused characteristics of each channel in each characteristic layer, and obtaining the fused characteristic pyramid network.
For example, adding the scaled features of each first channel to the scaled features of each corresponding second channel generates a fused feature for each channel in each feature layer.
It should be noted that the above examples are illustrative only and should not be taken as limiting the fusion characteristics of each channel in the embodiments of the present disclosure.
S406: and identifying each feature layer in the fused feature pyramid network to determine the corresponding identification result of each feature layer.
The identification result may include a type of the defect, a location of the defect, a size of the defect, etc., which is not limited in the present disclosure.
S407: and determining whether the product in the product image has defects according to the identification result corresponding to each feature layer and the weight value of each feature layer.
Wherein the weight value of each feature layer may be the same or different.
For example, the fused feature pyramid network has three feature layers, each feature layer corresponds to the same weight, and the weights corresponding to the three feature layers are all 1/3.
Alternatively, the weight may be assigned to each feature layer according to the number of features included in each feature layer, the importance degree, and the like.
It should be noted that the foregoing examples are only illustrative, and should not be taken as limiting the weight value of each feature layer in the embodiments of the present disclosure.
In this embodiment, each feature layer in the first feature pyramid network and the second feature pyramid network includes a plurality of channels, and the features of each first channel are scaled based on a first sub-scaling parameter corresponding to each first channel in each feature layer in each first feature pyramid network, so as to obtain scaled features of each first channel; scaling the features of the corresponding second channels based on the second sub-scaling parameters corresponding to each second channel in each feature pyramid network respectively to obtain the scaling features of each second channel, finally fusing the scaling features of each first channel with the scaling features of the corresponding second channel to generate fused features of each channel in each feature pyramid network, namely, the fused feature pyramid network can be obtained, each feature layer in the fused feature pyramid network is identified to determine the identification result corresponding to each feature layer, and whether the product in the product image has defects is determined according to the identification result corresponding to each feature layer and the weight value of each feature layer. Therefore, after different channels are scaled by adopting different scaling parameters, the scaled feature pyramid networks are fused to obtain the fused feature pyramid networks, and then each feature layer in the fused feature pyramid networks is respectively identified to determine whether the product in the product image has defects, so that the accuracy of the determined defect type and position can be further improved, and the accuracy of defect detection is improved.
The disclosure also provides a defect detection apparatus for implementing any one of the defect detection methods described above.
Fig. 5 is a schematic structural diagram of a defect detecting device according to an embodiment of the present disclosure, and as shown in fig. 5, the defect detecting device 500 includes:
a first obtaining module 510, configured to obtain a product image and a template image to be detected;
the first determining module 520 is configured to input the product image and the template image into the twin network to determine a first feature pyramid network corresponding to the product image and a second feature pyramid network corresponding to the template image;
the second obtaining module 530 is configured to perform scaling processing on each feature layer in the first feature pyramid network and each feature layer in the second feature pyramid network based on the first scaling parameter and the second scaling parameter generated by training, respectively, so as to obtain the first scaled feature pyramid network and the second scaled feature pyramid network;
a first generating module 540, configured to fuse each feature layer in the first scaled feature pyramid network with a corresponding feature layer in the second scaled feature pyramid network, respectively, so as to generate a fused feature pyramid network;
A second determining module 550, configured to identify each feature layer in the fused feature pyramid network, so as to determine whether a product in the product image has a defect.
In some embodiments of the present disclosure, each feature layer in the first feature pyramid network and the second feature pyramid network includes at least one channel, and the second obtaining module 503 is specifically configured to:
based on the first scaling parameters, respectively scaling the features of each first channel in each feature layer in the first feature pyramid network to obtain scaled features of each first channel;
based on the second scaling parameters, respectively scaling the features of each second channel in each feature layer in the second feature pyramid network to obtain scaled features of each second channel;
the first generating module 540 is specifically configured to fuse the scaling feature of each first channel with the scaling feature of the corresponding second channel to generate a fused feature of each channel in each feature layer.
In some embodiments of the present disclosure, each of the first feature pyramid network and the second feature pyramid network includes at least one channel, the first scaling parameter includes a first sub-scaling parameter corresponding to each channel in each feature layer in the first feature pyramid network, the second scaling parameter includes a second sub-scaling parameter corresponding to each channel in each feature layer in the second feature pyramid network, and the second obtaining module 530 is specifically configured to:
Scaling the features of the corresponding first channels based on first sub-scaling parameters corresponding to each first channel in each feature layer in each first feature pyramid network respectively to obtain scaled features of each first channel;
scaling the features of the corresponding second channels based on second sub-scaling parameters corresponding to each second channel in each feature layer in each second feature pyramid network respectively to obtain scaled features of each second channel;
the first generating module 540 is specifically configured to fuse the scaling feature of each first channel with the scaling feature of the corresponding second channel to generate a fused feature of each channel in each feature layer.
In some embodiments of the present disclosure, the second determining module 550 is specifically configured to:
identifying each feature layer in the fused feature pyramid network to determine the corresponding identification result of each feature layer;
and determining whether the product in the product image has defects according to the identification result corresponding to each feature layer and the weight value of each feature layer.
In some embodiments of the present disclosure, the first determining module 520 is specifically configured to:
Acquiring a training data set, wherein the training data set comprises a template image and a plurality of marked product images;
inputting the marked product image and the template image into an initial twin network to determine a third feature pyramid network corresponding to the marked product image and a fourth feature pyramid network corresponding to the template image;
respectively carrying out scaling treatment on each feature layer in the third feature pyramid network and each feature layer in the fourth feature pyramid network based on a preset third scaling parameter and a preset fourth scaling parameter so as to obtain the third scaling feature pyramid network and the fourth scaling feature pyramid network;
respectively fusing each feature layer in the third scaling feature pyramid network with a corresponding feature layer in the fourth scaling feature pyramid network to generate a fused scaling feature pyramid network;
identifying each feature layer in the fused scaled feature pyramid network respectively to determine a corresponding defect prediction result in the marked product image;
and respectively correcting the third scaling parameter, the fourth scaling parameter and the initial twin network according to the difference between the defect prediction result and the defect labeling result corresponding to the labeled product image until the difference between the corresponding defect prediction result and the labeling result in the labeled product image is smaller than a threshold value so as to determine the twin network, the first scaling parameter and the second scaling parameter.
In some embodiments of the present disclosure, the first obtaining module 510 is specifically configured to:
and preprocessing the product image so that the imaging angle of the product in the preprocessed product image is the same as that of the product in the template image.
In this embodiment, an obtained product image to be detected and a template image are input into a twin network to determine a first feature pyramid network corresponding to the product image and a second feature pyramid network corresponding to the template image, then scaling each feature layer in the first feature pyramid network and each feature layer in the second feature pyramid network respectively based on a first scaling parameter and a second scaling parameter generated by training, fusing each feature layer in the first scaling feature pyramid network and each feature layer in the second scaling feature pyramid network respectively to generate a fused feature pyramid network, and finally identifying each feature layer in the fused feature pyramid network respectively to determine whether a defect exists in a product in the product image. Therefore, each feature layer in the first feature pyramid network corresponding to the product image and the second feature pyramid network corresponding to the template image is subjected to scaling treatment respectively, and then the feature layers subjected to scaling treatment are subjected to feature fusion, so that the determined defect type and position can be more accurate, and the defect detection accuracy is improved.
According to embodiments of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.
Fig. 6 illustrates a schematic block diagram of an example electronic device 600 that may be used to implement embodiments of the present disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the disclosure described and/or claimed herein.
As shown in fig. 6, the apparatus 600 includes a computing unit 601 that can perform various appropriate actions and processes according to a computer program stored in a Read Only Memory (ROM) 602 or a computer program loaded from a storage unit 608 into a Random Access Memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the device 600 may also be stored. The computing unit 601, ROM 602, and RAM 603 are connected to each other by a bus 604. An input/output (I/O) interface 605 is also connected to bus 604.
Various components in the device 600 are connected to the I/O interface 605, including: an input unit 606 such as a keyboard, mouse, etc.; an output unit 607 such as various types of displays, speakers, and the like; a storage unit 608, such as a magnetic disk, optical disk, or the like; and a communication unit 609 such as a network card, modem, wireless communication transceiver, etc. The communication unit 609 allows the device 600 to exchange information/data with other devices via a computer network, such as the internet, and/or various telecommunication networks.
The computing unit 601 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of computing unit 601 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various computing units running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 601 performs the respective methods and processes described above, for example, a defect detection method. For example, in some embodiments, the defect detection method may be implemented as a computer software program tangibly embodied on a machine-readable medium, such as storage unit 608. In some embodiments, part or all of the computer program may be loaded and/or installed onto the device 600 via the ROM 602 and/or the communication unit 609. When a computer program is loaded into the RAM 603 and executed by the computing unit 601, one or more steps of the defect detection method described above may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to perform the defect detection method by any other suitable means (e.g. by means of firmware).
Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuit systems, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems On Chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.
Program code for carrying out methods of the present disclosure may be written in any combination of one or more programming languages. These program code may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus such that the program code, when executed by the processor or controller, causes the functions/operations specified in the flowchart and/or block diagram to be implemented. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.
In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and pointing device (e.g., a mouse or trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), the internet, and blockchain networks.
The computer system may include a client and a server. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical hosts and VPS service ("Virtual Private Server" or simply "VPS") are overcome. The server may also be a server of a distributed system or a server that incorporates a blockchain.
In this embodiment, an obtained product image to be detected and a template image are input into a twin network to determine a first feature pyramid network corresponding to the product image and a second feature pyramid network corresponding to the template image, then scaling each feature layer in the first feature pyramid network and each feature layer in the second feature pyramid network respectively based on a first scaling parameter and a second scaling parameter generated by training, fusing each feature layer in the first scaling feature pyramid network and each feature layer in the second scaling feature pyramid network respectively to generate a fused feature pyramid network, and finally identifying each feature layer in the fused feature pyramid network respectively to determine whether a defect exists in a product in the product image. Therefore, each feature layer in the first feature pyramid network corresponding to the product image and the second feature pyramid network corresponding to the template image is subjected to scaling treatment respectively, and then the feature layers subjected to scaling treatment are subjected to feature fusion, so that the determined defect type and position can be more accurate, and the defect detection accuracy is improved.
It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps recited in the present disclosure may be performed in parallel or sequentially or in a different order, provided that the desired results of the technical solutions of the present disclosure are achieved, and are not limited herein.
The above detailed description should not be taken as limiting the scope of the present disclosure. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present disclosure are intended to be included within the scope of the present disclosure.
Claims (15)
1. A defect detection method, comprising:
acquiring a product image to be detected and a template image;
inputting the product image and the template image into a twin network to determine a first feature pyramid network corresponding to the product image and a second feature pyramid network corresponding to the template image;
scaling each feature layer in the first feature pyramid network and each feature layer in the second feature pyramid network based on a first scaling parameter and a second scaling parameter generated by training respectively to obtain a first scaled feature pyramid network and a second scaled feature pyramid network;
Respectively fusing each feature layer in the first scaled feature pyramid network with a corresponding feature layer in the second scaled feature pyramid network to generate a fused feature pyramid network;
identifying each feature layer in the fused feature pyramid network respectively to determine whether a product in the product image has a defect;
each feature layer in the first feature pyramid network and the second feature pyramid network respectively comprises at least one channel, and each feature layer in the first feature pyramid network and each feature layer in the second feature pyramid network are scaled based on a first scaling parameter and a second scaling parameter generated by training respectively, so as to obtain a first scaled feature pyramid network and a second scaled feature pyramid network, which comprises the following steps:
based on the first scaling parameters, respectively scaling the features of each first channel in each feature layer in the first feature pyramid network to obtain scaling features of each first channel;
and based on the second scaling parameters, respectively scaling the features of each second channel in each feature layer in the second feature pyramid network to obtain the scaling features of each second channel.
2. The method of claim 1, wherein the fusing each feature layer in the first scaled feature pyramid network with a corresponding feature layer in the second scaled feature pyramid network, respectively, to generate a fused feature pyramid network comprises:
and fusing the scaling characteristics of each first channel with the scaling characteristics of the corresponding second channel to generate fusion characteristics of each channel in each characteristic layer.
3. The method of claim 1, wherein each of the first and second feature pyramid networks includes at least one channel, the first scaling parameter includes a first sub-scaling parameter corresponding to each channel in each of the first and second feature pyramid networks, the second scaling parameter includes a second sub-scaling parameter corresponding to each channel in each of the second feature pyramid networks, and the scaling processing is performed on each of the first and second feature pyramid networks based on the first and second scaling parameters generated by training, respectively, to obtain the first and second scaled feature pyramid networks, including:
Scaling the features of the corresponding first channels based on first sub-scaling parameters corresponding to each first channel in each feature layer in each first feature pyramid network respectively to obtain scaling features of each first channel;
scaling the features of the corresponding second channels based on second sub-scaling parameters corresponding to each second channel in each feature layer in each second feature pyramid network respectively so as to acquire scaling features of each second channel;
the fusing each feature layer in the first scaled feature pyramid network and the corresponding feature layer in the second scaled feature pyramid network respectively to generate a fused feature pyramid network, including:
and fusing the scaling characteristics of each first channel with the scaling characteristics of the corresponding second channel to generate fusion characteristics of each channel in each characteristic layer.
4. The method of claim 1, wherein the identifying each feature layer in the fused feature pyramid network to determine whether a product in the product image is defective comprises:
Identifying each feature layer in the fused feature pyramid network to determine an identification result corresponding to each feature layer;
and determining whether the product in the product image has defects or not according to the identification result corresponding to each feature layer and the weight value of each feature layer.
5. The method of any of claims 1-4, wherein prior to said inputting the product image and the template image into a twinning network, further comprising:
acquiring a training data set, wherein the training data set comprises the template image and a plurality of marked product images;
inputting the noted product image and the template image into an initial twin network to determine a third feature pyramid network corresponding to the noted product image and a fourth feature pyramid network corresponding to the template image;
scaling each feature layer in the third feature pyramid network and each feature layer in the fourth feature pyramid network based on a preset third scaling parameter and a preset fourth scaling parameter respectively to obtain a third scaling feature pyramid network and a fourth scaling feature pyramid network;
Respectively fusing each feature layer in the third scaling feature pyramid network with a corresponding feature layer in the fourth scaling feature pyramid network to generate a fused scaling feature pyramid network;
identifying each feature layer in the fused scaled feature pyramid network respectively to determine a corresponding defect prediction result in the marked product image;
and respectively correcting the third scaling parameter, the fourth scaling parameter and the initial twin network according to the difference between the defect prediction result and the defect labeling result corresponding to the labeled product image until the difference between the defect prediction result and the labeling result corresponding to the labeled product image is smaller than a threshold value, so as to determine the twin network, the first scaling parameter and the second scaling parameter.
6. The method of any of claims 1-4, wherein prior to said inputting the product image and the template image into a twinning network, further comprising:
and preprocessing the product image so that the imaging angle of the product in the preprocessed product image is the same as the imaging angle of the product in the template image.
7. A defect detection apparatus comprising:
the first acquisition module is used for acquiring a product image to be detected and a template image;
the first determining module is used for inputting the product image and the template image into a twin network to determine a first characteristic pyramid network corresponding to the product image and a second characteristic pyramid network corresponding to the template image;
the second obtaining module is used for carrying out scaling processing on each feature layer in the first feature pyramid network and each feature layer in the second feature pyramid network based on the first scaling parameter and the second scaling parameter generated by training respectively so as to obtain the first scaling feature pyramid network and the second scaling feature pyramid network;
the first generation module is used for respectively fusing each feature layer in the first scaling feature pyramid network with the corresponding feature layer in the second scaling feature pyramid network to generate a fused feature pyramid network;
the second determining module is used for respectively identifying each characteristic layer in the fused characteristic pyramid network so as to determine whether the product in the product image has defects or not;
Each feature layer in the first feature pyramid network and the second feature pyramid network respectively comprises at least one channel, and the second acquisition module is further configured to:
based on the first scaling parameters, respectively scaling the features of each first channel in each feature layer in the first feature pyramid network to obtain scaling features of each first channel;
and based on the second scaling parameters, respectively scaling the features of each second channel in each feature layer in the second feature pyramid network to obtain the scaling features of each second channel.
8. The defect detection apparatus of claim 7, the first generation module further configured to:
and fusing the scaling characteristics of each first channel with the scaling characteristics of the corresponding second channel to generate fusion characteristics of each channel in each characteristic layer.
9. The defect detection apparatus of claim 7, wherein each of the first feature pyramid network and the second feature pyramid network includes at least one channel, the first scaling parameter includes a first sub-scaling parameter corresponding to each channel in each feature layer in the first feature pyramid network, the second scaling parameter includes a second sub-scaling parameter corresponding to each channel in each feature layer in the second feature pyramid network, and the second obtaining module is further configured to:
Scaling the features of the corresponding first channels based on first sub-scaling parameters corresponding to each first channel in each feature layer in each first feature pyramid network respectively to obtain scaling features of each first channel;
scaling the features of the corresponding second channels based on second sub-scaling parameters corresponding to each second channel in each feature layer in each second feature pyramid network respectively so as to acquire scaling features of each second channel;
the first generation module is specifically configured to:
and fusing the scaling characteristics of each first channel with the scaling characteristics of the corresponding second channel to generate fusion characteristics of each channel in each characteristic layer.
10. The defect detection apparatus of claim 7, wherein the second determination module is further configured to:
identifying each feature layer in the fused feature pyramid network to determine an identification result corresponding to each feature layer;
and determining whether the product in the product image has defects or not according to the identification result corresponding to each feature layer and the weight value of each feature layer.
11. The defect detection apparatus of any of claims 7-10, wherein the first determination module is further configured to:
acquiring a training data set, wherein the training data set comprises the template image and a plurality of marked product images;
inputting the noted product image and the template image into an initial twin network to determine a third feature pyramid network corresponding to the noted product image and a fourth feature pyramid network corresponding to the template image;
respectively carrying out scaling treatment on each feature layer in the third feature pyramid network and each feature layer in the fourth feature pyramid network based on a preset third scaling parameter and a preset fourth scaling parameter so as to obtain the third scaling feature pyramid network and the fourth scaling feature pyramid network;
respectively fusing each feature layer in the third scaling feature pyramid network with a corresponding feature layer in the fourth scaling feature pyramid network to generate a fused scaling feature pyramid network;
identifying each feature layer in the fused scaled feature pyramid network respectively to determine a corresponding defect prediction result in the marked product image;
And respectively correcting the third scaling parameter, the fourth scaling parameter and the initial twin network according to the difference between the defect prediction result and the defect labeling result corresponding to the labeled product image until the difference between the defect prediction result and the labeling result corresponding to the labeled product image is smaller than a threshold value, so as to determine the twin network, the first scaling parameter and the second scaling parameter.
12. The defect detection apparatus of any of claims 7-10, wherein the first acquisition module is further configured to:
and preprocessing the product image so that the imaging angle of the product in the preprocessed product image is the same as the imaging angle of the product in the template image.
13. An electronic device, comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-6.
14. A non-transitory computer readable storage medium storing computer instructions for causing the computer to perform the method of any one of claims 1-6.
15. A computer program product comprising computer instructions which, when executed by a processor, implement the steps of the method according to any of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110554117.4A CN113344862B (en) | 2021-05-20 | 2021-05-20 | Defect detection method, device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110554117.4A CN113344862B (en) | 2021-05-20 | 2021-05-20 | Defect detection method, device, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113344862A CN113344862A (en) | 2021-09-03 |
| CN113344862B true CN113344862B (en) | 2024-04-12 |
Family
ID=77470294
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110554117.4A Active CN113344862B (en) | 2021-05-20 | 2021-05-20 | Defect detection method, device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113344862B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114708484B (en) * | 2022-03-14 | 2023-04-07 | 中铁电气化局集团有限公司 | Pattern analysis method suitable for identifying defects |
| KR20240023970A (en) * | 2022-08-16 | 2024-02-23 | (주)파이버폭스 | Apparatus and method for verifying optical fiber work using artificial intelligence |
| CN117314895B (en) * | 2023-11-27 | 2024-03-12 | 北京阿丘科技有限公司 | Defect detection method, apparatus, and computer-readable storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019020075A1 (en) * | 2017-07-28 | 2019-01-31 | 北京市商汤科技开发有限公司 | Image processing method, device, storage medium, computer program, and electronic device |
| CN111754513A (en) * | 2020-08-07 | 2020-10-09 | 腾讯科技（深圳）有限公司 | Product surface defect segmentation method, defect segmentation model learning method and device |
| CN111898615A (en) * | 2020-06-16 | 2020-11-06 | 济南浪潮高新科技投资发展有限公司 | Feature extraction method, device, equipment and medium of object detection model |
| CN112184752A (en) * | 2020-09-08 | 2021-01-05 | 北京工业大学 | Video target tracking method based on pyramid convolution |
| WO2021026804A1 (en) * | 2019-08-14 | 2021-02-18 | 深圳市大疆创新科技有限公司 | Cradle head-based target following method and apparatus, cradle head and computer storage medium |
-
2021
- 2021-05-20 CN CN202110554117.4A patent/CN113344862B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019020075A1 (en) * | 2017-07-28 | 2019-01-31 | 北京市商汤科技开发有限公司 | Image processing method, device, storage medium, computer program, and electronic device |
| WO2021026804A1 (en) * | 2019-08-14 | 2021-02-18 | 深圳市大疆创新科技有限公司 | Cradle head-based target following method and apparatus, cradle head and computer storage medium |
| CN111898615A (en) * | 2020-06-16 | 2020-11-06 | 济南浪潮高新科技投资发展有限公司 | Feature extraction method, device, equipment and medium of object detection model |
| CN111754513A (en) * | 2020-08-07 | 2020-10-09 | 腾讯科技（深圳）有限公司 | Product surface defect segmentation method, defect segmentation model learning method and device |
| CN112184752A (en) * | 2020-09-08 | 2021-01-05 | 北京工业大学 | Video target tracking method based on pyramid convolution |
Non-Patent Citations (2)
| Title |
|---|
| 基于孪生卷积神经网络的图像融合;杨雪;郑婷婷;戴阳;;计算机系统应用(05);全文 * |
| 基于深度学习的齿轮视觉微小缺陷检测;韩明;吴庆祥;曾雄军;;计算机系统应用(03);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113344862A (en) | 2021-09-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113344862B (en) | Defect detection method, device, electronic equipment and storage medium | |
| CN113177472B (en) | Dynamic gesture recognition method, device, equipment and storage medium | |
| CN113177468B (en) | Human behavior detection method and device, electronic equipment and storage medium | |
| CN112784765B (en) | Method, apparatus, device and storage medium for recognizing motion | |
| CN113065614B (en) | Training method of classification model and method for classifying target object | |
| CN113379813A (en) | Training method and device of depth estimation model, electronic equipment and storage medium | |
| CN113362314B (en) | Medical image recognition method, recognition model training method and device | |
| CN113095336A (en) | Method for training key point detection model and method for detecting key points of target object | |
| CN112861885B (en) | Image recognition method, device, electronic equipment and storage medium | |
| CN113361572B (en) | Training method and device for image processing model, electronic equipment and storage medium | |
| CN113705362A (en) | Training method and device of image detection model, electronic equipment and storage medium | |
| CN116403083A (en) | Image processing method and device, electronic equipment and storage medium | |
| CN115359308B (en) | Model training method, device, equipment, storage medium and program for identifying difficult cases | |
| CN113744268A (en) | Crack detection method, electronic device and readable storage medium | |
| CN112580666A (en) | Image feature extraction method, training method, device, electronic equipment and medium | |
| CN113627361B (en) | Training method and device for face recognition model and computer program product | |
| CN113379877B (en) | Face video generation method and device, electronic equipment and storage medium | |
| CN113643260A (en) | Method, apparatus, device, medium and product for detecting image quality | |
| CN113177466A (en) | Identity recognition method and device based on face image, electronic equipment and medium | |
| CN113781653B (en) | Object model generation method and device, electronic equipment and storage medium | |
| CN116052288A (en) | Living body detection model training method, living body detection device and electronic equipment | |
| CN113850203A (en) | Adhesion detection model training method, adhesion detection method and related device | |
| CN113378774A (en) | Gesture recognition method, device, equipment, storage medium and program product | |
| CN116128863B (en) | Medical image processing method, device and equipment | |
| CN114693995B (en) | Model training method applied to image processing, image processing method and device |
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 |