Detection method, storage medium and electronic equipment

Abstract

The invention relates to the technical field of electronics, and discloses a detection method, a storage medium and electronic equipment. According to the detection method disclosed by the invention, the area bonded by the back glue is scanned by using detection equipment such as computer tomography (CT) and the like to obtain a slice image. The method comprises the steps that firstly, the area of gum and the area of air in a gum bonding area are determined, finally, the ratio R of the area of the air to the area of the gum bonding area is calculated, and the activation state of the gum is determined based on the ratio R. In this way, the possibility that the electronic equipment is damaged due to the fact that the activation state of the gum is checked by disassembling the electronic equipment is avoided. And the accuracy of detecting the activation state of the gum can be improved through the ratio, so that the quality of the electronic equipment is improved.

Description

Technical Field

[0001] The present application relates to the field of electronic technology, and in particular to a detection method, a storage medium and an electronic device. Background Art

[0002] Adhesive bonding is a way to fix components in electronic products, in which the adhesive is activated after being squeezed, and after activation, the components can be firmly bonded. For example, the battery in a mobile phone can be fixed to the middle frame of the mobile phone by adhesive bonding. Due to insufficient activation (extrusion) space for the adhesive, poor flatness of the adhesive bonding surface with the electronic device, or poor surface activation energy (energy of surface molecules in a liquid) of the adhesive, the adhesive may not be properly activated, resulting in the components in the electronic device being loosely fixed by adhesive bonding, and the components may shake or even fall off. Therefore, the activation status of the adhesive in the electronic product needs to be checked before the electronic product leaves the factory.

[0003] At present, the activation status of the adhesive is checked by disassembling the electronic device and manually judging it. However, the electronic device may be damaged during the disassembly process, causing the electronic device to be unable to continue to work normally. In addition, manual judgment of the activation status of the adhesive depends on subjective judgment, without objective and quantitative judgment basis, which may lead to errors in judging the activation status of the adhesive, affecting the quality of the electronic device. Summary of the invention

[0004] In order to solve the above-mentioned problem that the electronic device may be damaged during the disassembly process and there is a lack of objective quantitative judgment of the activation state of the back glue, the present application provides a detection method, a storage medium and an electronic device.

[0005] In a first aspect, the present application provides a detection method, which is applied to an electronic device, to obtain a first slice image corresponding to a first bonding area between a first component and a second component of the electronic device, wherein the first slice image includes images corresponding to the first component, the second component and an adhesive, the first component and the second component are bonded by the adhesive, and the first slice image is obtained by scanning the first bonding area with a detection device; and the activation state of the adhesive is determined based on the first slice image.

[0006] It can be understood that the first component mentioned in the present application can be a middle frame of an electronic device, the second component can be a battery, or any component in the electronic device that needs to be bonded, and the adhesive can be a back glue. The first bonding area can be a back glue bonding area. The first slice image can be a slice image corresponding to the back glue bonding area. The detection device can be a detection device such as an electronic computer tomography.

[0007] By using the above method, the activation state of the adhesive backing is determined by using the slice image, thereby avoiding the possibility of damaging the electronic device by disassembling the electronic device to check the activation state of the adhesive backing.

[0008] In a possible implementation of the first aspect above, the detection device includes an X-ray detection device or a Y-ray detection device.

[0009] In a possible implementation of the first aspect above, determining the activation state of the adhesive based on the first slice image includes: determining a second slice image based on the first slice image, the second slice image being at least a partial image of the first slice image, and the second slice image including an image corresponding to the adhesive; determining a second bonding area based on the second slice image, the second bonding area being an area corresponding to the adhesive, and the second bonding area including an adhesive area and an air area; determining the area of ​​the air area in the second bonding area; and determining the activation state of the adhesive based on a ratio of the area of ​​the air area to the area of ​​the second bonding area.

[0010] It can be understood that the second slice image mentioned in the present application can be a partial image of the first slice image. The second bonding area can be the precise range of the adhesive bonding area.

[0011] Through the above method, the activation state of the adhesive is detected by calculating the ratio of the area of ​​air to the area of ​​the adhesive bonding area, which can improve the accuracy of detecting the activation state of the adhesive, thereby improving the quality of electronic equipment.

[0012] In a possible implementation of the first aspect above, the activation state of the adhesive includes unqualified, medium and excellent; the activation state of the adhesive is determined based on the ratio of the area of ​​the air region to the area of ​​the second bonding region; including: when the ratio of the area of ​​the air region to the area of ​​the second bonding region is less than a first threshold, the activation state of the adhesive is determined to be excellent; when the ratio of the area of ​​the air region to the area of ​​the second bonding region is greater than or equal to the first threshold and less than or equal to the second threshold, the activation state of the adhesive is determined to be medium; when the ratio of the area of ​​the air region to the area of ​​the second bonding region is greater than the second threshold, the activation state of the adhesive is determined to be unqualified.

[0013] It can be understood that the ratio of the area of ​​the air region to the area of ​​the second adhesive region is the ratio R mentioned in this application. The first threshold is the threshold for judging the activation state of the adhesive backing, for example, the size can be 5%; the second threshold is the threshold for judging the activation state of the adhesive backing, for example, the size can be 10%. When the ratio R is less than 5%, the activation state of the adhesive backing is judged to be good, that is, the state mentioned in this application is excellent; when 5% ≤ ratio R ≤ 10%, the activation state of the adhesive backing is judged to be medium, that is, the state mentioned in this application is medium; when the ratio R is greater than 10%, the activation state of the adhesive backing is judged to be poor, that is, the state mentioned in this application is unqualified.

[0014] In a possible implementation of the first aspect above, determining the second slice image based on the first slice image includes: improving the first image contrast of the first slice image to the second image contrast, and selecting an area in the first slice image with a grayscale value less than a first grayscale value threshold value of the second image contrast as the second slice image; determining the second bonding area based on the second slice image includes: acquiring grayscale value data corresponding to each pixel in the second slice image; acquiring a first Gaussian model based on the grayscale value data; determining a second grayscale value threshold range corresponding to the second bonding area based on a first parameter in the first Gaussian model, wherein the first parameter includes a mean and a standard deviation; and determining the second bonding area based on the second grayscale value threshold range.

[0015] It can be understood that since the imaging of the adhesive bonding area is darker and the grayscale value is small, after the contrast of the slice image is improved, the approximate range of the adhesive bonding area can be determined based on the grayscale value of the slice image and the area with small grayscale value in the slice image.

[0016] It can be understood that the first gray value threshold mentioned in this application is the threshold for distinguishing the adhesive bonding area and the battery area. The first Gaussian model is the selected optimal Gaussian mixture model, and the second gray value threshold range is the gray value range corresponding to the second adhesive area.

[0017] In a possible implementation of the first aspect above, obtaining a first Gaussian model based on gray value data includes: processing the gray value data based on an expected maximum algorithm to obtain a first parameter corresponding to a second Gaussian model corresponding to different sub-distributions; determining the Bayesian information criterion value corresponding to the first Gaussian model corresponding to different sub-distributions based on the first parameter corresponding to the second Gaussian model corresponding to different sub-distributions; and selecting a second Gaussian model whose Bayesian information criterion value is equal to a third threshold as the first Gaussian model.

[0018] It can be understood that the first Gaussian model mentioned in this application is a selected Gaussian mixture model, and the second Gaussian model is each Gaussian mixture model corresponding to different sub-distributions. The third threshold mentioned in this application is the data with the minimum value of the Bayesian Information Criterion.

[0019] In a possible implementation of the first aspect above, determining the area of ​​the air region in the second bonding area includes: determining training data based on the grayscale values ​​of the air region and the adhesive region in the second bonding area; standardizing the training data through a first Gaussian model to obtain standardized training data; training a first logistic regression model based on the standardized training data to obtain a second logistic regression model; determining a classification threshold of the air region and the adhesive region based on the second logistic regression model; determining the air region and the adhesive region based on the classification threshold; and determining the area of ​​the air region based on the air region in the second bonding area.

[0020] It can be understood that the classification threshold mentioned in the present application is the threshold for distinguishing the air area and the adhesive area. The first logistic regression model is a logistic regression model trained on standardized training data; the second logistic regression model is a selected logistic regression model.

[0021] In a possible implementation of the first aspect above, the first logistic regression model is trained based on the standardized training data to obtain the second logistic regression model, including: training each first logistic regression model based on the standardized training data to obtain each trained logistic regression model; validating each trained logistic regression model by a cross-validation method to obtain accuracy data corresponding to each trained model; determining a receiver operating characteristic curve based on the accuracy data corresponding to each logistic regression model; determining a data point at a first position in the receiver operating characteristic curve, and using a model trained with the accuracy data corresponding to the data point as the second logistic regression model.

[0022] It can be understood that the accuracy data mentioned in this application includes the true positive rate and the false positive rate, and the first position is the position closest to the upper left corner in the receiver operating characteristic curve.

[0023] In a second aspect, the present application provides a storage medium having instructions stored thereon, which, when executed on an electronic device, enables the electronic device to implement the above-mentioned first aspect and any possible detection method of the above-mentioned first aspect.

[0024] In a third aspect, the present application provides an electronic device, comprising: a memory for storing instructions executed by one or more processors in the electronic device; and a processor for executing the instructions stored in the memory to implement the above-mentioned first aspect and any possible detection method of the above-mentioned first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 According to some embodiments of the present application, a schematic diagram of a battery adhesive is shown;

[0026] Figure 2 According to some embodiments of the present application, a physical schematic diagram of a disassembled mobile phone 10 is shown;

[0027] Figure 3 According to some embodiments of the present application, a physical schematic diagram of adhesive in a battery compartment is shown;

[0028] Figure 4A According to some embodiments of the present application, a schematic flow chart of a detection method is shown;

[0029] Figure 4B According to some embodiments of the present application, a flow chart of a method for determining the activation state of a back glue based on a slice image is shown;

[0030] Figure 5 According to some embodiments of the present application, a schematic diagram of a 3D model obtained by CT scanning is shown;

[0031] Figure 6 According to some embodiments of the present application, a schematic diagram of a slice image is shown;

[0032] Figure 7 According to some embodiments of the present application, a schematic diagram of a local image in a slice image is shown;

[0033] Figure 8 According to some embodiments of the present application, a schematic diagram of a gray value distribution histogram is shown;

[0034] Fig. 9 According to some embodiments of the present application, a schematic diagram of a back adhesive bonding area is shown;

[0035] Fig.10 According to some embodiments of the present application, a schematic diagram of a ROC curve is shown;

[0036] Fig.11 According to some embodiments of the present application, a schematic diagram of an optimal logistic regression model curve is shown;

[0037] Fig.12 According to some embodiments of the present application, a schematic diagram of an activated state of a back glue is shown;

[0038] Fig.13 According to some embodiments of the present application, a schematic diagram of the hardware structure of a mobile phone 10 is shown. DETAILED DESCRIPTION

[0039] The illustrative embodiments of the present application include, but are not limited to, a detection method, a storage medium, and an electronic device.

[0040] The following is an introduction to the terms involved in this application:

[0041] (1) Gaussian Mixture Model

[0042] The Gaussian mixture model is a model that decomposes things into several Gaussian probability density functions (normal distribution curves).

[0043] (2) Logistic regression model

[0044] The logistic regression model is a generalized linear regression analysis model that estimates the probability of an event based on a given set of independent variable data.

[0045] (3) Expectation maximization

[0046] The maximum expectation algorithm is an optimization algorithm that performs maximum likelihood estimation through iteration. It is used to estimate parameters of probability models containing hidden variables or missing data.

[0047] (4) Bayesian Information Criterion (BIC)

[0048] BIC uses subjective probability to estimate some unknown states under incomplete intelligence, then uses the Bayesian formula to correct the probability of occurrence, and finally uses the expected value and corrected probability to make the optimal decision.

[0049] (5) Cross-validation

[0050] Cross-validation is to take out most of the samples from a given modeling sample to build the model, keep a small part of the samples to use the newly established model for prediction, and calculate the prediction error of this small part of the samples, and record the sum of their squares.

[0051] (6) Receiver operating characteristic (ROC)

[0052] The ROC curve refers to a coordinate graph with the false positive rate as the horizontal axis and the true positive rate as the vertical axis. It is a curve drawn under specific stimulus conditions due to different results obtained by using different judgment criteria.

[0053] The components in the electronic device mentioned in this application are any components in the electronic device bonded by adhesive, including but not limited to batteries, cameras, earpieces, and microphones. For the sake of ease of description, the following introduction is made using the battery as an example.

[0054] The electronic devices mentioned in this application include but are not limited to mobile phones, tablet computers, laptop computers, PDAs, mobile internet devices (MID), wearable devices, virtual reality (VR) devices, and augmented reality (AR) devices. For ease of description, the following takes the mobile phone 10 as an example for the electronic device.

[0055] The technical solution of the present application is introduced below in conjunction with the accompanying drawings.

[0056] As mentioned above, the battery in the mobile phone 10 can be fixed by adhesive bonding. Figure 1 A schematic diagram of a battery adhesive is shown, such as Figure 1As shown, the mobile phone 10 includes a middle frame 101 (the first component mentioned in this application) and a battery 102 (the second component mentioned in this application), and the battery 102 is fixed in the middle frame 101 by a back glue 103 (the adhesive mentioned in this application). It can be understood that the middle frame 101 is a fixing plate for fixing parts located between the front panel and the back shell (not marked in the figure) of the mobile phone 10. The back glue 103 is activated after being squeezed, and after activation, the battery 102 can be firmly bonded to the middle frame 101 in the mobile phone 10. Due to insufficient activation space of the back glue 103, poor flatness of the bonding surface between the back glue 103 and the middle frame 101, or poor surface activation energy of the back glue 103, the back glue 103 may be poorly activated, resulting in the battery 102 bonded by the back glue in the mobile phone 10 not being firmly fixed, and the battery 102 shaking or even falling. Therefore, it is necessary to check the activation state of the back glue 103 in the mobile phone 10.

[0057] Currently, the activation state of the adhesive 103 on the surface of the battery 102 is checked by disassembling the mobile phone 10 and manually judging. For example, Figure 2 FIG. 1 shows a schematic diagram of a disassembled mobile phone 10. Figure 2 As shown, the middle frame 101 has a battery compartment 201 for accommodating a battery 102, and the battery 102 is bonded to the battery compartment 201 in the middle frame 101 by means of adhesive 103. By disassembling the battery compartment 201 in the mobile phone 10, that is, removing the battery 102 from the battery compartment 201, the activation state of the adhesive 103 on the surface of the battery 102 or the surface of the battery compartment 201 can be checked manually based on experience. For example, Figure 3 A physical schematic diagram of the adhesive in the battery compartment is shown. It can be understood that in the battery compartment 201 of the mobile phone 10, the adhesive style of the adhesive 103 can be set as needed, and there is no restriction here. However, in the process of disassembling the mobile phone 10, the mobile phone 10 may be damaged, resulting in the problem that the mobile phone 10 cannot be used any more. In addition, judging the activation state of the adhesive 103 depends on manual experience, and there is no objective and quantitative basis for judgment, which may lead to errors in judging the activation state of the adhesive 103, affecting the quality of the mobile phone 10.

[0058] In order to solve the above problems, the present application proposes a detection method. Specifically, a detection device such as electronic computer tomography (CT) is used to scan the bonding area (such as the adhesive bonding area of ​​the battery and the middle frame), and a slice image corresponding to the bonding area is obtained, and the activation state of the adhesive is determined based on the slice image. Among them, the method of determining the activation state of the adhesive based on the slice image can be: first determine the bonding area based on the slice image, then determine the area of ​​the adhesive and the area of ​​the air in the bonding area, and finally, by calculating the ratio R of the area of ​​the air to the area of ​​the bonding area, the activation state of the adhesive is determined based on the ratio R. It can be understood that since the bonding area is composed of air and adhesive, the activation state of the adhesive can be effectively reflected by calculating the ratio of the area of ​​the air to the area of ​​the bonding area. Among them, when the area of ​​the air is smaller, it means that the effect of activation after the adhesive is squeezed is better. For example, when the ratio R of the area of ​​air to the area of ​​the bonding area is less than 5%, the activation state of the adhesive is judged to be good, that is, the activation state mentioned in this application is excellent; when 5% ≤ ratio R ≤ 10%, the activation state of the adhesive is judged to be medium; when the ratio R> 10%, the activation state of the adhesive is judged to be poor, that is, the activation state mentioned in this application is unqualified. Among them, 5%, 10%, and 15% are empirical values. In some embodiments, the ratio R in detecting the activation state of the adhesive can also have different ranges, which are not limited here.

[0059] It can be understood that the first threshold mentioned in the present application is a threshold for determining the activation state of the adhesive backing, for example, the value may be 5%; the second threshold is a threshold for determining the activation state of the adhesive backing, for example, the value may be 10%.

[0060] It is understood that the adhesive mentioned in the present application may be adhesive backing, or any adhesive substance used for bonding components in electronic devices. The bonding area mentioned in the present application may be an adhesive backing bonding area.

[0061] In this way, determining the activation state of the adhesive by slice image avoids the possibility of damaging the electronic device by disassembling the electronic device to check the activation state of the adhesive. And by calculating the ratio of the area of ​​air to the area of ​​the adhesive bonding area, the activation state of the adhesive can be detected, which can improve the accuracy of detecting the activation state of the adhesive, thereby improving the quality of the electronic device.

[0062] Combine the following Figure 4A , introduce the detection method provided by this application. Figure 4A A schematic diagram of a detection method is shown, such as Figure 4A As shown, the method comprises the following steps:

[0063] S101: Acquire a slice image corresponding to the adhesive bonding area.

[0064] In the present application, the adhesive bonding area of ​​the back glue can be scanned using a CT detection device to obtain a corresponding slice image (i.e., the first slice image mentioned in the present application). Specifically, the adhesive bonding area of ​​the back glue in the battery compartment (the first bonding area mentioned in the present application) is scanned using a CT detection device to obtain a three-dimensional (3D) model. For example, the CT detection device can be an X-ray CT device, a Y-ray CT device, etc.

[0065] Figure 5 A schematic diagram of a 3D model obtained by CT scanning is shown, Figure 5 As shown, the CT detection device displays the model of the adhesive bonding area from the overall view, the x-axis, i.e., the left and right space view, the y-axis, i.e., the front and back space view, and the z-axis, i.e., the upper and lower space view. Among them, the overall view displays the 3D model of the adhesive bonding area as a whole, and the length of the adhesive bonding area is 18,728.00μm, the width is 18,347.45μm, and the height is 17,922.77μm. The X-axis view displays the 3D model of the adhesive bonding area from the left and right space, with a scale of 1cm. The Y-axis view displays the 3D model of the adhesive bonding area from the front and back space, with a scale of 1cm. The Z-axis view displays the 3D model of the adhesive bonding area from the upper and lower space, with a scale of 1mm. In order to make the acquired 3D model clear, the resolution of the CT detection device should be less than 1 / 5 of the thickness of the adhesive in the battery, where 1 / 5 is empirical data. In some embodiments, the resolution of the CT detection device may also have other standards, which are not limited here.

[0066] Furthermore, the acquired 3D model is sliced ​​and analyzed to obtain a slice image (the first slice image mentioned in this application), and the original grayscale value information of the slice image is retained. For example, Figure 6 A schematic diagram of a slice image is shown, such as Figure 6 As shown, the slice image retains the original grayscale value information, the scale is 1 mm, and, since the adhesive bonding area is composed of air and adhesive, compared with the middle frame and the battery, the rays in the CT detection equipment have a higher transmittance to air and adhesive, and the imaging of the adhesive bonding area is darker.

[0067] S102: Determine the activation state of the adhesive based on the slice image.

[0068] Combine the following Figure 4B Introduce the contents of S102, for example, Figure 4B A schematic diagram of a method for determining the activation state of adhesive based on a slice image is shown. Figure 4B As shown, the method comprises the following steps:

[0069] S201: Determine the adhesive bonding area based on the slice image.

[0070] In the present application, the method for determining the adhesive bonding area based on the slice image can be to first obtain the grayscale values ​​of the global pixels of the slice image, extract the global maximum grayscale value and the global minimum grayscale value from the grayscale values ​​of the global pixels, and perform standardization on the slice image, that is, to improve the contrast of the slice image and determine the approximate range of the adhesive bonding area. Then, the local image in the slice image and the Gaussian mixture model algorithm are combined to determine the precise range of the adhesive bonding area.

[0071] For example, based on Figure 6 The slice image shown is used to determine the approximate range of the adhesive bonding area by improving the contrast of the slice image, for example, by improving the contrast of the first image to the second image. Since the imaging of the adhesive bonding area is darker and the grayscale value is small, after the contrast of the slice image is improved, the approximate range of the adhesive bonding area can be determined based on the grayscale value of the slice image and the area with a small grayscale value in the slice image. For example, the slice image can be standardized according to the following formula (1):

[0072] Standardization processing formula: (grayscale value of a certain point - minimum grayscale value of the entire domain) / (maximum grayscale value of the entire domain - minimum grayscale value of the entire domain) (1)

[0073] Among them, (grayscale value of a certain point - minimum grayscale value of the whole domain) represents the difference between the grayscale value of a certain point and the minimum grayscale value of the whole domain, and (maximum grayscale value of the whole domain - minimum grayscale value of the whole domain) represents the difference between the maximum grayscale value of the whole domain and the minimum grayscale value of the whole domain. The slice image is standardized according to the ratio of the difference between the grayscale value of a certain point and the minimum grayscale value of the whole domain divided by the difference between the maximum grayscale value of the whole domain and the minimum grayscale value of the whole domain. That is, according to formula (1), the grayscale value of each pixel in the image is standardized to obtain the slice image after the standardized processing, that is, to improve the contrast of the slice image.

[0074] After the contrast of the slice image is improved, the approximate range of the adhesive bonding area is determined based on the gray value of the slice image and the area with a small gray value in the slice image, for example, the gray value is less than the first gray value threshold, for example, the first gray value threshold can be 9000. Figure 6 In the slice image shown in FIG. 1 , an area whose gray value is less than the first gray value threshold is selected to be intercepted as shown in FIG. Figure 7 The local image shown (the second slice image mentioned in this application). Figure 7 A schematic diagram of a local image in a slice image is shown, such as Figure 7As shown, the horizontal axis represents X pixels, the vertical axis represents Y pixels, and the grayscale value diagram is shown on the right side of the image. The lighter the color, the larger the grayscale value, and the darker the color, the smaller the grayscale value. In the local image of the slice image, the area with a large grayscale value is the area of ​​the battery and the middle frame (i.e., the light-colored area), and the area with a small grayscale value (i.e., the dark-colored area) is the area where the back glue is bonded, that is, the approximate range of the back glue bonding area.

[0075] Then, based on the local image, the precise range of the adhesive bonding area (i.e., the second adhesive area mentioned in the present application) can be determined by obtaining the gray value data corresponding to each pixel point in the local image. For example, the gray value data corresponding to each pixel point in the local image can be generated as follows: Figure 8 The gray value distribution histogram shown; based on the gray value data, the optimal Gaussian mixture model (ie, the first Gaussian model mentioned in this application) is obtained; based on the mean μ in the optimal Gaussian mixture model m , standard deviation σ m Determine the gray value range corresponding to the precise range of the adhesive bonding area (ie, the second gray value threshold range mentioned in this application), for example, Figure 8 As shown, the grayscale value range can be [μ min -3σ min , μ min +3σ min ]; Determine the exact extent of the adhesive bonding area based on the grayscale value range.

[0076] The optimal Gaussian model based on the gray value data can be obtained by: processing the gray value data based on the expected maximum algorithm to obtain the mean μ corresponding to each Gaussian mixture model (i.e., the second Gaussian model mentioned in this application) corresponding to different sub-distributions m , standard deviation σ m ; Based on the mean μ corresponding to each Gaussian mixture model m , standard deviation σ m Determine the BIC value corresponding to each Gaussian mixture model; wherein the BIC value can be obtained according to formula (3); select the Gaussian mixture model with the smallest BIC value (that is, the second Gaussian model with a BIC value equal to the third threshold mentioned in this application) as the optimal Gaussian mixture model.

[0077] The following is a detailed description of the method for determining the precise range of the adhesive bonding area. Figure 7 The local image in the slice image shown in the figure extracts the grayscale values ​​of the global pixels in the local image and generates a grayscale value distribution histogram. Figure 8 A schematic diagram of a gray value distribution histogram is shown, such as Figure 8As shown in the figure, the horizontal axis represents the gray value, the vertical axis represents the probability density, and the vertical stripes represent the gray value distribution. The gray value distribution histogram describes the number of each gray value in the image, reflecting the frequency of each gray value in the image. When the gray value ranges from 7000 to 10000, the probability density gradually increases, and the probability density gradually increases from 0 to 0.15×10 -3 ; When the gray value is from 10000 to 10500, the probability density increases sharply, and when the gray value is 10500, the probability density increases to a maximum value of 0.85×10; when the gray value is from 10500 to 12500, the probability density decreases sharply, and when the gray value is 12500, the probability density is as small as 0. The gray value data of the global pixel points in the local image are fitted with a Gaussian mixture model, and the Gaussian mixture model curve is obtained, such as Figure 8 As shown in the figure, the curve that fits the gray value distribution is the Gaussian mixture model curve. The probability density function of the Gaussian mixture distribution refers to the following formula (2):

[0078]

[0079] Where m is the sub-distribution in the Gaussian mixture distribution, M is the number of sub-distributions m in the Gaussian mixture distribution, x is the gray value of the pixel, and c is the gray value of the pixel. m is the proportion of the Gaussian mixture distribution neutron distribution m, μ m is the mean of the Gaussian mixture distribution neutron distribution m, σ m is the standard deviation of the neutron distribution m of the Gaussian mixture distribution, and exp is the expected value.

[0080] In the Gaussian mixture model, by defining the number of different sub-distributions M, and combining the gray value data of the global pixel points and the expectation maximization algorithm, the mean μ of the Gaussian mixture model is calculated. m , standard deviation σ m Then, the BIC value of the Gaussian mixture model is calculated based on the parameter estimation results and the BIC formula (3).

[0081]

[0082] Where k is the number of model parameters, n is the number of samples, is the maximum likelihood value.

[0083] By comparing the BIC values ​​corresponding to the number M of different sub-distributions, the Gaussian mixture model algorithm shows that the smaller the BIC, the higher the degree of fit of the Gaussian mixture model and the fewer samples used. The number of sub-distributions M corresponding to the minimum BIC (that is, the Bayesian information criterion value mentioned in this application is equal to the third threshold, and the third threshold is the data with the minimum Bayesian information criterion value) is selected. The number of sub-distributions M is the optimal number of sub-distributions for the Gaussian mixture model (the first Gaussian model mentioned in this application), that is, the model with this number of sub-distributions is the optimal fitting model. In addition, the mean μ and standard deviation σ of all sub-distributions of the model are obtained in combination with formula (2).

[0084] Since the adhesive bonding area is composed of air and adhesive, the transmittance of the radiation in the CT detection equipment to air and adhesive is higher than that of the middle frame and battery, and the radiation receiving amount in the CT detection equipment is larger, resulting in darker imaging and smaller grayscale value in the adhesive bonding area. Therefore, among the means μ of all sub-distributions of the best fitting model, the minimum mean μ is selected. min The corresponding sub-distribution (the first Gaussian model mentioned in this application) is used to confirm the grayscale value range of the adhesive bonding area. Since, in the model curve of the Gaussian mixture, the probability that the grayscale value falls outside (μ-3σ, μ+3σ) is less than three thousandths, it is often believed in practice that the event that the grayscale value falls outside (μ-3σ, μ+3σ) will not occur, so the interval (μ-3σ, μ+3σ) can be regarded as the actual possible value interval of the grayscale value of the adhesive bonding area, which is called the "3σ" principle of normal distribution. For example, the grayscale value range of the adhesive bonding area is [μ min -3σ min , μ min +3σ min ](the second gray value threshold range mentioned in this application). Figure 8 As shown in the gray value distribution histogram, according to the minimum mean μ min , determine the gray value range of the adhesive bonding area. And according to the gray value range, Figure 7 The adhesive bonding area (the second bonding area mentioned in this application) is accurately extracted from the partial image shown. Fig. 9 A schematic diagram of a back adhesive bonding area is shown, such as Fig. 9 As shown, the area with darker image and smaller gray value is the area where the back glue is bonded, and the area with larger gray value is the area of ​​the middle frame and the battery.

[0085] S202: Determine the area of ​​the adhesive and the area of ​​air in the adhesive bonding area.

[0086] In the present application, the grayscale values ​​of the air area and the adhesive area (the adhesive area mentioned in the present application) can first be standardized using a Gaussian mixture model for training data, and then a logistic regression model can be trained on the standardized training data. The trained model can be verified by a cross-validation method, and the optimal logistic regression model in the verification result can be selected to determine the classification threshold, and the area of ​​the adhesive and the area of ​​the air in the adhesive bonding area can be determined based on the classification threshold.

[0087] Specifically, in Fig. 9 In the adhesive bonding area shown in the figure, the grayscale values ​​of the air area and the adhesive area are selected to establish the training data x, and the minimum mean μ generated by the Gaussian mixture model is fitted. min and the minimum standard deviation σ min , and combined with formula (4) to standardize the training data x to avoid the influence of gray value differences between different samples or different scans.

[0088]

[0089] Among them, μ min is the minimum mean, σ min is the minimum standard deviation, x is the training data, z scaled is the standardized training data.

[0090] Combined with formula (5) of the logistic regression model, and based on the standardized training data in formula (4), the logistic regression model is trained on the standardized training data (ie, the first logistic regression model mentioned in this application).

[0091]

[0092] Among them, z scaled is the standardized training data, w and b are the training parameters of the logistic regression model.

[0093] The training results of the logistic regression model are then verified by a cross-validation method. For example, the training results of the logistic regression model can be verified by a simple cross-validation, a leave-one-out cross-validation, or a leave-one-out cross-validation method, and the false positive rate (i.e., negative samples predicted by the model to be positive) and the true positive rate (i.e., positive samples predicted by the model to be negative) of each regression model (i.e., different combinations of training parameters w and b) are calculated based on the verification results to form an ROC curve.

[0094] For example, Fig.10 A schematic diagram of a ROC curve is shown, Fig.10As shown, the abscissa represents the false positive rate, the ordinate represents the true positive rate, and the area under the ROC curve (AUC) is 0.95. Among them, AUC is a simple evaluation of the performance of the logistic regression model. When AUC is above 0.9, the performance of the logistic model is excellent; when AUC is 0.7-0.9, the performance of the logistic model is good; when AUC is 0.5-0.7, the performance of the logistic model is poor. In this application, AUC is 0.95, indicating that the performance of the logistic regression model is excellent through parameter training. When the false positive rate is 0-0.05, the true positive rate rises linearly from 0, and when the false positive rate is 0.05, the true positive rate reaches 0.82 (i.e. point A) (accuracy data mentioned in this application); when the false positive rate is 0.05-1, the true positive rate slowly rises from 0.82 to 1. On the ROC curve, point A, which is closest to the upper left corner (i.e., the false positive rate is 0.05 and the true positive rate is 0.82), represents the maximum sum of sensitivity and specificity, that is, in the ROC curve, a higher true positive rate can be achieved with a lower false positive rate, that is, the regression model corresponding to point A (the first position mentioned in this application) has the highest prediction accuracy (the second logistic regression model mentioned in this application). Therefore, the optimal logistic regression model can be determined through point A.

[0095] Fig.11 A schematic diagram of an optimal logistic regression model curve is shown, such as Fig.11 As shown, the horizontal axis represents the standardized gray value, and the vertical axis represents the probability. When the standardized gray value is from -6 to -2, the probability rises slowly from 0 to 0.1; when the standardized gray value is from -2 to 0, the probability rises linearly from 0.1 to 0.9; when the standardized gray value is from 0 to 4, the probability rises slowly from 0.9 to 1. In the binary classification of logistic regression, when the probability is greater than 0.5, the probability value is considered to be 1, and when the probability is less than 0.5, the probability value is considered to be 0. The standardized gray value of 0.8502 corresponding to the probability value of 0.5 in the optimal logistic regression model is taken as the classification threshold (that is, the threshold for distinguishing the air area and the adhesive area in the adhesive bonding area). According to the regression model curve, it can be concluded that the classification threshold is u-0.8502σ. When the gray value is greater than the classification threshold, the part is adhesive, and the part with a gray value less than the classification threshold is air.

[0096] For example, to determine the area of ​​the air region within the precise range of the adhesive bonding area, the following can be done: determine the training data based on the grayscale values ​​of the air region and the adhesive region within the precise range of the adhesive bonding area; standardize the training data using an optimal Gaussian mixture model to obtain the standardized training data; train the logistic regression model based on the standardized training data (i.e., the first logistic regression model mentioned in the present application) to obtain the selected optimal logistic regression model (i.e., the second logistic regression model mentioned in the present application); determine the classification threshold of the air region and the adhesive region based on the selected optimal logistic regression model, wherein the classification threshold is the threshold for distinguishing the air region from the adhesive region; determine the air region and the adhesive region based on the classification threshold; and determine the area of ​​the air region based on the air region within the precise range of the adhesive bonding area.

[0097] The logistic regression model is trained based on the standardized training data to obtain the optimal logistic regression model. The method can include: training each logistic regression model based on the standardized training data to obtain each trained logistic regression model; validating each trained logistic regression model by a cross-validation method to obtain the true positive rate and false positive rate (i.e., the accuracy data mentioned in this application) corresponding to each trained model; determining the true positive rate and false positive rate corresponding to each logistic regression model based on the true positive rate and false positive rate. Figure 8 The ROC curve shown; determine the data point closest to the upper left corner (i.e., the first position mentioned in this application) in the ROC curve; and use the model trained with the true positive rate and false positive rate corresponding to the data point as the selected optimal logistic regression model.

[0098] S203: Calculate the ratio R of the area of ​​air to the area of ​​the adhesive bonding region, and determine the activation state of the adhesive based on the ratio R.

[0099] In this application, you can Fig. 9 In the adhesive bonding area shown, the activation state of the adhesive is determined by calculating the ratio R of the area of ​​air in the adhesive bonding area to the area of ​​the adhesive bonding area. For example, when the ratio R of the area of ​​air to the area of ​​the adhesive bonding area is less than 5%, the activation state of the adhesive is determined to be good; when 5%≤ratio R≤10%, the activation state of the adhesive is determined to be medium; when the ratio R>10%, the activation state of the adhesive is determined to be poor, wherein 5%, 10%, and 15% are empirical values. In some embodiments, the ratio R in detecting the activation state of the adhesive may also have different ranges, which are not limited here.

[0100] Fig.12 A schematic diagram showing a state of adhesive activation is shown, Fig.12 As shown in the figure, in area 1, by calculating the bonding area A and the air area A air, the ratio of the area of ​​air to the adhesive bonding area is 10%, indicating that the adhesive activation state at this location is poor. In area 2, the ratio of the area of ​​air to the adhesive bonding area is 3%, indicating that the adhesive activation at this location is good. Fig.12 As shown, the area of ​​the air region in region one is greater than that in region two. When the area of ​​the air is smaller, it means that the activation effect of the adhesive after being squeezed is better.

[0101] In this way, determining the activation state of the adhesive by slice image avoids the possibility of damaging the electronic device by disassembling the electronic device to check the activation state of the adhesive. And by calculating the ratio of the area of ​​air to the area of ​​the adhesive bonding area, the activation state of the adhesive can be detected, which can improve the accuracy of detecting the activation state of the adhesive, thereby improving the quality of the electronic device.

[0102] Let's take mobile phone 10 as an example and combine Fig.13 , introducing the hardware structure of the electronic device in this application. Fig.13 FIG. 1 shows a schematic diagram of the hardware structure of a mobile phone 10. Fig.13 As shown, the mobile phone 10 may include a processor 110, a memory 120, an interface module 130, a power module 140, a wireless communication module 150, a mobile communication module 160, an audio module 170, a sensor module 180, a button 190, a motor 191, and a display screen 192, etc.

[0103] The processor 110 may include one or more processing units, for example: the processor 110 may include an application processor (application processor, AP), a modem processor, a GPU, an image signal processor (image signal processor, ISP), a controller, a video codec, a digital signal processor (digital signal processor, DSP), a baseband processor, and / or a neural-network processing unit (neural-network processing unit, NPU), etc. Among them, different processing units can be independent devices or integrated in one or more processors. For example, in some embodiments of the present application, the processor 110 can execute the detection method mentioned in the present application.

[0104] The processor 110 may also be provided with a memory for storing instructions and data. In some embodiments, the memory in the processor 110 is a cache memory. The processor 110 may execute the detection method mentioned in the present application based on the instructions and data.

[0105] The memory 120 may be used to store computer executable program codes, which include instructions. The memory 120 may include a program storage area and a data storage area. The program storage area may store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc. The data storage area may store data created during the use of the mobile phone 10 (such as audio data, a phone book, etc.), etc.

[0106] The power module 140 is used to receive charging input from the charger and to power the processor 110, the memory 120, the display screen 192, and the wireless communication module 150. In some other embodiments, the power module 140 may also be disposed in the processor 110. In the present application, the power module 140 may include a battery bonded to the mobile phone 10 by adhesive.

[0107] It is to be understood that the structure illustrated in the embodiment of the present application does not constitute a specific limitation on the mobile phone 10. In other embodiments of the present application, the mobile phone 10 may include more or fewer components than shown in the figure, or combine some components, or separate some components, or arrange the components differently. The components shown in the figure may be implemented in hardware, software, or a combination of software and hardware.

[0108] In the accompanying drawings, some structural or method features may be shown in a specific arrangement and / or order. However, it should be understood that such a specific arrangement and / or order may not be required. Instead, in some embodiments, these features may be arranged in a manner and / or order different from that shown in the illustrative drawings. In addition, the inclusion of structural or method features in a particular figure does not mean that such features are required in all embodiments, and in some embodiments, these features may not be included or may be combined with other features.

[0109] It should be noted that the units / modules mentioned in the various device embodiments of the present application are all logical units / modules. Physically, a logical unit / module can be a physical unit / module, or a part of a physical unit / module, or can be implemented as a combination of multiple physical units / modules. The physical implementation method of these logical units / modules themselves is not the most important. The combination of functions implemented by these logical units / modules is the key to solving the technical problems proposed by the present application. In addition, in order to highlight the innovative part of the present application, the above-mentioned device embodiments of the present application do not introduce units / modules that are not closely related to solving the technical problems proposed by the present application, which does not mean that there are no other units / modules in the above-mentioned device embodiments.

[0110] It should be noted that in the examples and description of this patent, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "including one" do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0111] Although the present application has been illustrated and described with reference to certain preferred embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the scope of the present application.

Claims

1. A detection method, applied to electronic equipment, characterized in that: Acquire a first slice image corresponding to a first bonding area between a first component and a second component of the electronic device, wherein the first slice image includes images corresponding to the first component, the second component, and an adhesive, the first component and the second component are bonded by the adhesive, and the first slice image is acquired by scanning the first bonding area with a detection device; An activation state of the adhesive is determined based on the first slice image.

2. The detection method according to claim 1, characterized in that: The detection equipment includes an X-ray detection equipment or a Y-ray detection equipment.

3. The detection method according to claim 1, characterized in that: The step of determining the activation state of the adhesive based on the first slice image comprises: determining a second slice image based on the first slice image, wherein the second slice image is at least a partial image of the first slice image, and the second slice image includes an image corresponding to the adhesive; determining a second bonding area based on the second slice image, where the second bonding area is an area corresponding to the adhesive, and the second bonding area includes an adhesive area and an air area; determining the area of ​​the air region in the second bonding region; The activation state of the adhesive is determined based on the ratio of the area of ​​the air region to the area of ​​the second adhesive region.

4. The detection method according to claim 3, characterized in that: The activation status of the adhesive includes unqualified, medium and excellent; determining the activation state of the adhesive based on a ratio of the area of ​​the air-based region to the area of ​​the second adhesive region; include: When the ratio of the area of ​​the air region to the area of ​​the second bonding region is less than a first threshold, determining that the activation state of the adhesive is excellent; When the ratio of the area of ​​the air region to the area of ​​the second bonding region is greater than or equal to the first threshold and less than or equal to the second threshold, determining that the activation state of the adhesive is medium; When the ratio of the area of ​​the air region to the area of ​​the second bonding region is greater than the second threshold, it is determined that the activation state of the adhesive is unqualified.

5. The detection method according to claim 3, characterized in that: The determining the second slice image based on the first slice image comprises: improving the first image contrast of the first slice image to a second image contrast, and selecting a region of the first slice image having a grayscale value less than a first grayscale value threshold value at the second image contrast as the second slice image; The determining the second bonding area based on the second slice image comprises: Obtaining grayscale value data corresponding to each pixel in the second slice image; Acquire a first Gaussian model based on the gray value data; Determine a second gray value threshold range corresponding to the second bonding area based on a first parameter in the first Gaussian model, wherein the first parameter includes a mean value and a standard deviation; The second bonding area is determined based on the second gray value threshold range.

6. The detection method according to claim 5, characterized in that: Acquiring a first Gaussian model based on the gray value data includes: Processing the grayscale value data based on an expected maximum algorithm to obtain first parameters corresponding to second Gaussian models corresponding to different sub-distributions; Determine the Bayesian information criterion value corresponding to the first Gaussian model corresponding to the different sub-distributions based on the first parameter corresponding to the second Gaussian model corresponding to the different sub-distributions; A second Gaussian model whose Bayesian information criterion value is equal to the third threshold is selected as the first Gaussian model.

7. The detection method according to claim 5, characterized in that: The determining the area of ​​the air region in the second bonding region comprises: determining training data based on grayscale values ​​of the air region and the adhesive region in the second bonding region; Standardize the training data using the first Gaussian model to obtain standardized training data; Training the first logistic regression model based on the standardized training data to obtain a second logistic regression model; determining classification thresholds for the air region and the adhesive region based on a second logistic regression model; determining an air region and the adhesive region based on the classification threshold; Based on the air region in the second bonding region, the area of ​​the air region is determined.

8. The detection method according to claim 7, characterized in that: The step of training the first logistic regression model based on the standardized training data to obtain the second logistic regression model includes: Training each first logistic regression model based on the standardized training data to obtain each trained logistic regression model; Validate each trained logistic regression model by cross-validation to obtain accuracy data corresponding to each trained model; Determining a receiver operating characteristic curve based on the accuracy data corresponding to each of the logistic regression models; A data point at a first position in the receiver operating characteristic curve is determined, and a model trained with accuracy data corresponding to the data point is used as a second logistic regression model.

9. A storage medium, characterized in that: The storage medium stores instructions, which, when executed on an electronic device, enable the electronic device to implement the detection method according to any one of claims 1 to 8.

10. An electronic device, characterized in that: include: a memory for storing instructions executed by one or more processors in the electronic device; and a processor, configured to execute instructions stored in the memory to implement the detection method according to any one of claims 1 to 8.

Images