CN115880288A - Detection method and system for electronic component welding and computer equipment - Google Patents
Detection method and system for electronic component welding and computer equipment Download PDFInfo
- Publication number
- CN115880288A CN115880288A CN202310143229.XA CN202310143229A CN115880288A CN 115880288 A CN115880288 A CN 115880288A CN 202310143229 A CN202310143229 A CN 202310143229A CN 115880288 A CN115880288 A CN 115880288A
- Authority
- CN
- China
- Prior art keywords
- welding
- image
- detected
- standard
- area
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000003466 welding Methods 0.000 title claims abstract description 369
- 238000001514 detection method Methods 0.000 title claims abstract description 96
- 230000007547 defect Effects 0.000 claims abstract description 29
- 238000000034 method Methods 0.000 claims description 36
- 230000004927 fusion Effects 0.000 claims description 27
- 238000005476 soldering Methods 0.000 claims description 20
- 238000004364 calculation method Methods 0.000 claims description 16
- 238000007781 pre-processing Methods 0.000 claims description 15
- 230000011218 segmentation Effects 0.000 claims description 12
- 238000003709 image segmentation Methods 0.000 claims description 9
- 238000013528 artificial neural network Methods 0.000 claims description 4
- 238000001914 filtration Methods 0.000 claims description 3
- 238000004590 computer program Methods 0.000 claims description 2
- 229910000679 solder Inorganic materials 0.000 claims 2
- 238000013527 convolutional neural network Methods 0.000 abstract description 10
- 239000010410 layer Substances 0.000 description 62
- 238000010586 diagram Methods 0.000 description 11
- 230000008569 process Effects 0.000 description 9
- 230000000877 morphologic effect Effects 0.000 description 8
- 238000003860 storage Methods 0.000 description 6
- 230000006870 function Effects 0.000 description 4
- 238000000605 extraction Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 150000001875 compounds Chemical class 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- ATJFFYVFTNAWJD-UHFFFAOYSA-N Tin Chemical compound [Sn] ATJFFYVFTNAWJD-UHFFFAOYSA-N 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 125000004432 carbon atom Chemical group C* 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 239000002345 surface coating layer Substances 0.000 description 1
- 238000011179 visual inspection Methods 0.000 description 1
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of automatic detection, solves the technical problems of high detection cost and low efficiency, and particularly relates to a detection method, a system and computer equipment for electronic element welding, wherein the detection method comprises the following steps: acquiring a standard welding image and a welding image to be detected, which are acquired by image acquisition equipment; calculating the matching coefficient and the difference area of the standard welding image and the welding image to be detected; performing primary detection on the electronic element welding product to be detected according to a preset matching coefficient threshold value and a difference area threshold value; and searching and comparing a plurality of characteristic images with different resolutions of the welding image to be detected layer by layer according to the invariant moment characteristics of the image to obtain a welding area to be detected which is accurately matched with the standard welding image. According to the invention, by adopting a two-stage detection mode, products with obvious welding defects are firstly screened out, and then the improved convolutional neural network is utilized to carry out secondary detection on the products without the obvious welding defects, so that the detection efficiency is improved, and the cost is low.
Description
Technical Field
The invention relates to the technical field of automatic detection, in particular to a detection method and system for electronic element welding and computer equipment.
Background
When electronic components are welded, no matter manual or mechanical production, welding defects such as missing welding, tin shortage, askew pasting, bridging and the like occur with certain probability, and the defects seriously affect the reliability of electronic products, so that the detection of the welding quality of the electronic components is particularly important.
At present, the basic idea of using visual inspection technology to automatically inspect the welding quality of electronic components in the market is to acquire an image of a welding part by using image acquisition equipment, perform denoising and segmentation processing on the image of the welding part to obtain an image to be inspected which only contains a welding area, and input the image to be inspected into an inspection model to perform characteristic analysis so as to judge whether the image has defects and determine the types of the defects.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a detection method, a detection system and computer equipment for electronic element welding, solves the technical problems of high acquisition cost of high-resolution welding images and low detection efficiency caused by low transmission speed, and achieves the purpose of improving the detection efficiency and accuracy of the electronic element welding quality on the premise of not increasing the detection cost.
In order to solve the technical problems, the invention provides the following technical scheme: a method of detecting soldering of electronic components, comprising the steps of:
s1, acquiring a standard welding image acquired by image acquisition equipment and a welding image to be detected of an electronic element welding piece to be detected, wherein the standard welding image refers to an image which is qualified in welding quality and only comprises an electronic element and a welding area of the electronic element;
s2, calculating a matching coefficient r and a difference area Sc of the standard welding image and the welding image to be detected;
s3, performing primary detection on the electronic element welding product to be detected according to a preset matching coefficient threshold value Tr and a difference area threshold value Ts;
if the | r | is less than Tr and the | Sc | is less than Ts, determining that the product is unqualified and finishing the detection, otherwise, executing the step S4;
s4, searching and comparing a plurality of characteristic images with different resolutions of the welding image to be detected layer by layer according to the invariant moment characteristics of the image to obtain a welding area to be detected which is accurately matched with the standard welding image;
s5, performing image segmentation on the welding image to be detected according to the welding area to be detected to obtain a target image to be detected;
s6, preprocessing the target image to be detected to obtain a preprocessed image with clear welding spot outline and electronic element support body layers;
s7, extracting a welding spot shape line to be detected from the preprocessed image according to a segmentation threshold, and calculating a shape difference value X between the welding spot shape line to be detected and a preset standard welding spot shape line;
s8, performing secondary detection according to a comparison result of the form difference value X and a preset detection threshold Tx, and judging whether a welding defect exists or not;
and if the form difference value X exceeds the range of a preset detection threshold value Tx, determining the product as an unqualified product, otherwise, determining the product as a qualified product.
Further, the step S2 specifically includes the following steps:
s21, moving the standard welding image in the welding image to be detected, and calculating a matching coefficient r between the standard welding image and the welding image to be detected;
s22, the position where the maximum value of the matching coefficient appears is regarded as that the coarse registration of the standard welding image and the welding image to be detected is completed;
and S23, carrying out binarization processing on the standard welding image and the welding image to be detected after coarse registration, and carrying out differential operation to obtain a difference region area Sc.
Further, the step S4 specifically includes the following steps:
s41, inputting the welding image to be detected into an improved convolution neural network, and performing convolution downsampling operation by adopting convolution kernels with different sizes to obtain a plurality of characteristic images with different resolutions;
s42, forming a pyramid shape by the plurality of feature images with different resolutions, wherein the higher the level is, the lower the resolution of the feature images is;
s43, performing exhaustive comparison on the top-layer low-resolution characteristic image according to the welding spot area of the standard welding image to obtain a coarse positioning area;
and S44, searching the characteristic images layer by layer according to the sequence from top to bottom, and only searching the position near the upper layer of the coarse positioning area until the bottommost layer to obtain a welding area to be detected which is accurately matched with the standard welding image.
Further, the step S6 specifically includes the following steps:
s61, carrying out gray processing on the target image to be detected to obtain a gray image;
s62, denoising the gray level image by adopting non-local mean filtering to obtain a denoised image;
s63, sharpening the edge part of the de-noised image by adopting a Laplace operator to obtain a preprocessed image with clear welding spot outline and support body layers.
Further, in step S8, a formula for calculating a shape difference value X between the shape line of the welding spot to be detected and a preset standard shape line of the welding spot is as follows:
wherein,
represents the corresponding ^ th or greater part of the welding spot form line to be detected and the standard welding spot form line>
Height difference of several high points->
And the number of corresponding high points on the welding spot form line to be detected and the standard welding spot form line is shown.
Further, the step S5 is configured to:
and performing fusion reconstruction on a plurality of feature images with different resolutions by adopting a multi-scale feature fusion strategy to obtain a high-resolution image, and segmenting the high-resolution image according to the welding area to be detected to obtain a target image to be detected only containing the welding area.
Further, after the step S8, the method further includes:
and S9, outputting the detection result information and generating an instruction for recalling the unqualified electronic component welding product.
The invention also provides a technical scheme that: a system for implementing the above-mentioned detection method for electronic component soldering, comprising:
the welding device comprises an image acquisition module, a welding module and a welding module, wherein the image acquisition module is used for acquiring a standard welding image acquired by image acquisition equipment and a welding image to be detected of an electronic element welding piece to be detected, and the standard welding image refers to an image which is qualified in welding quality and only comprises an electronic element and a welding area of the electronic element;
the first calculation module is used for calculating a matching coefficient r and a difference area Sc of the standard welding image and the welding image to be detected;
the device comprises a preliminary detection module, a comparison module and a comparison module, wherein the preliminary detection module is used for preliminarily detecting an electronic element welding product to be detected according to a preset matching coefficient threshold value Tr and a difference area threshold value Ts;
the welding area positioning module is used for searching and comparing a plurality of characteristic images with different resolutions of the welding image to be detected layer by layer according to the invariant moment characteristics of the image to obtain a welding area to be detected which is accurately matched with the standard welding image;
the target image to be detected determining module is used for carrying out image segmentation on the welding image to be detected according to the welding area to be detected to obtain a target image to be detected;
the image preprocessing module is used for preprocessing the target image to be detected to obtain a preprocessed image with a well-arranged welding spot outline and an electronic element support body;
the second calculation module is used for extracting a welding spot morphological line to be detected from the preprocessed image according to a segmentation threshold value and calculating a morphological difference value X between the welding spot morphological line to be detected and a preset standard welding spot morphological line;
and the secondary detection module is used for carrying out secondary detection according to a comparison result of the form difference value X and a preset detection threshold Tx, and judging whether a welding defect exists or not.
Further, the method also comprises the following steps:
the detection result output module is used for outputting detection result information and generating an instruction for recalling an unqualified electronic component welding product;
the high-resolution image fusion module is used for carrying out fusion reconstruction on a plurality of feature images with different resolutions by adopting a multi-scale feature fusion strategy so as to obtain a high-resolution image.
The invention also provides a computer device comprising a processor and a memory for storing a computer program which, when executed by the processor, implements the above-described method of detecting soldering of electronic components.
By means of the technical scheme, the invention provides a detection method, a system and computer equipment for electronic component welding, which at least have the following beneficial effects:
1. according to the method, the obtained welding image to be detected and the standard welding image are subjected to rough registration and differential operation according to the welding defect type of the electronic element, whether the welding product to be detected has the welding defect is judged by combining a preset threshold value, the product with the obvious welding defect is screened out, and the product without the obvious welding defect is subjected to secondary detection by adopting the improved convolutional neural network, so that the detection efficiency of the welding product of the electronic element is improved.
2. The method extracts a plurality of characteristic images with different resolutions of the welding image to be detected without obvious defects by adopting the improved convolutional neural network, uses a layer-by-layer searching exhaustive comparison strategy according to the invariant moment characteristics of the images, and performs image segmentation on the welding image to be detected according to the welding area to be detected which is accurately positioned and matched with the standard welding image to obtain the target image to be detected which only comprises the welding area, so that the calculation amount of subsequent image processing can be reduced, the welding quality detection efficiency is further improved, and the system maintenance cost is also reduced.
3. The method and the device perform feature fusion reconstruction on a plurality of feature images with different resolutions of a to-be-detected welding area which is accurately positioned and matched with the standard welding image by adopting a multi-scale feature fusion strategy, and segment the fused and reconstructed high-resolution image according to the to-be-detected welding area to obtain the to-be-detected target image only comprising the welding area, thereby achieving the purpose of improving the detection efficiency and accuracy of the welding quality of the electronic element on the premise of not increasing the detection cost.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:
FIG. 1 is a flow chart of a method for detecting soldering of electronic components according to one embodiment of the present invention;
FIG. 2 is a schematic diagram illustrating a comparison between a standard welding image and a welding image to be detected according to a first embodiment of the present invention;
FIG. 3 is a flowchart illustrating a process of searching for a welding area to be detected according to a first embodiment of the present invention;
FIG. 4 is a flowchart illustrating a pre-processing of a target image to be detected according to a first embodiment of the present invention;
FIG. 5 is a schematic block diagram of a system for detecting soldering of electronic components according to one embodiment of the present invention;
FIG. 6 is a flowchart of a method for detecting electronic component soldering according to a second embodiment of the present invention;
FIG. 7 is a block diagram of an improved convolutional neural network according to a second embodiment of the present invention;
FIG. 8 is a schematic structural diagram of multi-scale image search and fusion according to a second embodiment of the present invention;
FIG. 9 is a schematic block diagram of a system for detecting soldering of electronic components according to a second embodiment of the present invention;
fig. 10 is a block diagram showing an internal configuration of a computer device according to an embodiment of the present invention.
In the figure: 101. an image acquisition module; 102. a first calculation module; 103. a preliminary detection module; 104. a welding area positioning module to be detected; 105. a target image to be detected determining module; 106. an image preprocessing module; 107. a second calculation module; 108. a secondary detection module; 109. a detection result output module; 110. and a high-resolution image fusion module.
Detailed Description
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below. Therefore, the realization process of how to apply technical means to solve the technical problems and achieve the technical effects can be fully understood and implemented.
Example one
Referring to fig. 1 to 4, a specific implementation manner of the present embodiment is shown, in the present embodiment, according to a welding defect type of an electronic component, rough registration and difference operation are performed on an acquired welding image to be detected and a standard welding image, and a preset threshold is combined to determine whether a welding product to be detected has a missing welding defect, so as to screen out a product with an obvious welding defect, and then an improved convolutional neural network is used to perform secondary detection on a product without an obvious welding defect, thereby improving detection efficiency of the welding product of the electronic component.
As shown in fig. 1, a method for detecting soldering of electronic components includes the following steps:
s1, acquiring a standard welding image acquired by image acquisition equipment and a welding image to be detected of an electronic element welding piece to be detected.
In the embodiment, an array camera with the model of MV-CA060-10GC gigabit Ethernet industrial area is adopted to obtain a standard welding image and a welding image to be detected of an electronic element welding piece to be detected, wherein the standard welding image refers to an image which is detected by professionals to be qualified in welding quality and only comprises an electronic element and a welding area thereof; in addition, the camera has the highest resolution of 3072 × 2048, the maximum frame rate of 17FPS, the exposure time of 27 μ s to 2.5s, and two linear LED light sources to ensure the quality of the acquired image.
S2, calculating a matching coefficient r and a difference area Sc of the standard welding image and the welding image to be detected. The method specifically comprises the following steps:
s21, moving the standard welding image in the welding image to be detected, and calculating a matching coefficient r between the standard welding image and the welding image to be detected, wherein the calculation formula of the matching coefficient r is as follows:
in the above-mentioned formula, the compound has the following structure,
representing pixel points on the standard welding image,/>
a mean value of the gray levels representing a standard welding image,
represents a pixel point on the welding image to be detected, and>
and representing the gray average value of the welding image to be detected.
And S22, regarding the position where the maximum value of the matching coefficient appears as a standard welding image and finishing coarse registration with the welding image to be detected.
The value range of the absolute value of the matching coefficient r is [0,1], and the larger the value is, the higher the registration accuracy of the standard welding image and the welding image to be detected is, otherwise, the lower the registration accuracy is. For this reason, in the present embodiment, the position where the maximum value of the matching coefficient occurs is regarded as the position where the coarse registration of the standard welding image and the welding image to be detected is completed.
And S23, carrying out binarization processing on the coarsely registered standard welding image and the welding image to be detected, and carrying out differential operation to obtain a difference region area Sc.
The following will be further described with reference to the schematic diagram of comparison between the standard welding image and the welding image to be detected shown in fig. 2, where a picture a in fig. 2 represents the standard welding image and B picture represents the welding image to be detected;
firstly, moving the picture A in the picture B, calculating a matching coefficient between the picture A and the picture B according to a calculation formula of the matching coefficient r, and when the maximum value of the matching coefficient r appears, taking the position where the maximum value appears as the position where the coarse registration of the standard welding image and the welding image to be detected is completed; and then, carrying out binarization processing on the images A and B after coarse registration, and carrying out difference operation on the corresponding positions of the image B and the image A to obtain a difference region area Sc between the image B and the image A.
S3, performing primary detection on the electronic element welding product to be detected according to a preset matching coefficient threshold value Tr and a difference area threshold value Ts;
in the embodiment, the matching coefficient threshold Tr and the difference area threshold Ts are respectively set to 0.45 and 0.08 according to the characteristics of the welded electronic element, if | r | < 0.45 and | Sc | < 0.08, the welded electronic element is judged to be an unqualified product, the specific defect is lack of welding, namely no electronic element exists in the welding position, and the detection is finished, otherwise, the step S4 is executed to perform secondary detection on the welded product to be welded, and finally whether the welding defect exists is determined.
It is to be noted that the matching coefficient threshold value Tr and the difference area threshold value Ts are closely related to the type of electronic component to be soldered, and for this reason, the matching coefficient threshold value Tr and the difference area threshold value Ts may be set to other values.
And S4, searching and comparing a plurality of characteristic images with different resolutions of the welding image to be detected layer by layer according to the invariant moment characteristics of the image to obtain a welding area to be detected which is accurately matched with the standard welding image.
The following will be further described with reference to the flowchart of fig. 3, which specifically includes the following steps:
s41, inputting the welding image to be detected into an improved convolution neural network, and performing convolution downsampling operation by adopting convolution kernels with different sizes to obtain a plurality of characteristic images with different resolutions;
in the embodiment, the improved convolutional neural network comprises an input layer, ten convolutional layers, a down-sampling layer, a full-connection layer and an output layer, when characteristic images are extracted, firstly, a welding image to be detected enters the convolutional layers through the input layer, the convolutional layers are subjected to down-sampling operation through the first to third convolutional layers to generate four characteristic diagrams, and then, high-frequency, medium-frequency and low-frequency characteristic information in the images are respectively extracted from each characteristic diagram through three different frequency band passages to obtain four characteristic images with different resolutions; the high-frequency band path part uses three convolution layers, the middle-frequency band path part uses two convolution layers, convolution kernels of the convolution layers used in the high-frequency band and the middle-frequency band are both 3 multiplied by 3, and the low-frequency band path part uses a convolution layer with a convolution kernel of 5 multiplied by 5.
The process of extracting a plurality of characteristic images with different resolutions from a low-resolution welding image to be detected can be expressed as the following expression:
in the above formula, the first and second carbon atoms are,
represents a welding image to be detected, <' > based on>
Respectively representing the filter weight and the offset, the sign->
Represents a convolution operation, <' > or>
The activation function ReLU used on the output feature image is indicated.
S42, forming a pyramid shape by a plurality of characteristic images with different resolutions, wherein the higher the level is, the lower the resolution of the characteristic image is;
s43, performing exhaustive comparison on the top-layer low-resolution characteristic image according to the welding spot area of the standard welding image to obtain a coarse positioning area;
and S44, searching the characteristic images layer by layer according to the sequence from top to bottom, and only searching the position near the upper layer of coarse positioning area until the bottommost layer to obtain the welding area to be detected which is accurately matched with the standard welding image.
As an exemplary further explanation, assume that the welding area of the standard welding image in the present embodiment
And a first +>
Layer search area>
Are respectively->
And &>
) Then, in the searching and positioning process, the welding area is selected
And a fifth->
Layer search area pick>
The area with the greatest degree of matching is taken as the matching area of the next layer of feature image, i.e. the ^ th ^ or ^ th ^>
Search area for a layer image>
The selection range is as follows:
in the above-mentioned formula, the compound has the following structure,
indicates to->
A rectangular neighborhood of central coordinates, based on a predetermined number of pixels>
Respectively represent->
Is at>
Length in the direction->
Respectively, that a rectangular neighborhood is->
The coefficient of change of direction length.
The characteristic images with different resolutions extracted from the welding image to be detected are searched layer by repeating the process, the searching subgraph generated at each candidate position is compared with the standard welding image in an exhaustive mode, the size of the characteristic image at the top layer is small, the generated searching subgraphs are few, the welding area to be detected of the welding image to be detected can be conveniently and rapidly and accurately located, and therefore the detection efficiency of the welding quality of the electronic element is improved.
And S5, carrying out image segmentation on the welding image to be detected according to the welding area to be detected to obtain a target image to be detected.
The welding image to be detected is subjected to image segmentation according to the accurately positioned welding area to be detected matched with the standard welding image, so that the target image to be detected only containing the welding area is obtained, the calculation amount of subsequent image processing can be reduced, and the welding quality detection efficiency is improved.
And S6, preprocessing the target image to be detected to obtain a preprocessed image with distinct welding spot outline and electronic element support body level.
Because the contrast between the electronic element and the support body is not strong, the characteristic information of the welding spot is difficult to extract, therefore, a series of pretreatment is needed to be carried out on an image of a welding piece before the welding quality detection is carried out to eliminate external noise and interference, and the surface characteristic of the welding spot is highlighted, so that the subsequent extraction of the characteristic of the welding spot is carried out. As shown in fig. 4, the specific process of performing the pretreatment includes the following steps:
and S61, carrying out graying processing on the target image to be detected to obtain a grayscale image, so that the conditions of dark image and poor contrast can be improved.
It should be noted that the target image to be detected includes three channels, namely red, green and blue, and the data value describing the color of each channel is 0 to 255, and subsequently, three components need to be processed respectively, so that the computation amount is greatly increased, and the processing efficiency is reduced; in the welding quality detection process, only the outline of a welding spot and the position information of a component are required to be extracted, and the color of the welding spot is not required to be concerned, so that the image of the welding part needs to be subjected to graying processing, the component values of the red channel, the green channel and the blue channel of the grayscale image are equal, the original data is simplified, and the subsequent operation efficiency is improved.
S62, denoising the gray level image by adopting non-local mean filtering to obtain a denoised image;
the uneven reflection of the surface coating layer of the electronic element support body and the influence of the surrounding environment and image acquisition equipment of a welding image to be detected in the acquisition process enable the welding image to be detected to have various noises and distortions.
S63, sharpening the edge part of the de-noised image by adopting a Laplace operator to obtain a preprocessed image with clear welding spot outline and support body layers.
After denoising processing, the image to be detected is more fuzzy than the original image, and the subsequent extraction of welding characteristics is inconvenient, so that the edge part of the denoised image, the shape of a welding area and the gray level jump part need to be sharpened by adopting a Laplace operator to obtain a preprocessed image with a well-arranged welding spot outline and a support body.
Through a series of preprocessing on the welding part image, the welding point outline is not lost, and the contrast between the electronic element and the supporting body is enhanced, so that a data base is laid for the subsequent electronic element welding quality detection.
And S7, extracting the shape lines of the welding spots to be detected from the preprocessed image according to the segmentation threshold, and calculating the shape difference value X between the shape lines of the welding spots to be detected and the preset standard shape lines of the welding spots.
In this embodiment, the preprocessed image is divided into two categories, namely a foreground category and a background category, variance values of the two categories are calculated through global search, so that a gray value with the largest variance value between the two categories is used as an optimal segmentation threshold, and finally, a shape line of the welding spot to be detected is extracted from the preprocessed image according to the segmentation threshold.
Calculating a shape difference value X between the shape line of the welding spot to be detected and the shape line of the preset standard welding spot according to a calculation formula, wherein the formula of the shape difference value X is as follows:
wherein,
represents the corresponding ^ th or greater part of the welding spot form line to be detected and the standard welding spot form line>
Height difference of several high points->
And the number of corresponding high points on the welding spot form line to be detected and the standard welding spot form line is shown.
S8, performing secondary detection according to a comparison result of the form difference value X and a preset detection threshold Tx, and judging whether a welding defect exists or not;
if the form difference value X exceeds the range of the preset detection threshold value Tx, determining the product as unqualified, otherwise, determining the product as qualified.
And S9, outputting the detection result information and generating an instruction for recalling the unqualified electronic component welding product.
So that the staff recalls the unqualified electronic product of welding according to the recall instruction, improve the product percent of pass to accord with modern enterprise's actual production demand more.
Referring to fig. 5, the present embodiment further provides a system for implementing the method for detecting soldering of an electronic component, including:
the image acquisition module 101 is configured to acquire a standard welding image acquired by an image acquisition device and a welding image to be detected of an electronic component welding part to be detected, where the standard welding image is an image that is qualified in welding quality and only includes an electronic component and a welding area thereof;
the first calculation module 102 is configured to calculate a matching coefficient r and a difference area Sc of the standard welding image and the welding image to be detected;
the preliminary detection module 103 is used for preliminarily detecting the electronic element welding product to be detected according to a preset matching coefficient threshold Tr and a difference area threshold Ts;
the welding area positioning module 104 is configured to perform layer-by-layer search and comparison on a plurality of characteristic images with different resolutions of the welding image to be detected according to the invariant moment characteristic of the image, so as to obtain a welding area to be detected, which is accurately matched with the standard welding image;
the target image to be detected determining module 105 is used for performing image segmentation on the welding image to be detected according to the welding area to be detected to obtain a target image to be detected;
the image preprocessing module 106 is used for preprocessing a target image to be detected to obtain a preprocessed image with clear welding spot outline and electronic element support body layers;
the second calculating module 107 is configured to extract a to-be-detected welding spot morphological line from the preprocessed image according to the segmentation threshold, and calculate a morphological difference value X between the to-be-detected welding spot morphological line and a preset standard welding spot morphological line;
the secondary detection module 108 is configured to perform secondary detection on the welding product to be detected according to a comparison result between the form difference value X and a preset detection threshold Tx, and determine whether a welding defect exists;
and the detection result output module 109 is used for outputting the detection result information and generating an instruction for recalling the unqualified electronic component welding product.
According to the embodiment, firstly, according to the welding defect type of the electronic element, rough registration and differential operation are carried out on an acquired welding image to be detected and a standard welding image, whether the welding product to be detected has the welding defect is judged by combining a preset threshold value, the product with the obvious welding defect is screened out, then, a plurality of characteristic images with different resolutions are extracted from the welding image to be detected by adopting an improved convolutional neural network, a welding area to be detected matched with the standard welding image is located by using a layer-by-layer searching strategy, and then the welding image to be detected is segmented according to the located welding area to be detected to obtain a target image to be detected.
Example two
The implementation provided by this embodiment is made on the basis of the first embodiment, and the same portions can solve the same technical problems and have the same beneficial effects, which are referred to each other, and detailed description will not be repeated in this embodiment.
Referring to fig. 6 to 8, a specific implementation manner of the present embodiment is shown, in which a plurality of feature images with different resolutions are extracted from an acquired low-resolution to-be-detected welding image, and the feature images with different resolutions are fused and reconstructed to obtain a high-resolution image, so as to achieve the purpose of improving the accuracy of detecting the welding quality of an electronic component without increasing the detection cost.
As shown in fig. 6, a method for detecting soldering of electronic components includes the steps of:
s1, acquiring a standard welding image acquired by image acquisition equipment and a welding image to be detected of an electronic element welding piece to be detected, wherein the standard welding image refers to an image which is qualified in welding quality and only comprises an electronic element and a welding area of the electronic element.
And S2, calculating a matching coefficient r and a difference area Sc of the standard welding image and the welding image to be detected.
S3, performing primary detection on the electronic element welding product to be detected according to a preset matching coefficient threshold value Tr and a difference area threshold value Ts;
if the | r | is less than Tr and the | Sc | is less than Ts, determining that the product is unqualified and finishing the detection, otherwise, executing the step S4.
And S4, searching and comparing a plurality of characteristic images with different resolutions of the welding image to be detected layer by layer according to the invariant moment characteristics of the image to obtain a welding area to be detected which is accurately matched with the standard welding image.
And S51, carrying out fusion reconstruction on a plurality of feature images with different resolutions by adopting a multi-scale feature fusion strategy to obtain a high-resolution image.
As shown in fig. 7, the improved convolutional neural network involved in this embodiment includes one input layer, ten convolutional layers and downsampling layers, one fully-connected layer, and one output layer. Firstly, introducing an improved convolutional neural network into a welding image to be detected through an input layer, performing convolutional downsampling operation through first to third convolutional layers to generate four characteristic graphs, and extracting medium-high frequency, medium-frequency and low-frequency characteristic information of the image through three different frequency band channels for each characteristic graph to obtain four characteristic images with different resolutions, wherein three convolutional layers are used in a high-frequency band channel part, two convolutional layers are used in a medium-frequency band channel part, the convolutional kernels of the convolutional layers used in the high-frequency band and the medium-frequency band are both 3 x 3, and one convolutional layer with the convolutional kernel of 5 x 5 is used in a low-frequency band channel part; then, a multi-scale image searching and fusing unit (shown in fig. 8) is used for searching the extracted characteristic images with different resolutions layer by layer, after a welding area to be detected of the welding image to be detected is accurately positioned, a nonlinear mapping mode is used for carrying out fusion reconstruction on the characteristic images with different resolutions so as to obtain a high-resolution image.
The specific process of reconstructing a high resolution image by fusion will be further described with reference to the structural diagram of multi-scale image search and fusion in fig. 8:
firstly, adopting maximum pooling downsampling for a first-grade characteristic image and calculating by a 1 multiplied by 1 convolution kernel;
secondly, after the secondary feature map is subjected to 1 multiplied by 1 convolution kernel to reduce the channel dimension and simplify the calculation, extracting the features and performing transverse link operation;
thirdly, fusing the characteristic graphs obtained in the first step and the second step and performing linear interpolation twice upsampling operation on a fused result;
fourthly, fusing the three feature images with different resolutions and the same size obtained in the first step to the third step to enhance the semantic information of the feature images, and calculating the fusion result by a 3 multiplied by 3 convolution kernel to eliminate the aliasing effect generated by multiple times of fusion;
and fifthly, repeating the first step to the fourth step until the feature extraction is completed, and obtaining a high-resolution image.
It should be noted that, the feature images with different resolutions from one level to four levels in fig. 8 correspond to Conv1 to Conv4_ x in the Resnet101 neural network, and the semantic information extracted from the images with different resolutions is effectively used to generate a final high-resolution image, so that the accuracy of the welding quality detection result is improved, and the misjudgment rate is reduced.
And S52, segmenting the high-resolution image according to the welding area to be detected to obtain a target image to be detected only comprising the welding area.
After the high-resolution image is obtained, the fused high-resolution image is segmented according to the welding area to be detected and the segmentation threshold value to obtain a target image to be detected only containing the welding area, and the purpose of improving the detection accuracy of the welding quality of the electronic element on the premise of not increasing the detection cost is achieved.
And S6, preprocessing the target image to be detected to obtain a preprocessed image with distinct welding spot outlines and electronic element support body layers.
S7, extracting a welding spot shape line to be detected from the preprocessed image according to the segmentation threshold, and calculating a shape difference value X between the welding spot shape line to be detected and a preset standard welding spot shape line;
s8, performing secondary detection according to a comparison result of the form difference value X and a preset detection threshold Tx, and judging whether welding defects exist or not;
if the form difference value X exceeds the range of the preset detection threshold Tx, determining the product as unqualified, otherwise determining the product as qualified.
And S9, outputting the detection result information and generating an instruction for recalling the unqualified electronic component welding product.
Referring to fig. 9, the present embodiment further provides a system for implementing the method for detecting soldering of an electronic component, including:
the image acquisition module 101 is configured to acquire a standard welding image acquired by an image acquisition device and a welding image to be detected of an electronic component welding part to be detected, where the standard welding image is an image that is qualified in welding quality and only includes an electronic component and a welding area thereof;
the first calculation module 102 is configured to calculate a matching coefficient r and a difference area Sc of the standard welding image and the welding image to be detected according to a calculation formula;
the preliminary detection module 103 is used for preliminarily detecting the electronic element welding product to be detected according to a preset matching coefficient threshold value Tr and a difference area threshold value Ts;
the welding area positioning module 104 is configured to perform layer-by-layer search and comparison on a plurality of characteristic images with different resolutions of the welding image to be detected according to the invariant moment characteristic of the image, so as to obtain a welding area to be detected, which is accurately matched with the standard welding image;
the high-resolution image fusion module 110 is configured to perform fusion reconstruction on a plurality of feature images with different resolutions by using a multi-scale feature fusion strategy to obtain a high-resolution image;
the target image to be detected determining module 105 is used for performing image segmentation on the high-resolution image according to the welding area to be detected to obtain a target image to be detected;
the image preprocessing module 106 is used for preprocessing a target image to be detected to obtain a preprocessed image with clear welding spot outline and electronic element support body layers;
a second calculating module 107, configured to extract a welding spot shape line to be detected from the preprocessed image according to the segmentation threshold, and calculate a shape difference value X between the welding spot shape line to be detected and a preset standard welding spot shape line;
the secondary detection module 108 is configured to perform secondary detection according to a comparison result between the form difference value X and a preset detection threshold Tx, and determine whether a welding defect exists;
and the detection result output module 109 is used for outputting the detection result information and generating an instruction for recalling the unqualified electronic component welding product.
According to the embodiment, the improved convolutional neural network is adopted to extract a plurality of characteristic images with different resolutions from the welding image to be detected, after the welding area to be detected matched with the standard welding image is located by using the layer-by-layer searching strategy, the characteristic fusion reconstruction is carried out on the plurality of characteristic images with different resolutions by using the multi-scale characteristic fusion strategy to obtain the high-resolution image, the fusion reconstructed high-resolution image is segmented according to the welding area to be detected, the target image to be detected which only comprises the welding area is obtained, and the purpose of improving the detection efficiency and accuracy of the welding quality of the electronic element on the premise of not increasing the detection cost is achieved.
The present embodiment also provides a computer apparatus, and the internal structure of the computer apparatus including a processor, a memory, a network interface, a display screen, and an input device connected via a system bus will be described below with reference to fig. 10.
The processor is a control center of the computer equipment, is connected with each part of the whole equipment by various interfaces and lines, and realizes various functions of the computer equipment by running or executing computer readable instructions and/or modules stored in the memory and calling data stored in the memory; the device can be a Central Processing Unit (CPU), and also can be other general processors, digital Signal Processors (DSPs), application Specific Integrated Circuits (ASICs), ready-made programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components and the like; wherein a general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The memory is used for storing computer readable instructions and/or modules, and mainly comprises a storage medium and an internal memory, wherein the storage medium can be a non-volatile storage medium or a volatile storage medium, the storage medium stores an operating system, and also can store computer readable instructions, and when the computer readable instructions are executed by the processor, the processor can realize the detection method of the welding of the electronic component. For example, steps S1 to S9 shown in fig. 1 and 6 and other extensions of the method and extensions of the related steps. Alternatively, the processor executes the computer readable instructions to implement the functions of the modules/units of the detection system for electronic component soldering in the above embodiments, such as the functions of the modules 101 to 109 shown in fig. 5, which are not described herein again for avoiding repetition.
The network interface is used for connecting and communicating with an external server through a network; the display screen can be a liquid crystal display screen or an electronic ink display screen; the input device may be a touch layer covered on a display screen, or may be a key, a track ball or a touch pad arranged on a casing of the computer device, or may be an external keyboard, a touch pad or a mouse, etc.
It should be noted that the memory may be integrated into the processor or provided separately from the processor, and the structure shown in fig. 10 is only a schematic block diagram of a part of the structure related to the present application, and does not constitute a limitation to the computer device to which the present application is applied, and a specific computer device may include more or less components than those shown in the figure, or combine some components, or adopt different component arrangements.
The above-mentioned serial numbers of the embodiments of the present application are merely for description and do not represent the merits of the embodiments. Through the description of the above embodiments, it is obvious to those skilled in the art that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better embodiment. Based on such understanding, the technical solution of the present application may be substantially or partially embodied in the form of a software product, which is stored in a storage medium (e.g., ROM/RAM, magnetic disk, optical disk) as described above and includes instructions for enabling a terminal device (e.g., a mobile phone, a computer, a server, or a network device) to execute the method according to the embodiments of the present application.
The present invention has been described in detail with reference to the foregoing embodiments, and the principles and embodiments of the present invention have been described herein with reference to specific examples, which are provided only to assist understanding of the methods and core concepts of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.
Claims (10)
1. A method for detecting the welding of electronic components is characterized by comprising the following steps:
s1, acquiring a standard welding image acquired by image acquisition equipment and a welding image to be detected of an electronic element welding piece to be detected, wherein the standard welding image refers to an image which is qualified in welding quality and only comprises an electronic element and a welding area of the electronic element;
s2, calculating a matching coefficient r and a difference area Sc of the standard welding image and the welding image to be detected;
s3, performing primary detection on the electronic element welding product to be detected according to a preset matching coefficient threshold value Tr and a difference area threshold value Ts;
if the | r | < Tr and the | Sc | < Ts, judging that the product is unqualified and finishing the detection, otherwise, executing the step S4;
s4, searching and comparing a plurality of characteristic images with different resolutions of the welding image to be detected layer by layer according to the invariant moment characteristics of the image to obtain a welding area to be detected which is accurately matched with the standard welding image;
s5, performing image segmentation on the welding image to be detected according to the welding area to be detected to obtain a target image to be detected;
s6, preprocessing the target image to be detected to obtain a preprocessed image with a well-arranged welding spot outline and an electronic element support body;
s7, extracting a welding spot shape line to be detected from the preprocessed image according to a segmentation threshold, and calculating a shape difference value X between the welding spot shape line to be detected and a preset standard welding spot shape line;
s8, performing secondary detection according to a comparison result of the form difference value X and a preset detection threshold Tx, and judging whether a welding defect exists or not;
and if the form difference value X exceeds the range of a preset detection threshold Tx, judging the product as an unqualified product, otherwise, judging the product as a qualified product.
2. The method for detecting the soldering of electronic components according to claim 1, wherein the step S2 specifically comprises the steps of:
s21, moving the standard welding image in the welding image to be detected, and calculating a matching coefficient r between the standard welding image and the welding image to be detected;
s22, the position where the maximum value of the matching coefficient appears is regarded as that the coarse registration of the standard welding image and the welding image to be detected is completed;
and S23, carrying out binarization processing on the standard welding image and the welding image to be detected after coarse registration, and carrying out differential operation to obtain a difference region area Sc.
3. The method for detecting the soldering of electronic components according to claim 1, wherein the step S4 specifically comprises the steps of:
s41, inputting the welding image to be detected into an improved convolution neural network, and performing convolution downsampling operation by adopting convolution kernels with different sizes to obtain a plurality of characteristic images with different resolutions;
s42, forming a plurality of characteristic images with different resolutions into a pyramid shape, wherein the higher the level is, the lower the resolution of the characteristic images is;
s43, performing exhaustive comparison on the top-layer low-resolution characteristic image according to the welding spot area of the standard welding image to obtain a coarse positioning area;
and S44, searching the characteristic images layer by layer according to the sequence from top to bottom, and only searching the position near the upper layer of the coarse positioning area until the bottommost layer to obtain a welding area to be detected which is accurately matched with the standard welding image.
4. The method for detecting the soldering of electronic components according to claim 1, wherein the step S6 specifically comprises the steps of:
s61, carrying out gray processing on the target image to be detected to obtain a gray image;
s62, denoising the gray level image by adopting non-local mean filtering to obtain a denoised image;
s63, sharpening the edge part of the de-noised image by adopting a Laplace operator to obtain a preprocessed image with clear welding spot outline and support body layers.
5. The method for detecting the soldering of an electronic component according to claim 1, wherein in step S8, the formula for calculating the form difference value X between the form line of the solder joint to be detected and the form line of the preset standard solder joint is as follows:
wherein,
represents the corresponding ^ th or greater part of the welding spot form line to be detected and the standard welding spot form line>
Height difference of several high points->
And the number of corresponding high points on the welding spot form line to be detected and the standard welding spot form line is shown.
6. The method for detecting soldering of electronic components according to claim 3, wherein the step S5 is provided as:
and performing fusion reconstruction on a plurality of feature images with different resolutions by adopting a multi-scale feature fusion strategy to obtain a high-resolution image, and segmenting the high-resolution image according to the welding area to be detected to obtain a target image to be detected only containing the welding area.
7. The method for detecting soldering of electronic components according to claim 1, wherein the step S8 is followed by further comprising:
and S9, outputting the detection result information and generating an instruction for recalling the unqualified electronic component welding product.
8. A system for implementing the method for detecting soldering of electronic components as claimed in any one of claims 1 to 7, comprising:
the welding device comprises an image acquisition module (101), wherein the image acquisition module (101) is used for acquiring a standard welding image acquired by image acquisition equipment and a welding image to be detected of an electronic element welding part to be detected, wherein the standard welding image refers to an image which is qualified in welding quality and only comprises the electronic element and a welding area of the electronic element;
the first calculation module (102), the first calculation module (102) is used for calculating a matching coefficient r and a difference area Sc of the standard welding image and the welding image to be detected;
the device comprises a preliminary detection module (103), wherein the preliminary detection module (103) is used for preliminarily detecting an electronic element welding product to be detected according to a preset matching coefficient threshold value Tr and a difference area threshold value Ts;
the welding area locating module (104) is used for searching and comparing a plurality of characteristic images with different resolutions of the welding image to be detected layer by layer according to the invariant moment characteristic of the image to obtain a welding area to be detected which is accurately matched with the standard welding image;
the target image to be detected determining module (105), the target image to be detected determining module (105) is used for carrying out image segmentation on the welding image to be detected according to the welding area to be detected to obtain a target image to be detected;
the image preprocessing module (106) is used for preprocessing the target image to be detected to obtain a preprocessed image with distinct welding spot outline and electronic element support body level;
the second calculation module (107) is used for extracting a welding spot shape line to be detected from the preprocessed image according to a segmentation threshold value, and calculating a shape difference value X between the welding spot shape line to be detected and a preset standard welding spot shape line;
and the secondary detection module (108) is used for carrying out secondary detection according to the comparison result of the form difference value X and a preset detection threshold Tx, and judging whether a welding defect exists or not.
9. The system of claim 8, further comprising:
the detection result output module (109), the detection result output module (109) is used for outputting detection result information and generating an instruction for recalling the unqualified electronic component welding product;
the high-resolution image fusion module (110) is used for carrying out fusion reconstruction on a plurality of feature images with different resolutions by adopting a multi-scale feature fusion strategy so as to obtain a high-resolution image.
10. A computer device comprising a processor and a memory for storing a computer program which, when executed by the processor, implements the method of detection of soldering of electronic components according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310143229.XA CN115880288B (en) | 2023-02-21 | 2023-02-21 | Detection method, system and computer equipment for electronic element welding |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310143229.XA CN115880288B (en) | 2023-02-21 | 2023-02-21 | Detection method, system and computer equipment for electronic element welding |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115880288A true CN115880288A (en) | 2023-03-31 |
| CN115880288B CN115880288B (en) | 2023-11-14 |
Family
ID=85761408
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310143229.XA Active CN115880288B (en) | 2023-02-21 | 2023-02-21 | Detection method, system and computer equipment for electronic element welding |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115880288B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117746077A (en) * | 2024-02-20 | 2024-03-22 | 深圳超盈智能科技有限公司 | Chip defect detection method, device, equipment and storage medium |
| CN117938626A (en) * | 2024-01-24 | 2024-04-26 | 南京脉期尔电气有限公司 | Network warning system triggered by applying power-on detection |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110517265A (en) * | 2019-09-04 | 2019-11-29 | 北京理工大学 | A kind of detection method of surface defects of products, device and storage medium |
| CN112116557A (en) * | 2020-08-12 | 2020-12-22 | 西安交通大学 | Radiographic image weld area defect detection method, storage medium and equipment |
| CN113362326A (en) * | 2021-07-26 | 2021-09-07 | 广东奥普特科技股份有限公司 | Method and device for detecting welding spot defects of battery |
| CN113379678A (en) * | 2021-05-14 | 2021-09-10 | 珠海格力智能装备有限公司 | Circuit board detection method and device, electronic equipment and storage medium |
| CN113888746A (en) * | 2020-07-01 | 2022-01-04 | 上海电力大学 | Welding seam image mode recognition method based on combination of Hu invariant moment and SVM |
| CN114240939A (en) * | 2022-02-24 | 2022-03-25 | 苏州浪潮智能科技有限公司 | Method, system, equipment and medium for detecting appearance defects of mainboard components |
| CN114693633A (en) * | 2022-03-28 | 2022-07-01 | 深圳市深视智能科技有限公司 | Welding defect detection method, device and system |
-
2023
- 2023-02-21 CN CN202310143229.XA patent/CN115880288B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110517265A (en) * | 2019-09-04 | 2019-11-29 | 北京理工大学 | A kind of detection method of surface defects of products, device and storage medium |
| CN113888746A (en) * | 2020-07-01 | 2022-01-04 | 上海电力大学 | Welding seam image mode recognition method based on combination of Hu invariant moment and SVM |
| CN112116557A (en) * | 2020-08-12 | 2020-12-22 | 西安交通大学 | Radiographic image weld area defect detection method, storage medium and equipment |
| CN113379678A (en) * | 2021-05-14 | 2021-09-10 | 珠海格力智能装备有限公司 | Circuit board detection method and device, electronic equipment and storage medium |
| CN113362326A (en) * | 2021-07-26 | 2021-09-07 | 广东奥普特科技股份有限公司 | Method and device for detecting welding spot defects of battery |
| CN114240939A (en) * | 2022-02-24 | 2022-03-25 | 苏州浪潮智能科技有限公司 | Method, system, equipment and medium for detecting appearance defects of mainboard components |
| CN114693633A (en) * | 2022-03-28 | 2022-07-01 | 深圳市深视智能科技有限公司 | Welding defect detection method, device and system |
Non-Patent Citations (1)
| Title |
|---|
| 陈智强;楼佩煌;钱晓明;: "硅电池片自动串焊表面缺陷在线视觉检测研究", 机电工程, no. 07 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117938626A (en) * | 2024-01-24 | 2024-04-26 | 南京脉期尔电气有限公司 | Network warning system triggered by applying power-on detection |
| CN117746077A (en) * | 2024-02-20 | 2024-03-22 | 深圳超盈智能科技有限公司 | Chip defect detection method, device, equipment and storage medium |
| CN117746077B (en) * | 2024-02-20 | 2024-05-10 | 深圳超盈智能科技有限公司 | Chip defect detection method, device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115880288B (en) | 2023-11-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115880288B (en) | Detection method, system and computer equipment for electronic element welding | |
| CN109785291B (en) | Lane line self-adaptive detection method | |
| CN111915704A (en) | Apple hierarchical identification method based on deep learning | |
| CN106846339A (en) | Image detection method and device | |
| CN114742799B (en) | Industrial scene unknown type defect segmentation method based on self-supervision heterogeneous network | |
| CN115830004A (en) | Surface defect detection method, device, computer equipment and storage medium | |
| CN116563237B (en) | Deep learning-based chicken carcass defect hyperspectral image detection method | |
| Ma et al. | An automatic detection method of Mura defects for liquid crystal display | |
| CN111428753B (en) | Training set acquisition method, electric power facility detection method and device | |
| CN115170552A (en) | Image quality evaluation method, device, equipment and storage medium | |
| CN115830031A (en) | Method and system for detecting circuit board patch and storage medium | |
| CN117152165A (en) | Photosensitive chip defect detection method and device, storage medium and electronic equipment | |
| CN117152182B (en) | Ultralow-illumination network camera image processing method and device and electronic equipment | |
| CN112833821B (en) | Differential geometric three-dimensional micro-vision detection system and method for high-density IC welding spots | |
| CN113674260A (en) | SMT welding spot defect detection method | |
| CN116542974B (en) | Method for detecting surface defects of copper-clad plate based on multi-scale gridding | |
| CN117541572A (en) | PCB component defect detection method | |
| CN110349129B (en) | Appearance defect detection method for high-density flexible IC substrate | |
| CN111415365B (en) | Image detection method and device | |
| CN113378847B (en) | Character segmentation method, system, computer device and storage medium | |
| CN107123105A (en) | Images match defect inspection method based on FAST algorithms | |
| CN111985535A (en) | Method and device for optimizing human body depth map through neural network | |
| Matsui et al. | Low-light image enhancement using a simple network structure | |
| CN113034432A (en) | Product defect detection method, system, device and storage medium | |
| CN110599398A (en) | Online image splicing and fusing method based on wavelet technology |
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 |