Intelligent electronic product manufacturing management method and system

Abstract

The invention discloses an intelligent electronic product manufacturing management method and system, and the method comprises the steps: obtaining a to-be-detected inductance image, extracting a winding image through a region growing algorithm, segmenting the winding image into a plurality of sub-images according to a winding mode, collecting normal and abnormal sample images to construct a data set, classifying the images through clustering analysis, and carrying out the recognition of the images. And determining the anomaly type of the inductor to be detected according to the distribution of the abnormal samples in the classification, identifying known and unknown anomalies, and optimizing the production process based on the detection result. According to the invention, the data set is formed by combining the normal and abnormal reference images, the abnormal type is identified by using clustering analysis, the detection precision and the automation degree are improved, the problem of inaccurate manual detection precision is effectively solved, and the product quality and the production efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to an intelligent electronic product manufacturing management method and system. Background Art

[0002] In the current field of electronic product manufacturing, inductors are one of the key components, and their performance directly affects the quality and reliability of the final product. Traditional inductor winding detection mainly relies on manual visual inspection, which has many limitations. First, due to the limitations of human eye recognition ability and subjective judgment, the detection accuracy is often not high, and it is difficult to find subtle or hidden defects. Secondly, manual inspection is inefficient and cannot meet the needs of modern high-density and large-scale production. In addition, as electronic equipment develops towards miniaturization and high performance, the design of inductors is becoming increasingly complex, and the quality requirements for the winding part are becoming higher and higher, which further aggravates the shortcomings of traditional detection methods. Summary of the invention

[0003] In order to solve at least one of the above-mentioned technical problems, the present invention provides an intelligent electronic product manufacturing management method and system.

[0004] In a first aspect, the present invention provides an intelligent electronic product manufacturing management method, the method comprising: Acquire an inductor image of the inductor to be tested; extract the winding part of the inductor image using a region growing algorithm to obtain a winding image; Based on the winding mode of the inductor to be detected, the winding image is divided using a corresponding preset spacing to obtain a plurality of first sub-winding images; Acquire a second sub-winding image as a normal reference and N third sub-winding images as abnormal references, wherein the N second inductance images correspond to the N abnormal types one by one; and integrate the first sub-winding image, the second sub-winding image and the third sub-winding image to obtain an image data set; Perform cluster analysis on the image data set to obtain M classified image sets; analyze the distribution of the second sub-winding images corresponding to the N abnormal types in the M classified image sets to determine the abnormal information of the first inductor image; Based on the abnormal information of the first inductance image, the product manufacturing process is optimized.

[0005] Preferably, analyzing the distribution of the second sub-winding images corresponding to the N abnormal types in the M classification image sets to determine the abnormal information of the first inductance image includes: Analyze the proportion of the second sub-winding image in the M classified image sets, determine the classified image set with the largest proportion of the second sub-winding image as a normal data set, and determine that the first sub-winding image in the corresponding classified image set is a normal type; If M=N+1, analyze the proportion of the third sub-winding images corresponding to different abnormal types in the M classification image sets, determine the abnormal type with the largest proportion in the classification image set as the abnormal type of the current classification image set in turn, and determine the abnormal type of the first sub-winding image in the corresponding classification image set; If M>N+1, analyze the proportion of the third sub-winding images corresponding to different abnormal types in the M classified image sets, determine the abnormal type with the largest proportion in the classified image set as the abnormal type of the current classified image set in turn, and determine the abnormal type of the first sub-winding image in the corresponding classified image set; mark the classified image set that does not contain the third sub-winding image as an unknown abnormal type, and determine the abnormal type of the first sub-winding image in the corresponding classified image set as an unknown abnormal type.

[0006] Preferably, after determining the abnormal type of the first sub-winding image in the classification image set based on the current classification image set, the method further includes: Analyze the position of the first sub-winding image corresponding to the abnormal type in the corresponding winding image to obtain abnormal position information; Determine the product manufacturing process optimization plan based on abnormal location information and historical maintenance data of the production line.

[0007] Preferably, the step of extracting the winding part of the inductor image by using a region growing algorithm to obtain the winding image includes: Grayscale the inductance image; Acquire the average grayscale of the third sub-winding image; Selecting seed points in the inductance image based on the average grayscale; The winding region is obtained by performing region growing according to the seed point, and the inductor image is cropped based on the winding region to obtain the winding image.

[0008] In a second aspect, the present invention further provides an intelligent electronic product manufacturing management system, the system comprising: The data acquisition module is used to acquire the inductance image of the inductor to be detected; the winding part of the inductance image is extracted by using a region growing algorithm to obtain a winding image; An image segmentation module, used to segment the winding image using a corresponding preset spacing based on the winding mode of the inductor to be detected, to obtain a plurality of first sub-winding images; A data integration module is used to obtain a second sub-winding image as a normal reference and N third sub-winding images as abnormal references, wherein the N second inductance images correspond to the N abnormal types one by one; and integrate the first sub-winding image, the second sub-winding image and the third sub-winding image to obtain an image data set; An image classification module is used to perform cluster analysis on the image data set to obtain M classified image sets; analyze the distribution of the second sub-winding images corresponding to N types of abnormalities in the M classified image sets to determine the abnormal information of the first inductor image; The product manufacturing process optimization module is used to optimize the product manufacturing process based on the abnormal information of the first inductance image.

[0009] Preferably, the image classification module is further used for: Analyze the proportion of the second sub-winding image in the M classified image sets, determine the classified image set with the largest proportion of the second sub-winding image as a normal data set, and determine that the first sub-winding image in the corresponding classified image set is a normal type; If M=N+1, analyze the proportion of the third sub-winding images corresponding to different abnormal types in the M classification image sets, determine the abnormal type with the largest proportion in the classification image set as the abnormal type of the current classification image set in turn, and determine the abnormal type of the first sub-winding image in the corresponding classification image set; If M>N+1, analyze the proportion of the third sub-winding images corresponding to different abnormal types in the M classified image sets, determine the abnormal type with the largest proportion in the classified image set as the abnormal type of the current classified image set in turn, and determine the abnormal type of the first sub-winding image in the corresponding classified image set; mark the classified image set that does not contain the third sub-winding image as an unknown abnormal type, and determine the abnormal type of the first sub-winding image in the corresponding classified image set as an unknown abnormal type.

[0010] Preferably, the image classification module is further used for: Analyze the position of the first sub-winding image corresponding to the abnormal type in the corresponding winding image to obtain abnormal position information; Determine the product manufacturing process optimization plan based on abnormal location information and historical maintenance data of the production line.

[0011] Preferably, the data acquisition module is also used for: Grayscale the inductance image; Acquire the average grayscale of the third sub-winding image; Selecting seed points in the inductance image based on the average grayscale; The winding region is obtained by performing region growing according to the seed point, and the inductor image is cropped based on the winding region to obtain the winding image.

[0012] In a third aspect, the present invention further provides an electronic device comprising a processor and a memory, wherein the memory is used to store computer program code, and the computer program code comprises computer instructions. When the processor executes the computer instructions, the electronic device executes the method as described in the first aspect above and any possible implementation thereof.

[0013] In a fourth aspect, the present invention further provides a computer-readable storage medium, wherein a computer program is stored in the computer-readable storage medium, wherein the computer program includes program instructions, and when the program instructions are executed by a processor of an electronic device, the processor executes the method as described in the first aspect above and any possible implementation thereof.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1) The present invention extracts the winding image by using a regional growing algorithm based on copper wire feature selection of seed points; retrieves the preset spacing from the database according to the winding method, and uses an image segmentation algorithm to segment the winding image into multiple first sub-winding images; then obtains normal and N types of abnormal reference sub-winding images, and integrates them to form an image data set; then uses the K-Means clustering algorithm to cluster and analyze the data set, and determines the abnormal information of the first inductor image according to the proportion of different sub-winding images in the classified image set; and then optimizes the product manufacturing process according to the abnormal information; the present application improves the detection accuracy, comprehensively detects abnormal types, optimizes the manufacturing process, and has a high degree of automation, which can effectively solve the problem of inaccurate manual detection of inductor windings and improve product quality and production efficiency.

[0015] 2) The present invention clusters the image data set into multiple classified image sets through the DBSCAN clustering algorithm, analyzes the distribution of normal and abnormal reference images in multiple classified image sets, accurately identifies known abnormal types, and can mark unknown abnormal types. It not only realizes high-precision automatic detection of inductor winding abnormalities, but also can comprehensively cover a variety of abnormal conditions including unknown fault modes, thereby significantly improving the product quality control level and the intelligence level of the manufacturing process.

[0016] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the background technology, the drawings required for use in the embodiments of the present invention or the background technology will be described below.

[0018] The drawings herein are incorporated into the specification and constitute a part of the specification. These drawings illustrate embodiments consistent with the present disclosure and are used to illustrate the technical solutions of the present disclosure together with the specification.

[0019] Figure 1 A flowchart of an intelligent electronic product manufacturing management method provided by an embodiment of the present invention; Figure 2 A schematic diagram of the structure of an intelligent electronic product manufacturing management system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0020] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0022] Existing inductor winding inspection methods mainly rely on manual visual inspection, which has the problems of low detection accuracy, low efficiency and difficulty in adapting to complex designs and diverse abnormal types.

[0023] See also Figure 1 , Figure 1 The following is a flow chart of an intelligent electronic product manufacturing management method provided by an embodiment of the present invention. Figure 1 As shown, the method includes: S100, obtaining an inductor image of the inductor to be tested; extracting the winding part of the inductor image by using a region growing algorithm to obtain a winding image; Use high-resolution industrial cameras to obtain full-range inductance images under specific lighting conditions.

[0024] The region growing algorithm is an algorithm for image segmentation based on the similarity between pixels in an image. In this step, first, one or more seed points are manually or automatically selected in the inductor image based on the color characteristics (usually golden yellow) and texture characteristics (with certain regularity) of the copper wire. These seed points are located in the copper wire winding part. Then, starting from the seed point, the adjacent pixels that are similar to the seed point pixel grayscale value, color, texture and other features and meet certain growth criteria (such as the set similarity threshold) are merged into the growth region. This process is repeated until no pixels that meet the conditions can be added to the growth region, thereby completely extracting the winding part in the inductor image and obtaining the winding image.

[0025] S200, based on the winding mode of the inductor to be detected, dividing the winding image by using a corresponding preset spacing to obtain a plurality of first sub-winding images; There are many winding methods for inductors, such as single-layer dense winding, multi-layer dense winding, and inter-winding. Different winding methods will result in different spacing between windings. The preset spacing is pre-stored in a computer database based on standard design data of different winding methods. During segmentation, the image recognition technology is used to first identify the winding method of the inductor to be tested, and then the corresponding preset spacing is retrieved from the database. The image segmentation algorithm is used to cut horizontally or vertically on the winding image according to the preset spacing, and the winding image is segmented into multiple first sub-winding images of equal width or length.

[0026] S300, acquiring a second sub-winding image as a normal reference and N third sub-winding images as abnormal references, wherein the N second inductance images correspond to the N abnormal types one by one; and integrating the first sub-winding image, the second sub-winding image and the third sub-winding image to obtain an image data set; The normal reference second sub-winding image is extracted and segmented from the inductor winding image that has been strictly tested and confirmed to be qualified. The N third sub-winding images used as abnormal references are obtained by simulating N different abnormal conditions (such as winding short circuit, open circuit, insufficient number of turns, uneven winding, etc.) in a laboratory environment, processing the inductor, obtaining its winding image and segmenting it. Each third sub-winding image corresponds to a specific abnormal type. The first sub-winding image obtained from the inductor to be tested, the normal reference second sub-winding image, and the N third sub-winding images of the abnormal reference are stored in a computer hard disk or other storage device according to a certain storage format (such as common image file formats JPEG, PNG, etc.) and organization method (such as establishing a folder structure to store different types of images respectively), forming an image data set containing multiple types of winding images for subsequent analysis and processing.

[0027] S400, performing cluster analysis on the image data set to obtain M classified image sets; analyzing the distribution of the second sub-winding images corresponding to the N abnormal types in the M classified image sets to determine the abnormal information of the first inductor image; The K-Means clustering algorithm is used, and the K value is set to N+1, where N corresponds to N known abnormal types, and the additional 1 represents the normal type. When performing the clustering operation, N+1 cluster centers are first randomly initialized. These cluster centers can be regarded as the initial representative points of the image features of each category. Then, for each image in the image data set, by calculating its distance from these N+1 cluster centers (Euclidean distance is selected as the distance measurement method, which can accurately measure the position difference of the image in the feature space), the image is divided into the category to which the nearest cluster center belongs. After completing a division, the feature mean of all images in each category is recalculated to update the position of the cluster center. This iteration is repeated until the position of the cluster center changes very little and reaches the convergence condition, and finally the image data set is successfully divided into N+1 classified image sets.

[0028] Check these N+1 classification image sets one by one, and determine the classification image set containing the largest number of normal reference second sub-winding images as the normal type data set. Because the image features in this classification image set are most similar to the normal second sub-winding image, it can be determined that the first sub-winding image in this classification image set is of normal type. For the remaining M classification image sets, the number of third sub-winding images corresponding to different abnormal types in each classification image set is counted respectively. The abnormal type corresponding to the third sub-winding image with the largest number in a certain classification image set is determined as the abnormal type of the classification image set, and then it can be determined that the first sub-winding image in the classification image set has the corresponding abnormal type.

[0029] S500, optimizing the product manufacturing process based on the abnormal information of the first inductance image.

[0030] According to the abnormal information of the first inductance image, the cause of the abnormality is analyzed. If it is a winding short circuit, it may be that the insulation layer of the copper wire is damaged during the winding process. It is necessary to check the pressure and speed parameters of the winding equipment and adjust the equipment to ensure that the insulation layer is not damaged during the winding process; if the number of turns is insufficient, it may be that the counting system of the winding equipment has deviations, and the counting system needs to be calibrated and maintained; if it is an unknown abnormal type, further in-depth research is required to find out the cause of the abnormality through experimental analysis, reference to materials, etc., and then optimize the corresponding links in the product manufacturing process in a targeted manner, such as improving process parameters, replacing equipment parts, strengthening quality inspection, etc., to improve product quality and production efficiency.

[0031] In this embodiment, a winding image is extracted by using a regional growing algorithm based on copper wire feature selection of seed points; a preset spacing is retrieved from a database according to the winding method, and an image segmentation algorithm is used to segment the winding image into multiple first sub-winding images; then normal and N types of abnormal reference sub-winding images are obtained and integrated to form an image data set; the K-Means clustering algorithm is then used to perform cluster analysis on the data set, and the abnormal information of the first inductor image is determined based on the proportion of different sub-winding images in the classified image set; and then the product manufacturing process is optimized based on the abnormal information; the present application improves detection accuracy, comprehensively detects abnormal types, optimizes manufacturing processes, and has a high degree of automation, which can effectively solve the problem of inaccurate manual detection of inductor windings and improve product quality and production efficiency.

[0032] Preferably, analyzing the distribution of the second sub-winding images corresponding to the N abnormal types in the M classification image sets to determine the abnormal information of the first inductance image includes: Analyze the proportion of the second sub-winding image in the M classified image sets, determine the classified image set with the largest proportion of the second sub-winding image as a normal data set, and determine that the first sub-winding image in the corresponding classified image set is a normal type; If M=N+1, analyze the proportion of the third sub-winding images corresponding to different abnormal types in the M classification image sets, determine the abnormal type with the largest proportion in the classification image set as the abnormal type of the current classification image set in turn, and determine the abnormal type of the first sub-winding image in the corresponding classification image set; If M>N+1, analyze the proportion of the third sub-winding images corresponding to different abnormal types in the M classified image sets, determine the abnormal type with the largest proportion in the classified image set as the abnormal type of the current classified image set in turn, and determine the abnormal type of the first sub-winding image in the corresponding classified image set; mark the classified image set that does not contain the third sub-winding image as an unknown abnormal type, and determine the abnormal type of the first sub-winding image in the corresponding classified image set as an unknown abnormal type.

[0033] In this embodiment, the DBSCAN clustering algorithm is used to divide the image data set. Before using the DBSCAN clustering algorithm, the image data needs to be preprocessed. The image is first grayed to remove color interference and highlight the texture and shape characteristics of the winding. Then, the image is denoised by Gaussian filtering to avoid the influence of noise on subsequent feature extraction and clustering. For the determination of the neighborhood radius and the minimum number of samples, the grid search method is used. Within the set reasonable parameter range, the image data set is clustered with different neighborhood radii and minimum number of samples. The clustering effect is evaluated using evaluation indicators such as the silhouette coefficient, and the parameter combination that optimizes the evaluation index is selected.

[0034] The DBSCAN clustering algorithm is used to cluster the image dataset. The DBSCAN algorithm identifies clusters based on density, which can effectively find clusters of any shape and mark noise points as data points that do not belong to any cluster. Through this method, the image dataset can be divided into multiple classified image sets, each of which represents a class of images with similar features.

[0035] When the number of classified image sets is exactly equal to the number of known abnormal types plus one normal classification (M=N+1), it means that except for one classification representing normality, each of the remaining classified image sets should theoretically correspond to a specific abnormal type. At this time, the distribution of the third sub-winding images corresponding to different abnormal types in all classified image sets is analyzed, and the abnormal type with the largest proportion in each classified image set is found, and it is designated as the main abnormal type of the classified image set. This step ensures that each known abnormal type can be accurately identified and accurately located in the corresponding classified image set.

[0036] When the number of classified image sets exceeds the number of known abnormal types plus one (M>N+1), it indicates that there are additional classified image sets that may contain unrecorded or new abnormal patterns. In this case, the known abnormal types are first identified and assigned to the respective classified image sets in the same manner as described above. Then, for those classified image sets that do not contain any third sub-winding images of known abnormal types, they are marked as unknown abnormal types, and the corresponding first sub-winding images in these classified image sets are also identified as unknown abnormal types. This strategy not only helps to identify common fault modes, but also captures potential new problems, thus providing direction for subsequent research.

[0037] In this embodiment, the DBSCAN clustering algorithm is used to cluster the image data set into multiple classified image sets, analyze the distribution of normal and abnormal reference images in the multiple classified image sets, accurately identify known abnormal types, and mark unknown abnormal types. This not only achieves high-precision automated detection of inductor winding abnormalities, but also comprehensively covers a variety of abnormal conditions including unknown fault modes, thereby significantly improving the product quality control level and the intelligence level of the manufacturing process.

[0038] Preferably, after determining the abnormal type of the first sub-winding image in the classification image set based on the current classification image set, the method further includes: Analyze the position of the first sub-winding image corresponding to the abnormal type in the corresponding winding image to obtain abnormal position information; Determine the product manufacturing process optimization plan based on abnormal location information and historical maintenance data of the production line.

[0039] After determining the abnormal type of the first sub-winding image in the classified image set, the specific positions of the first sub-winding images corresponding to these abnormal types in their original winding images are further analyzed. This process utilizes the method of marking the first sub-winding images according to the segmentation order when segmenting the winding image. For example, the position of the first segmented first sub-winding image is marked as 1, the second is marked as 2, and so on. This marking method not only helps to accurately identify the specific position of each first sub-winding image, but also can serve as an important basis for subsequent analysis. In this way, detailed abnormal position information can be obtained to clearly indicate which parts have problems. Based on the obtained abnormal position information and the historical maintenance data of the production line, a targeted product manufacturing process optimization plan is formulated. The historical maintenance data includes but is not limited to the location of similar faults in the past, the causes of these problems, and the solutions taken. Combined with this information, the possible weak links in the production process can be more accurately located, and the production process parameters or equipment settings can be adjusted accordingly, such as improving the working accuracy of the winding machine, adjusting the raw material supply specifications, etc., so as to effectively prevent the recurrence of the same or similar problems and improve the overall production efficiency and product quality.

[0040] In this embodiment, after determining the abnormal type of the first sub-winding image, the precise abnormal position information is obtained by analyzing the segmentation order marks of these abnormal images in their original winding images, and a targeted product manufacturing process optimization plan is formulated in combination with the historical maintenance data of the production line, thereby achieving accurate positioning of the abnormal position of the inductor winding, which not only improves the accuracy of fault identification, but also can perform effective preventive maintenance and process improvements based on historical data, significantly improving the accuracy and efficiency of product quality control, reducing the occurrence of repetitive faults, reducing maintenance costs, and promoting the transformation and upgrading of production processes to intelligence and automation, thereby enhancing the overall competitiveness and economic benefits of the enterprise.

[0041] Preferably, the step of extracting the winding part of the inductor image by using a region growing algorithm to obtain the winding image includes: Grayscale the inductance image; Acquire the average grayscale of the third sub-winding image; Selecting seed points in the inductance image based on the average grayscale; The winding region is obtained by performing region growing according to the seed point, and the inductor image is cropped based on the winding region to obtain the winding image.

[0042] The acquired inductance image is grayed out. Graying is the process of converting a color image into a grayscale image, which simplifies subsequent image processing tasks by reducing color information. The specific operation is to convert the RGB value of each pixel into a single grayscale value, usually calculated using a weighted average method, which can retain the main structural features of the original image and remove unnecessary color interference. The average grayscale value of the third sub-winding image as an abnormality reference is obtained. Here, the third sub-winding image refers to a set of image samples known to have a specific type of abnormality. By performing statistical analysis on the grayscale values ​​of these images, a representative average grayscale value is calculated.

[0043] Based on the average grayscale value obtained above, appropriate seed points are selected in the inductor image to be detected. The selection of seed points is crucial because they are the basis for the region growing algorithm to start expanding. In this method, the potential seed point position is determined based on the grayscale value similar to the abnormal winding part (i.e., close to the average grayscale value of the third sub-winding image). This step takes advantage of the difference in grayscale values ​​between copper wire and other materials to ensure that the selected seed point can accurately represent the starting position of the winding area. The region growing algorithm is executed based on the selected seed point, and it is gradually expanded until the preset conditions are met (such as reaching the boundary or the grayscale difference exceeds the threshold), thereby delineating the winding area. On this basis, the original inductor image is cropped to retain only the area containing the winding part, and finally a clear and focused winding image is obtained. This method not only improves the extraction accuracy of the winding part, but also lays a solid foundation for subsequent detailed analysis and fault detection.

[0044] In summary, the method provided in this embodiment can at least achieve the following effects: 1) The present invention extracts the winding image by using a regional growing algorithm based on copper wire feature selection of seed points; retrieves the preset spacing from the database according to the winding method, and uses an image segmentation algorithm to segment the winding image into multiple first sub-winding images; then obtains normal and N types of abnormal reference sub-winding images, and integrates them to form an image data set; then uses the K-Means clustering algorithm to cluster and analyze the data set, and determines the abnormal information of the first inductor image according to the proportion of different sub-winding images in the classified image set; and then optimizes the product manufacturing process according to the abnormal information; the present application improves the detection accuracy, comprehensively detects abnormal types, optimizes the manufacturing process, and has a high degree of automation, which can effectively solve the problem of inaccurate manual detection of inductor windings and improve product quality and production efficiency.

[0045] 2) The present invention clusters the image data set into multiple classified image sets through the DBSCAN clustering algorithm, analyzes the distribution of normal and abnormal reference images in multiple classified image sets, accurately identifies known abnormal types, and can mark unknown abnormal types. It not only realizes high-precision automatic detection of inductor winding abnormalities, but also can comprehensively cover a variety of abnormal conditions including unknown fault modes, thereby significantly improving the product quality control level and the intelligence level of the manufacturing process.

[0046] See also Figure 2 In one embodiment, an intelligent electronic product manufacturing management system is also provided, the system comprising: The data acquisition module 100 is used to acquire an inductance image of the inductor to be detected; a region growing algorithm is used to extract the winding part of the inductance image to obtain a winding image; An image segmentation module 200 is used to segment the winding image based on the winding mode of the inductor to be detected using a corresponding preset spacing to obtain a plurality of first sub-winding images; The data integration module 300 is used to obtain a second sub-winding image as a normal reference and N third sub-winding images as abnormal references, wherein the N second inductance images correspond to the N abnormal types one by one; and integrate the first sub-winding image, the second sub-winding image and the third sub-winding image to obtain an image data set; The image classification module 400 is used to perform cluster analysis on the image data set to obtain M classified image sets; analyze the distribution of the second sub-winding images corresponding to the N types of abnormalities in the M classified image sets to determine the abnormal information of the first inductor image; The product manufacturing process optimization module 500 is used to optimize the product manufacturing process based on the abnormal information of the first inductance image.

[0047] Preferably, the image classification module 400 is further used for: Analyze the proportion of the second sub-winding image in the M classified image sets, determine the classified image set with the largest proportion of the second sub-winding image as a normal data set, and determine that the first sub-winding image in the corresponding classified image set is a normal type; If M=N+1, analyze the proportion of the third sub-winding images corresponding to different abnormal types in the M classification image sets, determine the abnormal type with the largest proportion in the classification image set as the abnormal type of the current classification image set in turn, and determine the abnormal type of the first sub-winding image in the corresponding classification image set; If M>N+1, analyze the proportion of the third sub-winding images corresponding to different abnormal types in the M classified image sets, determine the abnormal type with the largest proportion in the classified image set as the abnormal type of the current classified image set in turn, and determine the abnormal type of the first sub-winding image in the corresponding classified image set; mark the classified image set that does not contain the third sub-winding image as an unknown abnormal type, and determine the abnormal type of the first sub-winding image in the corresponding classified image set as an unknown abnormal type.

[0048] Preferably, the image classification module 400 is further used for: Analyze the position of the first sub-winding image corresponding to the abnormal type in the corresponding winding image to obtain abnormal position information; Determine the product manufacturing process optimization plan based on abnormal location information and historical maintenance data of the production line.

[0049] Preferably, the data acquisition module 100 is also used for: Grayscale the inductance image; Acquire the average grayscale of the third sub-winding image; Selecting seed points in the inductance image based on the average grayscale; The winding region is obtained by performing region growing according to the seed point, and the inductor image is cropped based on the winding region to obtain the winding image.

[0050] It can be understood that the functions or modules included in the system provided in this embodiment can be used to execute the method described in the above method embodiment. Its specific implementation can refer to the description of the above method embodiment. For the sake of brevity, it will not be repeated here.

[0051] The present invention also provides an electronic device, including a processor and a memory, wherein the memory is used to store computer program code, and the computer program code includes computer instructions. When the processor executes the computer instructions, the electronic device executes a method as described in any possible implementation manner.

[0052] The present invention also provides a computer-readable storage medium, in which a computer program is stored. The computer program includes program instructions. When the program instructions are executed by a processor of an electronic device, the processor executes a method as described in any possible implementation manner.

[0053] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0054] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here. Those skilled in the art can also clearly understand that the descriptions of the various embodiments of the present invention have different focuses. For the convenience and brevity of description, the same or similar parts may not be repeated in different embodiments. Therefore, for parts not described or not described in detail in a certain embodiment, refer to the records of other embodiments.

Claims

1. An intelligent electronic product manufacturing management method, characterized in that: The method comprises: Acquire an inductor image of the inductor to be tested; extract the winding part of the inductor image using a region growing algorithm to obtain a winding image; Based on the winding mode of the inductor to be detected, the winding image is divided using a corresponding preset spacing to obtain a plurality of first sub-winding images; Acquire a second sub-winding image as a normal reference and N third sub-winding images as abnormal references, wherein the N second inductance images correspond to the N abnormal types one by one; and integrate the first sub-winding image, the second sub-winding image and the third sub-winding image to obtain an image data set; Perform cluster analysis on the image data set to obtain M classified image sets; analyze the distribution of the second sub-winding images corresponding to the N abnormal types in the M classified image sets to determine the abnormal information of the first inductor image; Based on the abnormal information of the first inductance image, the product manufacturing process is optimized.

2. The intelligent electronic product manufacturing management method according to claim 1, characterized in that: The analyzing the distribution of the second sub-winding images corresponding to the N abnormal types in the M classification image sets to determine the abnormal information of the first inductance image includes: Analyze the proportion of the second sub-winding image in the M classified image sets, determine the classified image set with the largest proportion of the second sub-winding image as a normal data set, and determine that the first sub-winding image in the corresponding classified image set is a normal type; If M=N+1, analyze the proportion of the third sub-winding images corresponding to different abnormal types in the M classification image sets, determine the abnormal type with the largest proportion in the classification image set as the abnormal type of the current classification image set in turn, and determine the abnormal type of the first sub-winding image in the corresponding classification image set; If M>N+1, analyze the proportion of the third sub-winding images corresponding to different abnormal types in the M classified image sets, determine the abnormal type with the largest proportion in the classified image set as the abnormal type of the current classified image set in turn, and determine the abnormal type of the first sub-winding image in the corresponding classified image set; mark the classified image set that does not contain the third sub-winding image as an unknown abnormal type, and determine the abnormal type of the first sub-winding image in the corresponding classified image set as an unknown abnormal type.

3. The intelligent electronic product manufacturing management method according to claim 2, characterized in that: After determining the abnormal type of the first sub-winding image in the classification image set based on the current classification image set, the method further includes: Analyze the position of the first sub-winding image corresponding to the abnormal type in the corresponding winding image to obtain abnormal position information; Determine the product manufacturing process optimization plan based on abnormal location information and historical maintenance data of the production line.

4. The intelligent electronic product manufacturing management method according to claim 1, characterized in that: The method of extracting the winding part of the inductor image by using a region growing algorithm to obtain the winding image includes: Grayscale the inductance image; Acquire the average grayscale of the third sub-winding image; Selecting seed points in the inductance image based on the average grayscale; The winding region is obtained by performing region growing according to the seed point, and the inductor image is cropped based on the winding region to obtain the winding image.

5. An intelligent electronic product manufacturing management system, characterized in that: The system comprises: The data acquisition module is used to acquire the inductance image of the inductor to be detected; the winding part of the inductance image is extracted by using a region growing algorithm to obtain a winding image; An image segmentation module, used to segment the winding image using a corresponding preset spacing based on the winding mode of the inductor to be detected, to obtain a plurality of first sub-winding images; A data integration module is used to obtain a second sub-winding image as a normal reference and N third sub-winding images as abnormal references, wherein the N second inductance images correspond to the N abnormal types one by one; and integrate the first sub-winding image, the second sub-winding image and the third sub-winding image to obtain an image data set; An image classification module is used to perform cluster analysis on the image data set to obtain M classified image sets; analyze the distribution of the second sub-winding images corresponding to N types of abnormalities in the M classified image sets to determine the abnormal information of the first inductor image; The product manufacturing process optimization module is used to optimize the product manufacturing process based on the abnormal information of the first inductance image.

6. The intelligent electronic product manufacturing management system according to claim 5, characterized in that: The image classification module is also used to: Analyze the proportion of the second sub-winding image in the M classified image sets, determine the classified image set with the largest proportion of the second sub-winding image as a normal data set, and determine that the first sub-winding image in the corresponding classified image set is a normal type; If M=N+1, analyze the proportion of the third sub-winding images corresponding to different abnormal types in the M classification image sets, determine the abnormal type with the largest proportion in the classification image set as the abnormal type of the current classification image set in turn, and determine the abnormal type of the first sub-winding image in the corresponding classification image set; If M>N+1, analyze the proportion of the third sub-winding images corresponding to different abnormal types in the M classification image sets, determine the abnormal type with the largest proportion in the classification image set as the abnormal type of the current classification image set in turn, and determine the abnormal type of the first sub-winding image in the corresponding classification image set; The classified image set that does not include the third sub-winding image is marked as an unknown abnormality type, and the abnormality type of the first sub-winding image in the corresponding classified image set is determined to be an unknown abnormality type.

7. The intelligent electronic product manufacturing management system according to claim 6, characterized in that: The image classification module is also used to: Analyze the position of the first sub-winding image corresponding to the abnormal type in the corresponding winding image to obtain abnormal position information; Determine the product manufacturing process optimization plan based on abnormal location information and historical maintenance data of the production line.

8. The intelligent electronic product manufacturing management system according to claim 5, characterized in that: The data acquisition module is also used for: Grayscale the inductance image; Acquire the average grayscale of the third sub-winding image; Selecting seed points in the inductance image based on the average grayscale; The winding region is obtained by performing region growing according to the seed point, and the inductor image is cropped based on the winding region to obtain the winding image.

9. An electronic device, characterized in that: include: A processor and a memory, wherein the memory is used to store computer program codes, wherein the computer program codes include computer instructions, and when the processor executes the computer instructions, the electronic device executes the intelligent electronic product manufacturing management method as described in any one of claims 1 to 4.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which includes program instructions. When the program instructions are executed by a processor of an electronic device, the processor executes the intelligent electronic product manufacturing management method according to any one of claims 1 to 4.

Images