Back adhesive quality detection method, device and system

Abstract

The invention provides a gum quality detection method, device and system, and relates to the technical field of gum quality detection.The method comprises the steps that an ultrasonic feature vector QA is obtained; obtaining a similarity list lambda; the lambda is traversed, and if lambda i is larger than rhoi, it is determined that the overall gum area of the to-be-detected component meets a preset qualified condition; otherwise, generating an overall to-be-detected gum image TE corresponding to the overall gum area according to an ultrasonic signal acquired by the phased array ultrasonic probe; extracting an area with gum in the TE to obtain a sub to-be-detected gum image list A corresponding to the TE; splicing the sub-ultrasonic feature vector corresponding to the Aj and the sub-image feature vector to obtain a mixed feature vector Hj corresponding to the Aj; according to the Hj, determining whether the quality of the overall gum area of the to-be-detected component meets a preset qualification condition; on the basis of improving the detection accuracy, the computing power consumption is reduced, and the efficiency of detecting the gum quality is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of adhesive backing quality detection, and in particular to a method, device and system for detecting adhesive backing quality. Background Art

[0002] In the field of electronic products, adhesive backing is usually used to fix components such as screens and batteries. Whether the adhesive backing has problems such as bubbles, breakage, and unevenness will directly affect the overall quality of the electronic product. For example, bubbles will cause the effective contact area between the adhesive backing and the substrate to decrease, and the local adhesion force will decrease, which may cause the component to loosen. The breakage of the adhesive backing will directly destroy the overall force balance. Especially when subjected to vibration and impact, the breakage point can easily become a stress concentration point, causing the component to fall off. In the existing technology, the detection of the adhesive backing is usually carried out through image processing, and the image of the adhesive backing is analyzed to determine whether there is a problem. However, this detection method has very high requirements on computing power and low detection efficiency. Summary of the Invention

[0003] In view of the above technical problems, the technical solution adopted by the present invention is:

[0004] According to a first aspect of the present application, a method for detecting adhesive quality is provided, the method comprising the following steps:

[0005] S100, in response to the component to be inspected reaching a preset position, obtaining an ultrasonic feature vector QA corresponding to the entire adhesive area of the component to be inspected; wherein QA is obtained by a preset phased array ultrasonic probe;

[0006] S200, obtaining the similarity between QA and each preset qualified ultrasonic feature vector to obtain a similarity list λ = (λ1, λ2, ..., λ i ,…,λ n ), i=1, 2,...,n; where, λ i is the similarity between QA and the i-th qualified ultrasonic feature vector, and n is the number of qualified ultrasonic feature vectors;

[0007] S300, traverse λ, if λ i >ρ i , it is determined that the overall adhesive area of the component to be inspected meets the preset qualification conditions; otherwise, the overall adhesive image TE corresponding to the overall adhesive area to be inspected is generated according to the ultrasonic signal collected by the phased array ultrasonic probe; ρ i is the similarity threshold corresponding to the i-th qualified ultrasonic feature vector;

[0008] S400, extracting the adhesive area in TE to obtain the sub-adhesive image list to be detected corresponding to TE = (A1, A2, ..., A j ,…,Am ), j = 1, 2, ..., m; where A j is the jth sub-adhesive image to be detected, and m is the number of sub-adhesive images to be detected;

[0009] S500, A j The corresponding sub-ultrasound feature vector is concatenated with the sub-image feature vector to obtain A j The corresponding mixed eigenvector H j ; A j The corresponding sub-ultrasonic eigenvector is obtained by A j The ultrasonic signal collected by the corresponding ultrasonic probe is obtained;

[0010] S600, according to H j , determine whether the quality of the overall adhesive area of the component to be inspected meets the preset qualification conditions.

[0011] According to another aspect of the present application, a device for detecting adhesive quality is provided, the device comprising:

[0012] An ultrasonic feature vector acquisition module is configured to acquire an ultrasonic feature vector QA corresponding to the entire adhesive area of the component to be inspected in response to the component to be inspected reaching a preset position; wherein QA is obtained by a preset phased array ultrasonic probe;

[0013] The similarity determination module is used to obtain the similarity between QA and each preset qualified ultrasonic feature vector to obtain a similarity list λ = (λ1, λ2, ..., λ i ,…,λ n ), i=1, 2,...,n; where, λ i is the similarity between QA and the i-th qualified ultrasonic feature vector, and n is the number of qualified ultrasonic feature vectors;

[0014] The qualified judgment module is used to traverse λ. If λ i >ρ i , it is determined that the overall adhesive area of the component to be inspected meets the preset qualification conditions; otherwise, the overall adhesive image TE corresponding to the overall adhesive area to be inspected is generated according to the ultrasonic signal collected by the phased array ultrasonic probe; ρ i is the similarity threshold corresponding to the i-th qualified ultrasonic feature vector;

[0015] The sub-to-be-detected adhesive image extraction module is used to extract the adhesive area in TE to obtain the sub-to-be-detected adhesive image list A corresponding to TE = (A1, A2, ..., A j ,…,A m ), j = 1, 2, ..., m; where A jis the jth sub-adhesive image to be detected, and m is the number of sub-adhesive images to be detected;

[0016] Feature vector concatenation module, used to concatenate A j The corresponding sub-ultrasound feature vector is concatenated with the sub-image feature vector to obtain A j The corresponding mixed eigenvector H j ; A j The corresponding sub-ultrasonic eigenvector is obtained by A j The ultrasonic signal collected by the corresponding ultrasonic probe is obtained;

[0017] Quality judgment module, used to judge the quality of j , determine whether the quality of the overall adhesive area of the component to be inspected meets the preset qualification conditions.

[0018] According to another aspect of the present application, a back glue quality detection system is also provided, including the back glue quality detection device mentioned above.

[0019] The present invention has at least the following beneficial effects:

[0020] The adhesive quality inspection method of this invention incorporates ultrasonic feature vectors for initial screening. This allows for rapid identification of the majority of qualified products during the feature matching phase, initiating the image analysis process only for samples suspected of defects. This hierarchical processing mechanism effectively reduces the frequency of image processing module calls, focusing system computing resources on critical problem areas. This significantly reduces the overall computational load compared to traditional solutions, making it particularly suitable for high-speed production line inspection scenarios.

[0021] Furthermore, a multi-dimensional defect detection system has been constructed; existing technologies are limited by two-dimensional image features and have difficulty identifying deep structural defects. This invention, through the collaborative analysis of ultrasonic feature vectors and image features, not only captures surface morphological anomalies but also reveals hidden defects such as bubbles and delamination within the adhesive layer through the attenuation characteristics of the ultrasonic signal. In particular, when hybrid feature vectors are used for local area analysis, multi-dimensional quality indicators such as adhesive layer thickness uniformity and bonding interface integrity can be simultaneously evaluated, significantly improving the defect detection rate. This improves detection accuracy while reducing computing power consumption and increasing detection efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0023] Figure 1This is a flow chart of a method for detecting adhesive quality provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0024] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present invention.

[0025] It should be noted that, based on this disclosure, those skilled in the art will appreciate that an aspect described herein can be implemented independently of any other aspect, and that two or more of these aspects can be combined in various ways. For example, any number of the aspects described herein can be used to implement an apparatus and / or practice a method. In addition, other structures and / or functionalities other than one or more of the aspects described herein can be used to implement such an apparatus and / or practice such a method.

[0026] The following will refer to Figure 1 The flowchart of the adhesive backing quality detection method shown in the figure introduces a adhesive backing quality detection method.

[0027] The adhesive quality detection method may include the following steps:

[0028] S100 , in response to the component to be inspected reaching a preset position, obtaining an ultrasonic feature vector QA corresponding to the entire adhesive area of the component to be inspected; wherein QA is obtained by a preset phased array ultrasonic probe.

[0029] In this embodiment, the component to be detected can be a mobile phone case with several pieces of adhesive backing. When the component to be detected reaches the preset position, an ultrasonic signal is emitted by a preset phased array ultrasonic probe and an ultrasonic signal returned by the entire adhesive backing area is received. The received ultrasonic signal is encoded and feature extracted to obtain QA; the phased array ultrasonic probe includes several ultrasonic probes, and the ultrasonic signal collected by each ultrasonic probe can be encoded and feature extracted separately to obtain a sub-ultrasonic feature vector corresponding to each ultrasonic probe, and then each sub-ultrasonic feature vector is spliced to obtain QA.

[0030] Phased array ultrasonic waves can emit sound waves at multiple angles, penetrate the adhesive layer and receive reflected signals. Signal characteristics (such as amplitude, phase, and propagation time) reflect the internal structure of the adhesive (whether there are defects such as bubbles and fractures). Phased array ultrasonic probes can quickly and non-contactly obtain the physical properties of the entire adhesive, making it suitable for real-time detection on automated production lines.

[0031] S200, obtaining the similarity between QA and each preset qualified ultrasonic feature vector to obtain a similarity list λ = (λ1, λ2, ..., λ i ,…,λ n ), i=1, 2,...,n; where, λ i is the similarity between QA and the i-th qualified ultrasonic feature vector, and n is the number of qualified ultrasonic feature vectors.

[0032] In this embodiment, cosine similarity, Euclidean distance, etc. can be used to measure the degree of match between the features to be tested and the qualified features. A benchmark is established by using historical qualified data to quickly screen out obviously qualified or unqualified parts.

[0033] Furthermore, the preset qualified ultrasonic feature vector can be obtained by the following steps:

[0034] S210, clustering the historical qualified ultrasonic feature vectors corresponding to several historical qualified parts that are the same as the part to be tested, to obtain a cluster list B = (B1, B2, ..., B i ,…,B n );B i is the i-th cluster obtained by clustering; the position of any historical qualified component relative to the phased array ultrasonic probe during detection is the same as or different from the position of the component to be detected relative to the phased array ultrasonic probe during detection.

[0035] In this embodiment, during the historical time period, when the adhesive backing area of each historical component is inspected, the historical qualified ultrasonic feature vector corresponding to the historical component that has passed the inspection can be obtained; it can be understood that when the historical qualified component arrives at the preset position, its posture is not necessarily the same, and there will be deviations in posture. When there are enough historical components, the posture of the component to be inspected when it arrives at the preset position can be covered; the historical qualified ultrasonic feature vector contained in each cluster can be understood as the ultrasonic feature vector corresponding to the historical qualified component with similar posture.

[0036] S220, B i The central vector of is determined as the i-th qualified ultrasonic feature vector.

[0037] In this embodiment, B i The center vector is B i The average vector of all historical qualified ultrasonic feature vectors in the cluster; by clustering and accommodating normal fluctuations of qualified products (such as detection differences under different postures), the rigidity of a single fixed standard is avoided and the detection robustness is improved.

[0038] S300, traverse λ, if λ i >ρ i, it is determined that the overall adhesive area of the component to be inspected meets the preset qualification conditions; otherwise, the overall adhesive image TE corresponding to the overall adhesive area to be inspected is generated according to the ultrasonic signal collected by the phased array ultrasonic probe; ρ i is the similarity threshold corresponding to the i-th qualified ultrasonic feature vector.

[0039] In this embodiment, if any λ i >ρ i , indicating QA and B i If the historical qualified ultrasonic feature vectors in are relatively similar, the overall adhesive area of the component to be inspected is determined to be qualified; otherwise, the overall adhesive image TE to be inspected is generated according to the ultrasonic signal; it should be noted that the similarity between the historical qualified ultrasonic feature vectors in each cluster is different, so each cluster corresponds to a similarity threshold to improve the accuracy of the judgment.

[0040] First, qualified parts are quickly filtered through feature vectors to reduce the amount of subsequent image processing calculations; for suspected unqualified parts, further detailed analysis is performed through images.

[0041] Furthermore, ρ i It can be obtained by following the steps below:

[0042] S310, obtain B i Each qualified ultrasonic feature vector in B i The similarity between the center vectors of i The corresponding similarity list η i =(η i，1 , η i，2, ,…,η i，j ,…,η i，f(i) ), j = 1, 2, ..., f(i); where η i，j For B i The jth historical qualified ultrasonic feature vector in B i The similarity between the center vectors of B, f(i) is i The number of historical qualified ultrasonic feature vectors in .

[0043] S320, η i The minimum similarity in is determined as ρ i .

[0044] In this embodiment, the similarity threshold represents the minimum similarity of qualified features within the cluster, ensuring that the detection standard covers all qualified cases and reducing false positives (i.e., "qualified features must at least reach the similarity of the worst qualified sample in the cluster").

[0045] S400, extracting the adhesive area in TE to obtain the sub-adhesive image list to be detected corresponding to TE = (A1, A2, ..., A j ,…,A m ), j = 1, 2, ..., m; where A j is the jth sub-adhesive image to be detected, and m is the number of sub-adhesive images to be detected.

[0046] In this embodiment, it can be understood that TE is an image of the overall adhesive area of the component to be inspected, and there will be areas without adhesive and areas with adhesive in the overall adhesive area; the preset image segmentation algorithm (such as threshold segmentation, edge detection) can be used to locate the adhesive area to obtain the sub-adhesive image list A to be inspected corresponding to TE, which is convenient for subsequent local defect detection.

[0047] S500, A j The corresponding sub-ultrasound feature vector is concatenated with the sub-image feature vector to obtain A j The corresponding mixed eigenvector H j ; A j The corresponding sub-ultrasonic eigenvector is obtained by A j The ultrasonic signal collected by the corresponding ultrasonic probe is obtained.

[0048] In this embodiment, for each sub-image, the corresponding sub-ultrasonic feature vector (from the adjacent probe) and sub-image feature vector are obtained and spliced into a mixed feature vector; multimodal data (ultrasonic physical features + image visual features) are fused to improve the comprehensiveness of defect recognition.

[0049] Furthermore, step S500 may include the following steps:

[0050] S510, obtain j The center point is horizontally distant from the ultrasonic signal PA collected by the nearest ultrasonic probe TU.

[0051] In this embodiment, the phased array ultrasonic probe is in the form of a matrix and is set above the preset position. The coverage area of each ultrasonic probe is limited. j The ultrasonic signal collected by the nearest ultrasonic probe TU can cover A j The corresponding area is large, so the probe TU with the closest horizontal distance to the center point of the sub-image is selected, whose signal is more accurate (the closer the distance, the less noise), and the ultrasonic signal PA collected by TU is obtained as the extraction A j The corresponding sub-ultrasonic eigenvector basis.

[0052] S520, extract features from PA to obtain A jThe corresponding sub-ultrasonic feature vector PB=(PB1,PB2,…,PB r ,…,PB s ), r=1, 2,…, s; where PB r A j The rth element in the corresponding sub-ultrasonic feature vector, s is A j The number of elements in the corresponding sub-ultrasound feature vector.

[0053] In this embodiment, the features of the ultrasonic signal corresponding to each ultrasonic probe have been extracted in S100 and can be directly used in S520.

[0054] S530, get A j The corresponding sub-image feature vector PC = (PC1, PC2, ..., PC x ,…,PC y ), x=1,2,…,y;where PC x A j The xth element in the corresponding sub-image feature vector, y is A j The number of elements in the corresponding sub-image feature vector.

[0055] In this embodiment, those skilled in the art can use existing image feature extraction methods according to actual needs to obtain A j The corresponding sub-image feature vector PC is not described here in detail.

[0056] S540, PB and PC are spliced to obtain H j =(PB, PC).

[0057] In this embodiment, ultrasonic characteristics (reflecting internal structure) and image characteristics (intuitively displaying defect locations) are combined to avoid the limitations of a single modality (e.g., pure ultrasonic characteristics may miss detecting surface unevenness, and pure image characteristics may miss detecting internal bubbles).

[0058] Furthermore, TU may be far away from A j The horizontal distance of the center point is far, and the corresponding ultrasonic signal may not be able to fully cover A j If the corresponding adhesive area is still spliced with the image features with equal weights, it may lead to an error in the final judgment result. Based on this, after step S530, the method further includes the following steps:

[0059] S550, get TU and A j The horizontal distance L from the center point TU and A j The maximum distance L between the boundary and the center point max .

[0060] S551, according to L TU and L max , determine the weight ω1 corresponding to the sub-ultrasound feature vector and the weight ω2 corresponding to the sub-image feature vector; where ω1=1 / 2-(L TU / L max )×ψ; ψ is the preset coefficient, ψ<0.5; ω1+ω2=1.

[0061] S552, according to ω1 and ω2, perform weighted concatenation on PB and PC to obtain H j =(ω1×PB,ω2×PC).

[0062] In this embodiment, the closer the distance (L TU The smaller the distance, the higher the ultrasonic feature weight; conversely, the higher the image feature weight. Dynamically adjusting weights based on the detection position prevents interference from long-distance probe signal distortion and achieves adaptive optimization of feature fusion. The weighted stitching strategy dynamically adjusts weights based on physical distance to optimize feature fusion, making it particularly suitable for complex detection scenarios (such as curved surface adhesives and multi-probe arrays).

[0063] S600, according to H j , determine whether the quality of the overall adhesive area of the component to be inspected meets the preset qualification conditions.

[0064] Analyze H by presetting the model j , judge whether there are defects in each sub-area, and then determine the overall adhesive quality.

[0065] Furthermore, step S600 may include the following steps:

[0066] S610, inputting QA into a preset initial quality defect classification model to obtain a confidence list μ = (μ1, μ2, ..., μ a ,…,μ b ), a=1, 2,...,b; among them, μ a is the confidence level that the quality defect of the entire adhesive backing area is the ath type of quality defect, b is the number of preset quality defect types; μ c >μ c+1 , c=1,2,…,b-1.

[0067] In this embodiment, an initial quality defect classification model is first used to output a confidence ranking of the overall quality defect types; the initial quality defect classification model can adopt a machine learning or deep learning model (such as CNN, random forest), and the training data is historical defect samples.

[0068] Furthermore, quality defects include: bubble defects, breakage defects and unevenness defects.

[0069] S620, obtaining each preset single quality defect detection model to obtain a single quality defect detection model list E=(E1, E2, ..., E a ,…,E b ); where E a It is a preset single quality defect detection model for detecting the ath type of quality defect.

[0070] In this embodiment, each preset single quality defect detection model is for one type of quality defect, and the arrangement order of the single quality defect detection models in E is the same as the arrangement order of the confidence levels in μ.

[0071] S630, obtaining a preset value N=1.

[0072] S640, H j Input to E N , and get H j The confidence level V that the corresponding adhesive area has the Nth type of quality defect j，N .

[0073] S650, if V j，N >ε, it is determined that the quality of the overall adhesive area of the component to be tested does not meet the preset qualification conditions; otherwise, it is determined that H j If the corresponding adhesive backing area does not have the Nth type of quality defect, the process proceeds to S660 ; ε is a preset confidence threshold.

[0074] In this embodiment, when H j When testing, first j Input the single quality defect detection model corresponding to the most likely quality defect. If the confidence of the corresponding type of quality defect output by the single quality defect detection model is greater than the preset confidence threshold, it can be directly determined that there is a quality problem in the entire adhesive backing area without the need for subsequent testing, thereby improving detection efficiency.

[0075] S660: If N < b, obtain N = N + 1 and proceed to S640; otherwise, exit the current process.

[0076] S670, if H j If the corresponding adhesive backing area does not have any type of quality defects, it is determined that the quality of the entire adhesive backing area of the component to be inspected meets the preset qualification conditions.

[0077] In this embodiment, layered detection improves efficiency (first rough screening of the whole, then fine inspection of the local area); special models are designed for different types of quality defects to improve detection accuracy (for example, bubbles appear as low reflection in ultrasonic images, and fractures appear as signal interruption, and the model can be trained in a targeted manner).

[0078] The method of this embodiment has at least the following beneficial effects:

[0079] 1. Multimodal fusion improves detection accuracy

[0080] Combining ultrasonic features (reflecting internal structure) with image features (intuitively displaying defect locations) avoids the limitations of a single modality (e.g., pure ultrasonic features may miss surface unevenness, and pure image features may miss internal bubbles). The weighted stitching strategy dynamically adjusts weights based on physical distance to optimize feature fusion effects, making it particularly suitable for complex inspection scenarios (e.g., curved surface adhesives, multi-probe arrays).

[0081] 2. Data-driven adaptive standards

[0082] By clustering historical qualified data, a dynamic benchmark (qualified feature vector and threshold) is generated to adapt to normal fluctuations in the production process (such as slight equipment offset and material batch differences) and reduce the misjudgment rate; the threshold is based on the minimum similarity within the cluster to ensure that "the qualified standard is not lower than the worst qualified sample in history" and improve the rationality of the detection standard.

[0083] 3. Layered detection improves efficiency

[0084] First, qualified parts (S200-S300) are quickly screened through the overall feature vector, and only suspected unqualified parts are subjected to complex image segmentation and sub-region detection, which greatly reduces the amount of calculation and is suitable for high-speed production lines. The staged defect classification (initial model coarse classification + single defect model fine inspection) avoids the inefficiency of uniformly processing all defect types and improves detection speed.

[0085] 4. Targeted defect detection capabilities

[0086] Clearly cover typical defects such as bubbles, fractures, and unevenness. Use the penetrating power of phased array ultrasound to detect internal defects (such as bubbles) and image analysis to detect surface defects (such as unevenness), achieving full coverage of all defect types. The probe distance weighting strategy ensures the detection accuracy of local defects (such as edge fractures) to avoid missed detections due to signal attenuation.

[0087] 5. Automation and industrial adaptability

[0088] Non-contact detection, preset position triggering, and automatic model judgment are fully compatible with automated production lines and support real-time online detection. Its reliance on historical data (clustering and model training) enables self-optimization capabilities, continuously improving detection performance as production data accumulates.

[0089] Furthermore, although the steps of the method of the present disclosure are described in a particular order in the accompanying drawings, this does not require or imply that the steps must be performed in this particular order, or that all steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps.

[0090] An embodiment of the present invention further provides a device for detecting adhesive quality, the device comprising:

[0091] The ultrasonic feature vector acquisition module is used to obtain the ultrasonic feature vector QA corresponding to the entire adhesive area of the component to be inspected in response to the component to be inspected reaching a preset position; wherein QA is obtained through a preset phased array ultrasonic probe.

[0092] The similarity determination module is used to obtain the similarity between QA and each preset qualified ultrasonic feature vector to obtain a similarity list λ = (λ1, λ2, ..., λ i ,…,λ n ), i=1, 2,...,n; where, λ i is the similarity between QA and the i-th qualified ultrasonic feature vector, and n is the number of qualified ultrasonic feature vectors.

[0093] The qualified judgment module is used to traverse λ. If λ i >ρ i , it is determined that the overall adhesive area of the component to be inspected meets the preset qualification conditions; otherwise, the overall adhesive image TE corresponding to the overall adhesive area to be inspected is generated according to the ultrasonic signal collected by the phased array ultrasonic probe; ρ i is the similarity threshold corresponding to the i-th qualified ultrasonic feature vector.

[0094] The sub-to-be-detected adhesive image extraction module is used to extract the adhesive area in TE to obtain the sub-to-be-detected adhesive image list A corresponding to TE = (A1, A2, ..., A j ,…,A m ), j = 1, 2, ..., m; where A j is the jth sub-adhesive image to be detected, and m is the number of sub-adhesive images to be detected.

[0095] Feature vector concatenation module, used to concatenate A j The corresponding sub-ultrasound feature vector is concatenated with the sub-image feature vector to obtain A j The corresponding mixed eigenvector H j ; A j The corresponding sub-ultrasonic eigenvector is obtained by A jThe ultrasonic signal collected by the corresponding ultrasonic probe is obtained.

[0096] Quality judgment module, used to judge the quality of j , determine whether the quality of the overall adhesive area of the component to be inspected meets the preset qualification conditions.

[0097] An embodiment of the present invention further provides a back glue quality detection system, comprising the back glue quality detection device as described in the above embodiment.

[0098] An embodiment of the present invention further provides an electronic device including a processor and the aforementioned non-transitory computer-readable storage medium.

[0099] The electronic device is merely an example and should not limit the functions and scope of use of the embodiments of the present application.

[0100] The electronic device is implemented as a general-purpose computing device. Components of the electronic device may include, but are not limited to, the aforementioned at least one processor, the aforementioned at least one memory, and a bus connecting different system components (including the memory and the processor).

[0101] The memory stores program codes, which can be executed by the processor, so that the processor performs the steps of various embodiments described in this specification.

[0102] The memory may include readable media in the form of volatile memory, such as random access memory (RAM) and / or cache memory, and may further include read-only memory (ROM).

[0103] The memory may also include a program / utility having a set (at least one) of program modules, such program modules including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0104] The bus may represent one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor, or a local bus using any of a variety of bus architectures.

[0105] The electronic device may also communicate with one or more external devices (e.g., keyboards, pointing devices, Bluetooth devices, etc.), one or more devices that enable a user to interact with the electronic device, and / or any device that enables the electronic device to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface. Furthermore, the electronic device may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via a network adapter. The network adapter communicates with other modules of the electronic device via a bus. It should be understood that, although not shown in the figures, other hardware and / or software modules may be used in conjunction with the electronic device, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0106] Through the description of the above embodiments, it is easy for those skilled in the art to understand that the example embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution according to the embodiments of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the embodiments of the present disclosure.

[0107] An embodiment of the present invention further provides a computer program product comprising program code. When the program product is run on an electronic device, the program code is used to enable the electronic device to execute the steps of the method according to various exemplary embodiments of the present invention described above in this specification.

[0108] Although some specific embodiments of the present invention have been described in detail by way of examples, it should be understood by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the present invention. It should also be understood by those skilled in the art that various modifications may be made to the embodiments without departing from the scope and spirit of the present invention.

Claims

1. A method for detecting adhesive quality, characterized in that: The method comprises the following steps: S100, in response to the component to be inspected reaching a preset position, obtaining an ultrasonic feature vector QA corresponding to the entire adhesive area of the component to be inspected; wherein QA is obtained by a preset phased array ultrasonic probe; S200, obtaining the similarity between QA and each preset qualified ultrasonic feature vector to obtain a similarity list λ = (λ1, λ2, ..., λ i ,…,λ n ), i=1, 2,...,n; where, λ i is the similarity between QA and the i-th qualified ultrasonic feature vector, and n is the number of qualified ultrasonic feature vectors; S300, traverse λ, if λ i >ρ i , it is determined that the overall adhesive area of the component to be inspected meets the preset qualification conditions; otherwise, the overall adhesive image TE corresponding to the overall adhesive area to be inspected is generated according to the ultrasonic signal collected by the phased array ultrasonic probe; ρ i is the similarity threshold corresponding to the i-th qualified ultrasonic feature vector; S400, extracting the adhesive area in TE to obtain the sub-adhesive image list to be detected corresponding to TE = (A1, A2, ..., A j ,…,A m ), j = 1, 2, ..., m; where A j is the jth sub-adhesive image to be detected, and m is the number of sub-adhesive images to be detected; S500, A j The corresponding sub-ultrasound feature vector is concatenated with the sub-image feature vector to obtain A j The corresponding mixed eigenvector H j ; A j The corresponding sub-ultrasonic eigenvector is obtained by A j The ultrasonic signal collected by the corresponding ultrasonic probe is obtained; S600, according to H j , determine whether the quality of the overall adhesive area of the component to be inspected meets the preset qualification conditions.

2. The adhesive quality detection method according to claim 1, characterized in that: The preset qualified ultrasonic feature vector is obtained by the following steps: S210, clustering the historical qualified ultrasonic feature vectors corresponding to several historical qualified parts that are the same as the part to be tested, to obtain a cluster list B = (B1, B2, ..., B i ,…,B n );B i is the i-th cluster obtained by clustering; the position of any historical qualified component relative to the phased array ultrasonic probe during testing is the same as or different from the position of the component to be tested relative to the phased array ultrasonic probe during testing; S220, B i The central vector of is determined as the i-th qualified ultrasonic feature vector.

3. The adhesive quality detection method according to claim 2, characterized in that: ρ i Obtained through the following steps: S310, obtain B i Each qualified ultrasonic feature vector in history is i The similarity between the center vectors of i The corresponding similarity list η i =(η i，1 , η i，2, ,…,η i，j ,…,η i，f(i) ), j = 1, 2, ..., f(i); where η i，j B i The jth historical qualified ultrasonic feature vector in B i The similarity between the center vectors of B, f(i) is i The number of historical qualified ultrasonic feature vectors in; S320, η i The minimum similarity in is determined as ρ i .

4. The adhesive quality detection method according to claim 1, characterized in that: Step S500 includes the following steps: S510, obtain j The center point is horizontally distant from the ultrasonic signal PA collected by the nearest ultrasonic probe TU; S520, extract features from PA to obtain A j The corresponding sub-ultrasonic feature vector PB=(PB1,PB2,…,PB r ,…,PB s ), r=1, 2,…, s; where PB r A j The rth element in the corresponding sub-ultrasonic feature vector, s is A j The number of elements in the corresponding sub-ultrasound feature vector; S530, get A j The corresponding sub-image feature vector PC = (PC1, PC2, ..., PC x ,…,PC y ), x=1,2,…,y;where PC x A j The xth element in the corresponding sub-image feature vector, y is A j The number of elements in the corresponding sub-image feature vector; S540, PB and PC are spliced to obtain H j =(PB, PC).

5. The adhesive quality detection method according to claim 4, characterized in that: After step S530, the method further includes the following steps: S550, get TU and A j The horizontal distance L from the center point TU and A j The maximum distance L between the boundary and the center point max ; S551, according to L TU and L max , determine the weight ω1 corresponding to the sub-ultrasound feature vector and the weight ω2 corresponding to the sub-image feature vector; where ω1=1 / 2-(L TU / L max )×ψ; ψ is the preset coefficient, ψ<0.5; ω1+ω2=1; S552, according to ω1 and ω2, perform weighted concatenation on PB and PC to obtain H j =(ω1×PB,ω2×PC).

6. The adhesive quality detection method according to claim 1, characterized in that: Step S600 includes the following steps: S610, inputting QA into a preset initial quality defect classification model to obtain a confidence list μ = (μ1, μ2, ..., μ a ,…,μ b ), a=1, 2,...,b; among them, μ a is the confidence level that the quality defect of the entire adhesive backing area is the ath type of quality defect, b is the number of preset quality defect types; μ c >μ c+1 , c=1,2,…,b-1; S620, obtaining each preset single quality defect detection model to obtain a single quality defect detection model list E=(E1, E2, ..., E a ,…,E b ); where E a A single quality defect detection model is preset to detect the ath type of quality defect; S630, obtaining a preset value N=1; S640, H j Input to E N , get H j The confidence level V that the corresponding adhesive area has the Nth type of quality defect j，N ; S650, if V j，N >ε, it is determined that the quality of the overall adhesive area of the component to be tested does not meet the preset qualification conditions; otherwise, it is determined that H j If the corresponding adhesive backing area does not have the Nth type of quality defect, the process proceeds to S660 ; ε is a preset confidence threshold; S660, if N < b, obtain N = N + 1 and enter S640; otherwise, jump out of the current process; S670, if H j If the corresponding adhesive backing area does not have any type of quality defects, it is determined that the quality of the entire adhesive backing area of the component to be inspected meets the preset qualification conditions.

7. The adhesive quality detection method according to claim 6, characterized in that: The quality defects include: bubble defects, breakage defects and unevenness defects.

8. A device for detecting adhesive quality according to any one of claims 1 to 7, characterized in that: The device comprises: An ultrasonic feature vector acquisition module is configured to acquire an ultrasonic feature vector QA corresponding to the entire adhesive area of the component to be inspected in response to the component to be inspected reaching a preset position; wherein QA is obtained by a preset phased array ultrasonic probe; The similarity determination module is used to obtain the similarity between QA and each preset qualified ultrasonic feature vector to obtain a similarity list λ = (λ1, λ2, ..., λ i ,…,λ n ), i=1, 2,...,n; where, λ i is the similarity between QA and the i-th qualified ultrasonic feature vector, and n is the number of qualified ultrasonic feature vectors; The qualified judgment module is used to traverse λ. If λ i >ρ i , it is determined that the overall adhesive area of the component to be inspected meets the preset qualification conditions; otherwise, the overall adhesive image TE corresponding to the overall adhesive area to be inspected is generated according to the ultrasonic signal collected by the phased array ultrasonic probe; ρ i is the similarity threshold corresponding to the i-th qualified ultrasonic feature vector; The sub-to-be-detected adhesive image extraction module is used to extract the adhesive area in TE to obtain the sub-to-be-detected adhesive image list A corresponding to TE = (A1, A2, ..., A j ,…,A m ), j = 1, 2, ..., m; where A j is the jth sub-adhesive image to be detected, and m is the number of sub-adhesive images to be detected; Feature vector concatenation module, used to concatenate A j The corresponding sub-ultrasound feature vector is concatenated with the sub-image feature vector to obtain A j The corresponding mixed eigenvector H j ; A j The corresponding sub-ultrasonic eigenvector is obtained by A j The ultrasonic signal collected by the corresponding ultrasonic probe is obtained; Quality judgment module, used to judge the quality of j , determine whether the quality of the overall adhesive area of the component to be inspected meets the preset qualification conditions.

9. A back glue quality detection system, characterized in that: It includes the adhesive quality detection device as described in claim 8.

Images