CN117078677B - Defect detection method and system for starting sheet - Google Patents
Defect detection method and system for starting sheet Download PDFInfo
- Publication number
- CN117078677B CN117078677B CN202311330676.2A CN202311330676A CN117078677B CN 117078677 B CN117078677 B CN 117078677B CN 202311330676 A CN202311330676 A CN 202311330676A CN 117078677 B CN117078677 B CN 117078677B
- Authority
- CN
- China
- Prior art keywords
- detected
- image
- images
- target
- defect
- 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 144
- 238000001514 detection method Methods 0.000 title claims abstract description 50
- 238000000034 method Methods 0.000 claims abstract description 24
- 238000003860 storage Methods 0.000 claims description 13
- 238000013527 convolutional neural network Methods 0.000 claims description 12
- 238000013135 deep learning Methods 0.000 claims description 12
- 238000011176 pooling Methods 0.000 claims description 4
- 238000012216 screening Methods 0.000 claims description 4
- 238000004590 computer program Methods 0.000 claims description 3
- 206010027146 Melanoderma Diseases 0.000 claims description 2
- 238000002372 labelling Methods 0.000 claims description 2
- 230000037390 scarring Effects 0.000 claims 1
- 238000004519 manufacturing process Methods 0.000 abstract description 11
- 238000010586 diagram Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 1
- 239000012467 final product Substances 0.000 description 1
- 239000013067 intermediate product Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000009856 non-ferrous metallurgy Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 231100000241 scar Toxicity 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012549 training Methods 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
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
- G06T3/4038—Scaling the whole image or part thereof for image mosaicing, i.e. plane images composed of plane sub-images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2200/00—Indexing scheme for image data processing or generation, in general
- G06T2200/32—Indexing scheme for image data processing or generation, in general involving image mosaicing
-
- 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]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
-
- 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
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
- Y02E60/10—Energy storage using batteries
Abstract
The invention discloses a defect detection method and a defect detection system for a starting sheet, wherein the method comprises the following steps: acquiring at least two first images to be detected and at least two second images to be detected associated with the at least two first images to be detected, and splicing the at least two first images to be detected with the corresponding at least two second images to be detected to obtain at least two spliced images to be detected; comparing at least two spliced images to be detected with the target images to be detected, and respectively marking defect areas of the target images to be detected according to comparison results; taking difference of at least two marked target to-be-detected images, and removing defect areas with different positions to obtain at least one target defect area of the target to-be-detected images; and performing defect detection on at least one target defect area. The problem that a trained image recognition model in the prior art cannot be suitable for recognizing the starting sheet on a plurality of production lines is solved.
Description
Technical Field
The invention belongs to the technical field of defect detection, and particularly relates to a defect detection method and system for a starting sheet.
Background
The quality of the surface quality of the starting sheet, which is one of important intermediate products in the nonferrous metallurgy industry, influences the performance and quality of the final product.
In the defect recognition process of the starting sheet, a defect-free starting sheet and starting sheets with different defect types are generally adopted to train an image recognition model, and then the trained image recognition model is used for recognizing the starting sheet of a specific production line.
Disclosure of Invention
The invention provides a defect detection method and a defect detection system for a starting sheet, which are used for solving the technical problems that the method and the system are not suitable for the identification of the starting sheet on a plurality of production lines and are easy to cause low identification accuracy.
In a first aspect, the present invention provides a defect detection method for a starting sheet, including:
acquiring a plurality of images to be detected of the starting sheet in a period of continuous time, defining a sequence composed of all images to be detected before a target image to be detected as a first image sub-sequence to be detected, and defining a sequence composed of all images to be detected after the target image to be detected as a second image sub-sequence to be detected, wherein the target image to be detected is an image containing the complete starting sheet in the plurality of images to be detected;
establishing an association relation between each first to-be-detected image in the first to-be-detected image sub-sequence and each second to-be-detected image in the second to-be-detected image sub-sequence;
acquiring at least two first images to be detected and at least two second images to be detected associated with the at least two first images to be detected, and splicing the at least two first images to be detected and the corresponding at least two second images to be detected based on a preset splicing rule to obtain at least two spliced images to be detected;
comparing the at least two spliced images to be detected with the target image to be detected, and respectively marking the defect areas of the target image to be detected according to the comparison result to obtain at least two marked target images to be detected;
taking difference of at least two marked target to-be-detected images, and removing defect areas with different positions to obtain at least one target defect area of the target to-be-detected images;
and inputting the at least one target defect area into a preset deep learning convolutional neural network to detect defects, and obtaining the defect type of the target image to be detected.
In a second aspect, the present invention provides a defect detection system for a starting sheet, comprising:
the acquisition module is configured to acquire a plurality of images to be detected of the starting sheet in a continuous time, and define a sequence composed of all the images to be detected before a target image to be detected as a first image subsequence to be detected, and define a sequence composed of all the images to be detected after the target image to be detected as a second image subsequence to be detected, wherein the target image to be detected is an image containing the complete starting sheet in the plurality of images to be detected;
the establishing module is configured to establish the association relation between each first to-be-detected image in the first to-be-detected image sub-sequence and each second to-be-detected image in the second to-be-detected image sub-sequence;
the splicing module is configured to acquire at least two first images to be detected and at least two second images to be detected associated with the at least two first images to be detected, splice the at least two first images to be detected with the corresponding at least two second images to be detected based on preset splicing rules, and obtain at least two spliced images to be detected;
the comparison module is configured to compare the at least two spliced images to be detected with the target image to be detected, and respectively label the defect areas of the target image to be detected according to the comparison result to obtain at least two labeled target images to be detected;
the screening module is configured to make difference between at least two marked target to-be-detected images, and remove defect areas with different positions to obtain at least one target defect area of the target to-be-detected images;
the detection module is configured to input the at least one target defect area into a preset deep learning convolutional neural network to detect defects, and obtain the defect type of the target image to be detected.
In a third aspect, there is provided an electronic device, comprising: the device comprises at least one processor and a memory communicatively connected with 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 steps of the defect detection method for a starting sheet of any of the embodiments of the present invention.
In a fourth aspect, the present invention also provides a computer readable storage medium having stored thereon a computer program, the program instructions, when executed by a processor, cause the processor to perform the steps of the defect detection method for a starting sheet of any of the embodiments of the present invention.
The defect detection method and system for the starting sheet have the following beneficial effects:
acquiring at least two first images to be detected and at least two second images to be detected associated with the at least two first images to be detected, and splicing the at least two first images to be detected and the corresponding at least two second images to be detected based on preset splicing rules to obtain at least two spliced images to be detected; comparing the at least two spliced images to be detected with the target images to be detected, and respectively marking defect areas of the target images to be detected according to the comparison result to obtain at least two marked target images to be detected; taking difference of at least two marked target to-be-detected images, and removing defect areas with different positions to obtain at least one target defect area of the target to-be-detected images; at least one target defect area is input into a preset deep learning convolutional neural network for defect detection, the defect type of a target to-be-detected image is obtained, the current starting sheet can be used as a comparison object, the problem that a trained image recognition model in the prior art cannot be suitable for recognition of the starting sheet on a plurality of production lines is solved, and recognition accuracy is effectively improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flowchart of a defect detection method for a starting sheet according to an embodiment of the present invention;
FIG. 2 is a block diagram of a defect detection system for a starting sheet according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the embodiments of the present invention will be described in further detail with reference to the accompanying drawings. The exemplary embodiments of the present invention and their descriptions herein are for the purpose of explaining the present invention, but are not to be construed as limiting the invention.
In the description of the present specification, the terms "comprising," "including," "having," "containing," and the like are open-ended terms, meaning including, but not limited to. Reference to the terms "one embodiment," "a particular embodiment," "some embodiments," "for example," etc., means that a particular feature, structure, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present application. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiments or examples. Furthermore, the particular features, structures, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. The sequence of steps involved in the embodiments is used to schematically illustrate the practice of the present application, and is not limited thereto and may be appropriately adjusted as desired.
The embodiment of the invention provides a defect detection method for a starting sheet, and an execution body of the method can be at least one of a mobile phone, a tablet computer, a personal computer (Personal Computer, a PC) and the like which can be configured to execute the method provided by the embodiment of the invention, or the execution body of the method can also be an operating system or an Application (APP), or the execution body of the method can also be a server, and the like.
In an application scene, a plurality of production lines for producing the starting sheet are arranged, the starting sheets of the production lines are summarized, and finally a defect detection device is arranged at the terminal position of the production line for producing the starting sheet, and the defect detection device comprises an image acquisition unit and a defect detection unit connected with the image acquisition unit. Specifically, the image capturing unit may be an image capturing device in which the defect detection method of the present application can be performed. The image of the starting sheet transmitted by the transmission mechanism is acquired through the camera device, and the photographed image of the starting sheet can be analyzed whether the current photographed starting sheet has defects or not and the specific defect type after being subjected to defect identification of the defect detection unit.
Based on the application scenario, a defect-free starting sheet and starting sheets with different defect types are generally adopted to train an image recognition model in the defect recognition process, and then the trained image recognition model is used for recognizing the starting sheets of a specific production line.
The defect detection method for the starting sheet according to the embodiment of the present invention will be described in detail below with reference to the accompanying drawings.
As shown in fig. 1, the defect detection method for the starting sheet specifically includes the following steps:
step S101, a plurality of images to be detected of the starting sheet in a period of continuous time are obtained, a sequence composed of all images to be detected before a target image to be detected is defined as a first image sub-sequence to be detected, and a sequence composed of all images to be detected after the target image to be detected is defined as a second image sub-sequence to be detected, wherein the target image to be detected is an image containing the complete starting sheet among the plurality of images to be detected.
Firstly, a plurality of images to be detected of a starting sheet in a period of continuous time are acquired, the images to be detected are arranged based on the acquired time sequence, then a sequence composed of images to be detected before the image of the complete starting sheet is defined as a first image sub-sequence to be detected, and a sequence composed of all images to be detected after the image of the complete starting sheet is defined as a second image sub-sequence to be detected.
Generally, the photographing starting sheet is performed based on a preset time interval for a continuous period of time. For example, a camera is adopted to shoot the starting sheet, when the starting sheet is positioned below the camera, the starting sheet is shot continuously based on a time interval of 5s until the current starting sheet is far away from the camera, so that a plurality of images to be detected of the starting sheet in a period of continuous time are obtained.
Step S102, establishing an association relationship between each first to-be-detected image in the first to-be-detected image sub-sequence and each second to-be-detected image in the second to-be-detected image sub-sequence.
In the step, aligning a first to-be-detected image in a first to-be-detected image sub-sequence with a first to-be-detected image in a second to-be-detected image sub-sequence; and associating each first to-be-detected image in the first to-be-detected image sub-sequence with each second to-be-detected image in the second to-be-detected image sub-sequence based on the arrangement order.
For example, a first one of the first sub-sequences of images to be detected is associated with a first one of the second sub-sequences of images to be detected, and then a second one of the first sub-sequences of images to be detected is associated with a second one of the second sub-sequences of images to be detected until all of the first ones of the sub-sequences of images to be detected are associated.
Step S103, obtaining at least two first to-be-detected images and at least two second to-be-detected images associated with the at least two first to-be-detected images, and stitching the at least two first to-be-detected images and the corresponding at least two second to-be-detected images based on a preset stitching rule, so as to obtain at least two stitched to-be-detected images.
After at least two first to-be-detected images and at least two second to-be-detected images associated with the at least two first to-be-detected images are acquired, the at least two first to-be-detected images and the corresponding at least two second to-be-detected images are spliced based on preset splicing rules, and at least two spliced to-be-detected images are obtained.
Specifically, the splicing rule is: taking the long side of a starting sheet contained in a target image to be detected as a splicing direction; defining a head and a tail of a certain first to-be-detected image along the splicing direction and defining a head and a tail of a certain second to-be-detected image associated with the certain first to-be-detected image respectively; and splicing the head part of a certain first image to be detected and the tail part of a certain second image to be detected to obtain a certain spliced image to be detected, wherein the certain spliced image to be detected is different from the target image to be detected.
For example, an X-Y axis coordinate system is established, the target to-be-detected image is placed in the X-Y axis coordinate system, and the long side of the starting sheet included in the target to-be-detected image is parallel to the Y axis, so the Y axis is defined as the stitching direction. And then, placing the long sides of part of the starting sheet in a certain first to-be-detected image in an X-Y axis coordinate system in parallel with the Y axis direction, defining the upper part of the certain first to-be-detected image as a head part, defining the lower part of the certain first to-be-detected image as a tail part, defining the head part and the tail part of the certain second to-be-detected image as well, and finally splicing the tail part of the certain first to-be-detected image with the head part of the certain second to-be-detected image to obtain a spliced to-be-detected image.
And step S104, comparing the at least two spliced images to be detected with the target image to be detected, and respectively marking the defect areas of the target image to be detected according to the comparison result to obtain at least two marked target images to be detected.
In the step, inputting a certain spliced to-be-detected image and the target to-be-detected image into a preset image recognition model to obtain the characteristic variation of the target to-be-detected image relative to the certain spliced to-be-detected image; judging whether the characteristic variation is larger than a preset threshold value or not; marking a primary defect area of the target to-be-detected image according to the target characteristic variation relative to the certain spliced to-be-detected image to obtain the primary defect area of the target to-be-detected image, wherein the target characteristic variation is a characteristic variation larger than a preset threshold value; performing secondary defect region labeling on the target to-be-detected image according to the target characteristic variation relative to the other spliced to-be-detected image to obtain a secondary defect region of the target to-be-detected image; and obtaining two marked target images to be detected according to the primary defect area and the secondary defect area.
It should be noted that, an X-Y axis coordinate system is established, and the target to-be-detected image and a certain splicing to-be-detected image are overlapped and placed in the X-Y axis coordinate system, so that the feature variation of the target to-be-detected image relative to the certain splicing to-be-detected image is the difference between the feature point of the target to-be-detected image at the coordinates (X, Y) and the feature point of the certain splicing to-be-detected image at the coordinates (X, Y).
Specifically, the image recognition model is a VGGNet network model, wherein the VGGNet network model is obtained by machine learning using a plurality of sets of training images and detection images.
Step S105, the difference is made between at least two marked target to-be-detected images, and defect areas with different positions are removed, so that at least one target defect area of the target to-be-detected images is obtained.
And S106, inputting the at least one target defect area into a preset deep learning convolutional neural network for defect detection to obtain the defect type of the target image to be detected.
In this step, the deep learning convolutional neural network includes a convolutional layer, a unit layer, a pooling layer, a full-connection layer, and a SOFTMAX layer; the convolution layer is used for extracting pixel characteristics in at least one target defect area; the unit layer is used for extracting texture and defect outline characteristics in at least one target defect area, combining input and output of the unit layer, inputting the combined result into the unit layer again, continuing extracting the characteristics, and finally outputting advanced characteristics of at least one target defect area; the pooling layer is used for simplifying the advanced features; the full connection layer is used for judging the defect type close to the advanced feature; the SOFTMAX layer is used for outputting a detection result of at least one target defect area.
The defect type includes any one of a scar defect, a black spot defect, and a streak defect.
In summary, compared with the mode of directly carrying out recognition through the recognition model in the prior art, although the purpose of quick recognition cannot be achieved, in the application scene of recognizing different starting sheets after a plurality of production lines are combined, because the starting sheets are continuously moved through the transmission device, considerable time is required to be spent in the moving process, even if a quick recognition method is adopted, the detection time is shortened, the efficiency of the production line cannot be improved, the detection time is shortened to a certain extent, the defect detection precision is emphasized, and the workload of manual sorting of subsequent workers is reduced. According to the method, at least two first images to be detected and at least two second images to be detected which are associated with the at least two first images to be detected are obtained, and the at least two first images to be detected and the corresponding at least two second images to be detected are spliced based on a preset splicing rule, so that at least two spliced images to be detected are obtained; comparing the at least two spliced images to be detected with the target images to be detected, and respectively marking defect areas of the target images to be detected according to the comparison result to obtain at least two marked target images to be detected; taking difference of at least two marked target to-be-detected images, and removing defect areas with different positions to obtain at least one target defect area of the target to-be-detected images; at least one target defect area is input into a preset deep learning convolutional neural network for defect detection, the defect type of a target to-be-detected image is obtained, the current starting sheet can be used as a comparison object, the problem that a trained image recognition model in the prior art cannot be suitable for recognition of the starting sheet on a plurality of production lines is solved, and recognition accuracy is effectively improved.
Referring to fig. 2, a block diagram of a defect detection system for a starting sheet of the present application is shown.
As shown in fig. 2, the defect detection system 200 includes an acquisition module 210, an establishment module 220, a stitching module 230, a comparison module 240, a screening module 250, and a detection module 260.
The acquiring module 210 is configured to acquire a plurality of images to be detected of the starting sheet within a continuous time, and define a sequence composed of all images to be detected before a target image to be detected as a first image sub-sequence, and define a sequence composed of all images to be detected after the target image to be detected as a second image sub-sequence, where the target image to be detected is an image including the complete starting sheet among the plurality of images to be detected; the establishing module 220 is configured to establish an association relationship between each first to-be-detected image in the first to-be-detected image sub-sequence and each second to-be-detected image in the second to-be-detected image sub-sequence; the stitching module 230 is configured to obtain at least two first to-be-detected images and at least two second to-be-detected images associated with the at least two first to-be-detected images, and stitch the at least two first to-be-detected images and the corresponding at least two second to-be-detected images based on a preset stitching rule, so as to obtain at least two stitched to-be-detected images; the comparison module 240 is configured to compare the at least two spliced images to be detected with the target image to be detected, and respectively label the defect areas of the target image to be detected according to the comparison result, so as to obtain at least two labeled target images to be detected; the screening module 250 is configured to perform difference on at least two marked target to-be-detected images, and remove defect areas with different positions to obtain at least one target defect area of the target to-be-detected images; the detection module 260 is configured to input the at least one target defect area into a preset deep learning convolutional neural network to perform defect detection, so as to obtain a defect type of the target image to be detected.
It should be understood that the modules depicted in fig. 2 correspond to the various steps in the method described with reference to fig. 1. Thus, the operations and features described above for the method and the corresponding technical effects are equally applicable to the modules in fig. 2, and are not described here again.
In other embodiments, embodiments of the present invention further provide a computer readable storage medium having stored thereon a computer program, where the program instructions, when executed by a processor, cause the processor to perform the defect detection method for a starting sheet in any of the method embodiments described above;
as one embodiment, the computer-readable storage medium of the present invention stores computer-executable instructions configured to:
acquiring a plurality of images to be detected of the starting sheet in a period of continuous time, defining a sequence composed of all images to be detected before a target image to be detected as a first image sub-sequence to be detected, and defining a sequence composed of all images to be detected after the target image to be detected as a second image sub-sequence to be detected, wherein the target image to be detected is an image containing the complete starting sheet in the plurality of images to be detected;
establishing an association relation between each first to-be-detected image in the first to-be-detected image sub-sequence and each second to-be-detected image in the second to-be-detected image sub-sequence;
acquiring at least two first images to be detected and at least two second images to be detected associated with the at least two first images to be detected, and splicing the at least two first images to be detected and the corresponding at least two second images to be detected based on a preset splicing rule to obtain at least two spliced images to be detected;
comparing the at least two spliced images to be detected with the target image to be detected, and respectively marking the defect areas of the target image to be detected according to the comparison result to obtain at least two marked target images to be detected;
taking difference of at least two marked target to-be-detected images, and removing defect areas with different positions to obtain at least one target defect area of the target to-be-detected images;
and inputting the at least one target defect area into a preset deep learning convolutional neural network to detect defects, and obtaining the defect type of the target image to be detected.
The computer readable storage medium may include a storage program area and a storage data area, wherein the storage program area may store an operating system, at least one application program required for a function; the storage data area may store data created according to the use of a defect detection system for the starting sheet, etc. In addition, the computer-readable storage medium may include high-speed random access memory, and may also include memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, the computer readable storage medium optionally includes memory remotely located with respect to the processor, which may be connected to the defect detection system for the starting sheet via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
Fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the present invention, as shown in fig. 3, where the device includes: a processor 310 and a memory 320. The electronic device may further include: an input device 330 and an output device 340. The processor 310, memory 320, input device 330, and output device 340 may be connected by a bus or other means, for example in fig. 3. Memory 320 is the computer-readable storage medium described above. The processor 310 executes various functional applications of the server and data processing by running non-volatile software programs, instructions and modules stored in the memory 320, i.e., implements the defect detection method for the starting sheet of the above-described method embodiment. The input device 330 may receive input numeric or character information and generate key signal inputs related to user settings and function control of the defect detection system for the starting sheet. The output device 340 may include a display device such as a display screen.
The electronic equipment can execute the method provided by the embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method. Technical details not described in detail in this embodiment may be found in the methods provided in the embodiments of the present invention.
As an embodiment, the electronic device is applied to a defect detection system for a starting sheet, and is used for a client, and includes: 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, the instructions being executable by the at least one processor to enable the at least one processor to:
acquiring a plurality of images to be detected of the starting sheet in a period of continuous time, defining a sequence composed of all images to be detected before a target image to be detected as a first image sub-sequence to be detected, and defining a sequence composed of all images to be detected after the target image to be detected as a second image sub-sequence to be detected, wherein the target image to be detected is an image containing the complete starting sheet in the plurality of images to be detected;
establishing an association relation between each first to-be-detected image in the first to-be-detected image sub-sequence and each second to-be-detected image in the second to-be-detected image sub-sequence;
acquiring at least two first images to be detected and at least two second images to be detected associated with the at least two first images to be detected, and splicing the at least two first images to be detected and the corresponding at least two second images to be detected based on a preset splicing rule to obtain at least two spliced images to be detected;
comparing the at least two spliced images to be detected with the target image to be detected, and respectively marking the defect areas of the target image to be detected according to the comparison result to obtain at least two marked target images to be detected;
taking difference of at least two marked target to-be-detected images, and removing defect areas with different positions to obtain at least one target defect area of the target to-be-detected images;
and inputting the at least one target defect area into a preset deep learning convolutional neural network to detect defects, and obtaining the defect type of the target image to be detected.
From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on such understanding, the foregoing technical solutions may be embodied essentially or in part in the form of a software product, which may be stored in a computer-readable storage medium, such as a ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to perform the various embodiments or methods of some parts of the embodiments.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (9)
1. A defect detection method for a starting sheet, comprising:
acquiring a plurality of images to be detected of the starting sheet in a period of continuous time, defining a sequence composed of all images to be detected before a target image to be detected as a first image sub-sequence to be detected, and defining a sequence composed of all images to be detected after the target image to be detected as a second image sub-sequence to be detected, wherein the target image to be detected is an image containing the complete starting sheet in the plurality of images to be detected;
establishing an association relationship between each first to-be-detected image in the first to-be-detected image sub-sequence and each second to-be-detected image in the second to-be-detected image sub-sequence, wherein the establishing an association relationship between each first to-be-detected image in the first to-be-detected image sub-sequence and each second to-be-detected image in the second to-be-detected image sub-sequence comprises:
aligning a first one of the first sub-sequences of images to be detected with a first one of the second sub-sequences of images to be detected;
associating each first to-be-detected image in the first to-be-detected image sub-sequence with each second to-be-detected image in the second to-be-detected image sub-sequence based on the arrangement order;
acquiring at least two first images to be detected and at least two second images to be detected associated with the at least two first images to be detected, and splicing the at least two first images to be detected and the corresponding at least two second images to be detected based on a preset splicing rule to obtain at least two spliced images to be detected;
comparing the at least two spliced images to be detected with the target image to be detected, and respectively marking the defect areas of the target image to be detected according to the comparison result to obtain at least two marked target images to be detected;
taking difference of at least two marked target to-be-detected images, and removing defect areas with different positions to obtain at least one target defect area of the target to-be-detected images;
and inputting the at least one target defect area into a preset deep learning convolutional neural network to detect defects, and obtaining the defect type of the target image to be detected.
2. The method for detecting defects of a starting sheet according to claim 1, wherein the stitching the at least two first images to be detected with the corresponding at least two second images to be detected based on a preset stitching rule, to obtain at least two stitched images to be detected, includes:
taking the long side of the starting sheet contained in the target image to be detected as a splicing direction;
defining a head and a tail of a certain first image to be detected along the splicing direction, and defining a head and a tail of a certain second image to be detected associated with the certain first image to be detected;
and splicing the head part of the certain first image to be detected and the tail part of the certain second image to be detected to obtain a certain splicing image to be detected, wherein the certain splicing image to be detected is different from the target image to be detected.
3. The method for detecting defects of a starting sheet according to claim 1, wherein comparing the at least two stitched images to be detected with the target image to be detected comprises:
inputting a certain splicing to-be-detected image and the target to-be-detected image into a preset image recognition model to obtain the characteristic variation of the target to-be-detected image relative to the certain splicing to-be-detected image;
and judging whether the characteristic variation is larger than a preset threshold value.
4. The method for detecting defects of a starting sheet according to claim 3, wherein the marking the defect areas of the target to-be-detected images according to the comparison result respectively, and obtaining at least two marked target to-be-detected images comprises:
marking a primary defect area of the target to-be-detected image according to the target characteristic variation relative to the certain spliced to-be-detected image to obtain the primary defect area of the target to-be-detected image, wherein the target characteristic variation is a characteristic variation larger than a preset threshold value;
performing secondary defect region labeling on the target to-be-detected image according to the target characteristic variation relative to the other spliced to-be-detected image to obtain a secondary defect region of the target to-be-detected image;
and obtaining two marked target images to be detected according to the primary defect area and the secondary defect area.
5. The defect detection method for a starting sheet of claim 1, wherein the deep learning convolutional neural network comprises a convolutional layer, a unit layer, a pooling layer, a fully-connected layer, and a SOFTMAX layer;
the convolution layer is used for extracting pixel characteristics in the at least one target defect area; the unit layer is used for extracting texture and defect contour characteristics in the at least one target defect area, combining input and output of the unit layer, inputting the combined result into the unit layer again, continuously extracting the characteristics, and finally outputting high-level characteristics of the at least one target defect area; the pooling layer is used for simplifying the advanced features; the full connection layer is used for judging the defect type close to the advanced feature; the SOFTMAX layer is used for outputting a detection result of the at least one target defect area.
6. The defect detection method for a starting sheet according to claim 5, wherein the defect type includes any one of a scarring defect, a black spot defect, and a streak defect.
7. A defect detection system for a starting sheet, comprising:
the acquisition module is configured to acquire a plurality of images to be detected of the starting sheet in a continuous time, and define a sequence composed of all the images to be detected before a target image to be detected as a first image subsequence to be detected, and define a sequence composed of all the images to be detected after the target image to be detected as a second image subsequence to be detected, wherein the target image to be detected is an image containing the complete starting sheet in the plurality of images to be detected;
the establishing module is configured to establish an association relationship between each first to-be-detected image in the first to-be-detected image sub-sequence and each second to-be-detected image in the second to-be-detected image sub-sequence, wherein the establishing an association relationship between each first to-be-detected image in the first to-be-detected image sub-sequence and each second to-be-detected image in the second to-be-detected image sub-sequence includes:
aligning a first one of the first sub-sequences of images to be detected with a first one of the second sub-sequences of images to be detected;
associating each first to-be-detected image in the first to-be-detected image sub-sequence with each second to-be-detected image in the second to-be-detected image sub-sequence based on the arrangement order;
the splicing module is configured to acquire at least two first images to be detected and at least two second images to be detected associated with the at least two first images to be detected, splice the at least two first images to be detected with the corresponding at least two second images to be detected based on preset splicing rules, and obtain at least two spliced images to be detected;
the comparison module is configured to compare the at least two spliced images to be detected with the target image to be detected, and respectively label the defect areas of the target image to be detected according to the comparison result to obtain at least two labeled target images to be detected;
the screening module is configured to make difference between at least two marked target to-be-detected images, and remove defect areas with different positions to obtain at least one target defect area of the target to-be-detected images;
the detection module is configured to input the at least one target defect area into a preset deep learning convolutional neural network to detect defects, and obtain the defect type of the target image to be detected.
8. 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 to 6.
9. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the method of any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311330676.2A CN117078677B (en) | 2023-10-16 | 2023-10-16 | Defect detection method and system for starting sheet |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311330676.2A CN117078677B (en) | 2023-10-16 | 2023-10-16 | Defect detection method and system for starting sheet |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117078677A CN117078677A (en) | 2023-11-17 |
| CN117078677B true CN117078677B (en) | 2024-01-30 |
Family
ID=88710156
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311330676.2A Active CN117078677B (en) | 2023-10-16 | 2023-10-16 | Defect detection method and system for starting sheet |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117078677B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117437221B (en) * | 2023-12-18 | 2024-04-26 | 浙江祥晋汽车零部件股份有限公司 | Method and system for detecting bright decorative strip based on image detection |
Citations (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111985584A (en) * | 2020-09-30 | 2020-11-24 | 平安科技（深圳）有限公司 | Disease auxiliary detection equipment, method, device and medium based on multi-mode data |
| CN112001841A (en) * | 2020-07-14 | 2020-11-27 | 歌尔股份有限公司 | Image to-be-detected region extraction method and device and product defect detection system |
| CN112598627A (en) * | 2020-12-10 | 2021-04-02 | 广东省大湾区集成电路与系统应用研究院 | Method, system, electronic device and medium for detecting image defects |
| CN112748120A (en) * | 2020-12-29 | 2021-05-04 | 中国联合网络通信集团有限公司 | Defect detection system, defect detection method, defect detection device, defect detection equipment and storage medium |
| WO2021136510A1 (en) * | 2019-12-31 | 2021-07-08 | 杭州海康机器人技术有限公司 | Data acquisition system, method, device, electronic apparatus, and storage medium |
| WO2021196925A1 (en) * | 2020-03-31 | 2021-10-07 | 清华大学 | Method and apparatus for detecting and tracking moving object |
| CN113505781A (en) * | 2021-06-01 | 2021-10-15 | 北京旷视科技有限公司 | Target detection method and device, electronic equipment and readable storage medium |
| CN114299066A (en) * | 2022-03-03 | 2022-04-08 | 清华大学 | Defect detection method and device based on salient feature pre-extraction and image segmentation |
| CN114359055A (en) * | 2022-03-21 | 2022-04-15 | 湖南大学 | Image splicing method and related device for multi-camera shooting screen body |
| CN114565567A (en) * | 2022-02-15 | 2022-05-31 | 清华大学 | Method and device for detecting defects of complex texture lace cloth |
| CN114972174A (en) * | 2022-04-06 | 2022-08-30 | 电子科技大学中山学院 | Defect detection method and device, electronic equipment and storage medium |
| CN114998186A (en) * | 2022-03-31 | 2022-09-02 | 南通恒立机械设备有限公司 | Image processing-based method and system for detecting surface scab defect of copper starting sheet |
| CN115078392A (en) * | 2022-07-29 | 2022-09-20 | 深圳市泽信智能装备有限公司 | Semiconductor chip defect detection equipment and method |
| CN115131283A (en) * | 2022-04-15 | 2022-09-30 | 腾讯科技（深圳）有限公司 | Defect detection and model training method, device, equipment and medium for target object |
| CN115205247A (en) * | 2022-07-13 | 2022-10-18 | 上海商汤智能科技有限公司 | Method, device and equipment for detecting defects of battery pole piece and storage medium |
| CN115239719A (en) * | 2022-09-22 | 2022-10-25 | 南昌昂坤半导体设备有限公司 | Defect detection method, system, electronic device and storage medium |
| CN115294063A (en) * | 2022-08-05 | 2022-11-04 | 中国电信股份有限公司 | Method, device, system, electronic device and medium for detecting defect of electronic component |
| CN218097731U (en) * | 2022-06-20 | 2022-12-20 | 江西天鑫冶金装备技术有限公司 | Starting sheet sag detection device |
| CN115526892A (en) * | 2022-11-29 | 2022-12-27 | 南方电网数字电网研究院有限公司 | Image defect duplicate removal detection method and device based on three-dimensional reconstruction |
| WO2023087741A1 (en) * | 2021-11-18 | 2023-05-25 | 上海商汤智能科技有限公司 | Defect detection method and apparatus, and electronic device, storage medium and computer program product |
| CN116503414A (en) * | 2023-06-29 | 2023-07-28 | 深圳思谋信息科技有限公司 | Screen defect detection method, device, computer equipment and storage medium |
| CN116718616A (en) * | 2023-08-04 | 2023-09-08 | 苏州视谷视觉技术有限公司 | Machine vision detection system and detection method for flaw detection |
| CN116848547A (en) * | 2021-12-07 | 2023-10-03 | 宁德时代新能源科技股份有限公司 | Image processing method and system |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180211373A1 (en) * | 2017-01-20 | 2018-07-26 | Aquifi, Inc. | Systems and methods for defect detection |
| CN113947560A (en) * | 2020-07-17 | 2022-01-18 | 京东方科技集团股份有限公司 | Product detection method, device, system and computer storage medium |
-
2023
- 2023-10-16 CN CN202311330676.2A patent/CN117078677B/en active Active
Patent Citations (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021136510A1 (en) * | 2019-12-31 | 2021-07-08 | 杭州海康机器人技术有限公司 | Data acquisition system, method, device, electronic apparatus, and storage medium |
| WO2021196925A1 (en) * | 2020-03-31 | 2021-10-07 | 清华大学 | Method and apparatus for detecting and tracking moving object |
| CN112001841A (en) * | 2020-07-14 | 2020-11-27 | 歌尔股份有限公司 | Image to-be-detected region extraction method and device and product defect detection system |
| CN111985584A (en) * | 2020-09-30 | 2020-11-24 | 平安科技（深圳）有限公司 | Disease auxiliary detection equipment, method, device and medium based on multi-mode data |
| CN112598627A (en) * | 2020-12-10 | 2021-04-02 | 广东省大湾区集成电路与系统应用研究院 | Method, system, electronic device and medium for detecting image defects |
| CN112748120A (en) * | 2020-12-29 | 2021-05-04 | 中国联合网络通信集团有限公司 | Defect detection system, defect detection method, defect detection device, defect detection equipment and storage medium |
| CN113505781A (en) * | 2021-06-01 | 2021-10-15 | 北京旷视科技有限公司 | Target detection method and device, electronic equipment and readable storage medium |
| WO2023087741A1 (en) * | 2021-11-18 | 2023-05-25 | 上海商汤智能科技有限公司 | Defect detection method and apparatus, and electronic device, storage medium and computer program product |
| CN116848547A (en) * | 2021-12-07 | 2023-10-03 | 宁德时代新能源科技股份有限公司 | Image processing method and system |
| CN114565567A (en) * | 2022-02-15 | 2022-05-31 | 清华大学 | Method and device for detecting defects of complex texture lace cloth |
| CN114299066A (en) * | 2022-03-03 | 2022-04-08 | 清华大学 | Defect detection method and device based on salient feature pre-extraction and image segmentation |
| CN114359055A (en) * | 2022-03-21 | 2022-04-15 | 湖南大学 | Image splicing method and related device for multi-camera shooting screen body |
| CN114998186A (en) * | 2022-03-31 | 2022-09-02 | 南通恒立机械设备有限公司 | Image processing-based method and system for detecting surface scab defect of copper starting sheet |
| CN114972174A (en) * | 2022-04-06 | 2022-08-30 | 电子科技大学中山学院 | Defect detection method and device, electronic equipment and storage medium |
| CN115131283A (en) * | 2022-04-15 | 2022-09-30 | 腾讯科技（深圳）有限公司 | Defect detection and model training method, device, equipment and medium for target object |
| CN218097731U (en) * | 2022-06-20 | 2022-12-20 | 江西天鑫冶金装备技术有限公司 | Starting sheet sag detection device |
| CN115205247A (en) * | 2022-07-13 | 2022-10-18 | 上海商汤智能科技有限公司 | Method, device and equipment for detecting defects of battery pole piece and storage medium |
| CN115078392A (en) * | 2022-07-29 | 2022-09-20 | 深圳市泽信智能装备有限公司 | Semiconductor chip defect detection equipment and method |
| CN115294063A (en) * | 2022-08-05 | 2022-11-04 | 中国电信股份有限公司 | Method, device, system, electronic device and medium for detecting defect of electronic component |
| CN115239719A (en) * | 2022-09-22 | 2022-10-25 | 南昌昂坤半导体设备有限公司 | Defect detection method, system, electronic device and storage medium |
| CN115526892A (en) * | 2022-11-29 | 2022-12-27 | 南方电网数字电网研究院有限公司 | Image defect duplicate removal detection method and device based on three-dimensional reconstruction |
| CN116503414A (en) * | 2023-06-29 | 2023-07-28 | 深圳思谋信息科技有限公司 | Screen defect detection method, device, computer equipment and storage medium |
| CN116718616A (en) * | 2023-08-04 | 2023-09-08 | 苏州视谷视觉技术有限公司 | Machine vision detection system and detection method for flaw detection |
Non-Patent Citations (2)
| Title |
|---|
| Research on the Surface Defect Detection of Magnetic Sheet for Industrial Manufacture;Wenbo Zhang 等;《International Journal of Hybrid Information Technology》;第9卷(第1期);第253-266页 * |
| 基于红外图像的铜电解精炼过程状态检测研究;赵仁涛;《CNKI学位论文》;第2016年卷(第04期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117078677A (en) | 2023-11-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117078677B (en) | Defect detection method and system for starting sheet | |
| CN113239930B (en) | Glass paper defect identification method, system, device and storage medium | |
| CN114240939B (en) | Method, system, equipment and medium for detecting appearance defects of mainboard components | |
| CN110781877B (en) | Image recognition method, device and storage medium | |
| CN104239909A (en) | Method and device for recognizing images | |
| CN111339902B (en) | Liquid crystal display indication recognition method and device for digital display instrument | |
| CN113284154B (en) | Steel coil end face image segmentation method and device and electronic equipment | |
| CN115661161B (en) | Defect detection method, device, storage medium, apparatus and program product for parts | |
| CN114782286A (en) | Defect repairing method, optical repairing device, electronic device and storage medium | |
| CN112116582A (en) | Cigarette detection and identification method under stock or display scene | |
| CN115272204A (en) | Bearing surface scratch detection method based on machine vision | |
| CN110599453A (en) | Panel defect detection method and device based on image fusion and equipment terminal | |
| CN111311487B (en) | Rapid splicing method and system for photovoltaic module images | |
| CN113066088A (en) | Detection method, detection device and storage medium in industrial detection | |
| CN115239734A (en) | Model training method, device, storage medium, equipment and computer program product | |
| CN107862314B (en) | Code spraying identification method and device | |
| CN113780484A (en) | Industrial product defect detection method and device | |
| CN114626118A (en) | Building indoor model generation method and device | |
| CN115457252A (en) | Method and device for positioning welding key points | |
| WO2017219562A1 (en) | Method and apparatus for generating two-dimensional code | |
| CN105184244B (en) | Video human face detection method and device | |
| CN111935480B (en) | Detection method for image acquisition device and related device | |
| CN109325441B (en) | Method for identifying insulator object of power transmission line | |
| CN110135426B (en) | Sample labeling method and computer storage medium | |
| CN111556251A (en) | Electronic book generation method, device and medium |
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 |