An X-ray inspection image data management system and method for ceramic wine bottles

Abstract

This invention discloses an X-ray inspection image data management system and method for ceramic wine bottles, relating to the field of image processing technology. The management method includes the following steps: acquiring initial image data of the ceramic wine bottle and preprocessing it to generate standardized images; summarizing the standardized images to generate an inspection log; acquiring defect regions for each standardized image in any inspection log and determining the defect type in any defect region; extracting the X-ray shooting angle corresponding to any standardized image, generating matching parameters that present the defect situation, and obtaining the optimal shooting angle for any defect type in different defect regions; obtaining the X-ray shooting path when performing defect detection on the ceramic wine bottle in real time; setting the optimal shooting angle for X-ray, analyzing the actual defect situation presented by the optimal shooting angle, and making anomaly judgments for defect identification; reducing the manual visual judgment step and avoiding subjective errors.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, specifically to an X-ray detection image data management system and method for ceramic wine bottles. Background Technology

[0002] Currently, quality control is of paramount importance in the production of ceramic wine bottles. Traditional methods for inspecting the quality of ceramic wine bottles, such as manual visual inspection, are inefficient, inaccurate, and unable to detect minute defects. However, non-destructive testing using X-ray imaging technology can effectively identify internal defects in ceramic wine bottles, improving inspection efficiency and accuracy. However, existing X-ray imaging-based methods for inspecting ceramic wine bottles still have shortcomings. A single shooting angle cannot cover the entire bottle body, and defects on the bottle shoulder and bottom require a specific tilt angle to be visible. These methods rely heavily on manual parameter adjustment and visual judgment, which is time-consuming and affects the comprehensive presentation of defects, seriously impacting the production process. Summary of the Invention

[0003] The purpose of this invention is to provide an X-ray inspection image data management system and method for ceramic wine bottles, so as to solve the problems raised in the prior art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a method for managing X-ray inspection image data of ceramic wine bottles, the method comprising the following steps: Step S100: Acquire the initial image data generated by X-Ray during any inspection of the ceramic wine bottle, preprocess the initial image data to generate standardized images; summarize the various standardized images generated during the inspection process to generate an inspection log of the inspection process; Step S200: Perform defect detection on each standardized image in any inspection log to obtain the defect area of ​​the ceramic wine bottle in any inspection log; analyze the defect situation in any defect area to determine the defect type in any defect area; Step S300: Extract the shooting angle of X-Ray corresponding to any standardized image, and generate matching parameters for the defect situation based on the defect situation presented by the extracted standardized image; analyze the difference of defect area in different standardized images of the same defect type, and obtain the best shooting angle of any defect type in different defect areas. Step S400: When performing real-time defect detection on ceramic wine bottles, the X-Ray shooting path is obtained based on the occurrence frequency of each defect area; the defect area is identified based on the defect situation identified on the shooting path, and the optimal shooting angle for X-Ray is set according to the defect type in the defect area. Step S500: Predict the expected recognition of any target defect area at the optimal shooting angle; analyze the actual defect situation presented by the optimal shooting angle, and make anomaly judgments on the defect recognition situation.

[0005] Furthermore, step S100 includes the following steps: Step S101: Set a preset illumination angle for the X-Ray relative to the ceramic bottle, pass the X-Ray through the ceramic bottle at the preset illumination angle, and simultaneously start the flat panel detector to capture the image of the X-Ray passing through the ceramic bottle, thereby obtaining the initial image data corresponding to the preset illumination angle; by acquiring multiple frames of images from the flat panel detector, and averaging the gray values ​​of each corresponding pixel in each frame to obtain a detection image, the initial image data is obtained. Step S102: Perform Gamma transformation, USM enhancement, and contrast adjustment on the initial image data sequentially to generate a corresponding standardized image. To improve the efficiency of Gamma transformation, a transformation mapping table is pre-generated, and the transformation is quickly implemented by looking up the table. The image after Gamma transformation is then enhanced using the USM algorithm: the image after Gamma transformation is set as X, and Gaussian filtering is performed. The filtering parameters are manually adjusted and determined through the software interface. The filtered image is set as G, then the mask M = XG, the enhanced image Y = X + λ × M, and the enhancement coefficient λ is also manually adjusted and determined through the software interface. The gray values ​​of the enhanced image are highly concentrated in a small gray range. The high and low thresholds of the gray range are manually selected through the software interface, and the gray values ​​between the high and low thresholds are linearly mapped to between 0 and 255. Through the above steps, the defects in the standardized image are made visible instead of invisible. Step S103: Continuously adjust the X-Ray illumination angle to obtain standardized images at each illumination angle. Summarize the standardized images at all illumination angles to generate a detection log for one detection process. Adjusting the X-Ray illumination angle is necessary because different defects in different areas cannot be fully identified by a fixed illumination angle. It is necessary to take images in different areas to better capture all defects.

[0006] Furthermore, step S200 includes the following steps: Step S201: Pre-build a defect database, which stores several defect types. For any defect type, there is a corresponding set of defect feature groups. The defect feature group consists of defect attributes and defect parameters. Summarize the defect attributes of all defect types to obtain a set of defect attributes. Step S202: Randomly select a detection log and randomly extract a standardized image from the selected detection log. Randomly segment the extracted standardized image to obtain several candidate regions. Randomly extract a candidate region and obtain the defect parameters of each defect attribute in the extracted candidate region. If a defect attribute has corresponding defect parameters, then the extracted candidate region is set as an abnormal region. Step S203: Obtain each abnormal region from the extracted standardized image. If there is a common boundary between two abnormal regions, compare the defect attributes of the two abnormal regions to obtain the number of identical defect attributes between the two abnormal regions as A. sim The correlation between the two anomalous regions was calculated as G=A. sim / Min(A1,A2), where Min() is the minimum value function, A1 is the number of defective attributes in one of the abnormal regions, and A2 is the number of defective attributes in another abnormal region; a preset correlation threshold G is used. th If G≥G th If the two abnormal regions are merged into a new defect region, several defect regions will be regenerated from the extracted standardized image. Since the initial region division is random, there may be cases where the defect is divided into two regions. It is necessary to compare the defect attributes of the two adjacent regions. If the defect attributes are highly similar, it means that the two regions have identified the same defect. Therefore, the two regions need to be merged to make the subsequent defect analysis more complete and accurate. Step S204: Randomly select a defect area, and generate actual defect feature groups from the defect attributes of the selected defect area that contain defect parameters; randomly select a defect type from the defect database, and compare each actual defect feature group of the selected defect area with the defect feature group set of the selected defect type. If there is a defect feature group in the defect feature group set whose defect attribute is the same as the defect attribute of the actual defect feature group, and the defect parameter contains the defect parameter of the actual defect feature group, then set the actual defect feature group as a target defect feature group of the selected defect type. Step S205: Count the number of target defect feature groups for each defect type in the selected defect region, and set the number of target defect feature groups for the i-th defect type as b. i A preset threshold b for the number of feature groups is given. th If b i ≥b th Then, the i-th defect type is set as a defect type of the selected defect area, resulting in several defect types of the selected defect area. Multiple defect types may exist in one area at the same time, so all of them need to be captured, which is beneficial for subsequent adjustment of the shooting angle.

[0007] Furthermore, step S300 includes the following steps: Step S301: Randomly select a detection log, randomly select a standardized image from the selected detection log, obtain the illumination angle corresponding to the selected standardized image, and extract the defect region in the selected standardized image to obtain several defect types for each defect region; arbitrarily select a defect type, obtain the actual defect feature group of the selected defect type in the selected standardized image, and summarize the illumination angle, the selected defect type, the defect region where the selected defect type is located, and the actual defect feature group corresponding to the selected defect type to generate a matching parameter group of the selected defect type in the corresponding defect region; Step S302: Randomly select a defect type, obtain each matching parameter group of the selected defect type in the selection detection log, if two matching parameter groups are the same in the defect area, set the two matching parameter groups as the same area parameter group, and divide the selected defect type into the same area parameter group set of each defect area in the selection detection log. Step S303: Obtain the defect area of ​​the selected defect type in different detection logs. If there is an overlapping area between the defect areas in different detection logs, the area of ​​the overlapping area is S. over The defect areas from the two detection logs are combined to obtain a merged region, the area of ​​which is S. merge The degree of overlap between the defect areas in the two detection logs was calculated as C=S. over / S merge A preset overlap threshold C is used. th If C≥C th If the defect areas of the two detection logs are set as the same defect area, the matching parameter group of the same defect area in the two detection logs is set as the same area parameter group. Since the area division in different detection logs will be different, because the defects in different detection logs are also different, but X-Ray imaging has a shooting range, under the same shooting angle, if the overlap area of ​​the defect areas is large, the two defect areas can be reasonably approximated as the same area. Step S304: Obtain the parameter groups of each region in the selected detection log for the selected defect type. Randomly select the i-th defect attribute from the actual defect feature group of the region parameter group to obtain the defect parameter of the i-th defect attribute in each region parameter group. Obtain the illumination angle of each region parameter group and sort the region parameter groups according to the illumination angle from smallest to largest. Select the region parameter group with the largest defect parameter corresponding to the i-th defect attribute as the high-quality parameter group, and divide all region parameter groups into two groups according to the high-quality parameter group. Any defect attribute will increase or decrease with the change of shooting angle. Therefore, the peak value is the shooting situation of the best shooting angle, which can reflect the defect degree to the greatest extent. Step S305: Randomly select two adjacent parameter groups from any set of parameters in the same region, obtain the difference in illumination angle ΔT and the difference in defect parameters ΔD between the two adjacent parameter groups, and calculate the parameter change rate η = ΔD / ΔT between the two adjacent parameter groups; obtain the parameter change rate between any two adjacent parameter groups in the two sets of parameters in the same region, and calculate the average change rate η between the two sets of parameters in the same region. ave ; Step S306: Obtain the average change rate of the high-quality parameter group and the two sets of parameters in the same area corresponding to each defect attribute in each detection log; arbitrarily select an illumination angle and obtain the illumination angle difference ΔT between the high-quality parameter group of the i-th defect attribute and the selected illumination angle. ’ The defect parameter D is obtained from the high-quality parameter group of the i-th defect attribute. i The expected defect parameter of the i-th defect attribute at the selected illumination angle is calculated to be (D). i ) ex =D i -ΔT ’ ×(η i ) ave , where (η i ) ave Let be the average rate of change of the i-th defect attribute; obtain the expected defect parameters of each defect attribute at the selected illumination angle, and sum them to obtain the defect feature value D at the selected illumination angle. T The defect feature values ​​of each illumination angle are obtained, and the illumination angle with the largest defect feature value is selected as the best shooting angle for the defect type in the defect area.

[0008] Furthermore, step S400 includes the following steps: Step S401: Obtain the defect areas of the ceramic wine bottle in each inspection log, and count the number of defects occurring in each defect area in each inspection log. Let m be the number of defects occurring in the j-th defect area. jThe occurrence frequency f=m of the j-th defect region is calculated. j / M, where M is the total number of detection logs; sort each defect area from high to low according to the frequency of occurrence, and generate the X-Ray imaging path; Step S402: The mobile X-Ray takes pictures of each area according to the generated shooting path to obtain real-time standardized images of each area. The defect parameters of each defect attribute in any real-time standardized image are obtained. If a defect attribute has corresponding defect parameters, the area where it is located is set as the target defect area. Step S403: Obtain the defect feature groups of each defect attribute in the target defect area, compare each defect feature group with the defect feature group set of various defect types in the defect database to obtain the defect type of the target defect area; if there is only one defect type in the target defect area, then set the best shooting angle of the unique defect type as the best shooting angle of the target defect area; if there are several defect types in the target defect area, then set the corresponding best shooting angles as the best shooting angles of the target defect area.

[0009] Furthermore, step S500 includes the following steps: Step S501: Arbitrarily select a target defect area. When the X-Ray image is taken of the selected target defect area from the optimal shooting angle of the previous target defect area, a first standardized image is generated. Set the optimal shooting angle of the previous target defect area as the initial shooting angle, and obtain the defect parameters corresponding to each defect attribute of the first standardized image at the initial shooting angle. Set the defect parameter corresponding to the i-th defect attribute as D1. i ; Step S502: Obtain the defect type of the selected target defect area and obtain the optimal shooting angle T for the selected target defect area. best The initial shooting angle is set to T0, and the change in shooting angle is calculated to be ΔT. best =T best -T0, setting the average rate of change of the i-th defect attribute as (η) i ) ave The expected defect parameter D2 of the i-th defect attribute under the optimal shooting angle is calculated. i The expected defect feature value (D) is obtained by summing the expected defect parameters of each defect attribute. T ) ex ; Step S503: Acquire the actual defect feature groups of the selected target defect area at the optimal shooting angle, obtain the defect parameters corresponding to each defect attribute, and sum them to obtain the actual defect feature value (D). T ) acThe defect detection deviation rate δ = |(D) of the selected target defect area was calculated. T ) ex -(D T ) ac | / (D T ) ex Preset a deviation rate threshold δ th If δ≥δ th If so, an abnormal alert will be issued regarding the defect identification status of the selected target defect area.

[0010] To better implement the above methods, an X-Ray inspection image data management system is also proposed. The management system includes an image generation and processing module, a defect identification and classification module, an image shooting parameter matching module, an angle adjustment and optimization module, and an anomaly identification and judgment module. The image generation and processing module is used to acquire initial image data generated by X-Ray during any inspection of ceramic wine bottles, preprocess the initial image data to generate standardized images, and summarize the various standardized images generated during the inspection process to generate an inspection log. The defect identification and segmentation module is used to perform defect detection on each standardized image in any inspection log to obtain the defect area of ​​the ceramic wine bottle in any inspection log; and to analyze the defect situation in any defect area to determine the defect type in any defect area. The shooting parameter matching module is used to extract the shooting angle of X-Ray corresponding to any standardized image, and generate matching parameters based on the defect situation presented in the extracted standardized image; it analyzes the difference of defect area in different standardized images of the same defect type to obtain the optimal shooting angle of any defect type in different defect areas. The angle adjustment and optimization module is used to obtain the X-Ray shooting path based on the occurrence frequency of each defect area when performing real-time defect detection on ceramic wine bottles; to identify the defect area based on the defect situation identified on the shooting path; and to set the optimal shooting angle for X-Ray according to the defect type in the defect area. The anomaly detection module is used to predict the expected recognition of any target defect area under the optimal shooting angle; analyze the actual defect situation presented by the optimal shooting angle, and make anomaly judgments on the defect recognition situation.

[0011] Furthermore, the defect identification and classification module includes an image defect identification unit and a defect type classification unit; The image defect recognition unit is used to perform defect detection on each standardized image in any inspection log to obtain the defect area of ​​the ceramic wine bottle in any inspection log; the defect type classification unit is used to analyze the defect situation in any defect area and determine the defect type in any defect area.

[0012] Furthermore, the shooting parameter matching module includes a shooting angle extraction unit and a shooting angle matching unit; The shooting angle extraction unit is used to extract the shooting angle of X-Ray corresponding to any standardized image, and generate matching parameters for the defect situation based on the defect situation presented in the extracted standardized image; the shooting angle matching unit is used to analyze the difference of defect area in different standardized images of the same defect type, and obtain the optimal shooting angle of any defect type in different defect areas.

[0013] Furthermore, the angle adjustment and optimization module includes an actual defect detection unit and an optimal angle setting unit; The actual defect detection unit is used to obtain the X-Ray shooting path based on the occurrence frequency of each defect area when performing real-time defect detection on ceramic wine bottles; the optimal angle setting unit is used to identify defect areas based on the defect situation identified on the shooting path, and set the optimal shooting angle for X-Ray according to the defect type in the defect area.

[0014] Compared with the prior art, the beneficial effects of the present invention are: This invention solves the problem that a single angle cannot accurately identify defects in different areas of a wine bottle by establishing a matching relationship between different defect types and the optimal shooting angle; it also automatically selects the shooting angle with the largest defect feature value, significantly improving defect identification efficiency. 2. This invention dynamically generates a detection path based on the frequency of defect occurrence in the defect area. It first scans high-frequency defect areas to reduce invalid images and can capture defects most quickly. Furthermore, it automatically calls the pre-stored optimal shooting angle according to the identified defect type, without the need for manual intervention in parameter adjustment, thus improving the automation level of the detection process. 3. This invention reduces the need for manual visual judgment by moving from image preprocessing and defect database matching to optimal angle decision-making, thus avoiding subjective errors and achieving fully automated management of the entire process, significantly improving production efficiency. Attached Figure Description

[0015] Figure 1 A schematic diagram illustrating the steps of an X-ray inspection image data management method for ceramic wine bottles; Figure 2 This is a schematic diagram of the structure of an X-ray detection image data management system for a ceramic wine bottle. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Example: Figures 1 to 2 As shown, this invention provides a method for managing X-ray inspection image data of ceramic wine bottles. The management method includes the following steps: Step S100: Acquire the initial image data generated by X-Ray during any inspection of the ceramic wine bottle, preprocess the initial image data to generate standardized images; summarize the various standardized images generated during the inspection process to generate an inspection log of the inspection process; Step S100 includes the following steps: Step S101: Set a preset illumination angle for X-Ray relative to the ceramic bottle, pass X-Ray through the ceramic bottle at the preset illumination angle, and simultaneously start the flat panel detector to capture the image of X-Ray passing through the ceramic bottle to obtain the initial image data corresponding to the preset illumination angle. Step S102: Perform Gamma transformation, USM enhancement, and contrast adjustment on the initial image data in sequence to generate the corresponding standardized image; Step S103: Continuously adjust the X-Ray illumination angle to obtain standardized images at each illumination angle. Summarize the standardized images at all illumination angles to generate a detection log for one detection process.

[0018] Step S200: Perform defect detection on each standardized image in any inspection log to obtain the defect area of ​​the ceramic wine bottle in any inspection log; analyze the defect situation in any defect area to determine the defect type in any defect area; Step S200 includes the following steps: Step S201: Pre-build a defect database, which stores several defect types. For any defect type, there is a corresponding set of defect feature groups. The defect feature group consists of defect attributes and defect parameters. Summarize the defect attributes of all defect types to obtain a set of defect attributes. Step S202: Randomly select a detection log and randomly extract a standardized image from the selected detection log. Randomly segment the extracted standardized image to obtain several candidate regions. Randomly extract a candidate region and obtain the defect parameters of each defect attribute in the extracted candidate region. If a defect attribute has corresponding defect parameters, then the extracted candidate region is set as an abnormal region. Step S203: Obtain each abnormal region from the extracted standardized image. If there is a common boundary between two abnormal regions, compare the defect attributes of the two abnormal regions to obtain the number of identical defect attributes between the two abnormal regions as A. sim The correlation between the two anomalous regions was calculated as G=A. sim / Min(A1,A2), where Min() is the minimum value function, A1 is the number of defective attributes in one of the abnormal regions, and A2 is the number of defective attributes in another abnormal region; a preset correlation threshold G is used. th If G≥G th Then, the two abnormal regions are merged into a new defect region, and several defect regions are regenerated from the extracted standardized image. Example 1: The number of defect attributes in defect region 1 is set to 5, namely bubble diameter, crack length, density, thickness and impurity quantity, and the number of defect attributes in defect region 2 is set to 3, namely bubble diameter, crack length and density. The two defect regions have a common boundary, and the number of the same defect attributes in the two defect regions is 3. The correlation degree G is 3 / 3 = 100%. The preset correlation degree threshold is 80%. Therefore, defect region 1 and defect region 2 are merged. Step S204: Randomly select a defect area, and generate actual defect feature groups from the defect attributes of the selected defect area that contain defect parameters; randomly select a defect type from the defect database, and compare each actual defect feature group of the selected defect area with the defect feature group set of the selected defect type. If there is a defect feature group in the defect feature group set whose defect attribute is the same as the defect attribute of the actual defect feature group, and the defect parameter contains the defect parameter of the actual defect feature group, then set the actual defect feature group as a target defect feature group of the selected defect type. Step S205: Count the number of target defect feature groups for each defect type in the selected defect region, and set the number of target defect feature groups for the i-th defect type as b. i A preset threshold b for the number of feature groups is given. th If b i ≥b th Then, the i-th defect type is set as a defect type for the selected defect area, resulting in several defect types for the selected defect area.

[0019] Step S300: Extract the shooting angle of X-Ray corresponding to any standardized image, and generate matching parameters for the defect situation based on the defect situation presented by the extracted standardized image; analyze the difference of defect area in different standardized images of the same defect type, and obtain the best shooting angle of any defect type in different defect areas. Step S300 includes the following steps: Step S301: Randomly select a detection log, randomly select a standardized image from the selected detection log, obtain the illumination angle corresponding to the selected standardized image, and extract the defect region in the selected standardized image to obtain several defect types for each defect region; arbitrarily select a defect type, obtain the actual defect feature group of the selected defect type in the selected standardized image, and summarize the illumination angle, the selected defect type, the defect region where the selected defect type is located, and the actual defect feature group corresponding to the selected defect type to generate a matching parameter group of the selected defect type in the corresponding defect region; Step S302: Randomly select a defect type, obtain each matching parameter group of the selected defect type in the selection detection log, if two matching parameter groups are the same in the defect area, set the two matching parameter groups as the same area parameter group, and divide the selected defect type into the same area parameter group set of each defect area in the selection detection log. Step S303: Obtain the defect area of ​​the selected defect type in different detection logs. If there is an overlapping area between the defect areas in different detection logs, the area of ​​the overlapping area is S. over The defect areas from the two detection logs are combined to obtain a merged region, the area of ​​which is S. merge The degree of overlap between the defect areas in the two detection logs was calculated as C=S. over / S merge A preset overlap threshold C is used. th If C≥C th If so, the defect areas of the two detection logs will be set to the same defect area, and the matching parameter group of the same defect area in the two detection logs will be set to the same area parameter group. Step S304: Obtain the parameter groups of each same area in the selected detection log for the selected defect type. Randomly select the i-th defect attribute from the actual defect feature group of the same area parameter group to obtain the defect parameter of the i-th defect attribute in each same area parameter group. Obtain the illumination angle of each same area parameter group respectively, and sort the same area parameter groups according to the illumination angle from smallest to largest. Select the same area parameter group with the largest defect parameter corresponding to the i-th defect attribute as the high-quality parameter group, and divide all same area parameter groups into two same area parameter groups according to the high-quality parameter group. Step S305: Randomly select two adjacent parameter groups from any set of parameters in the same region, obtain the difference in illumination angle ΔT and the difference in defect parameters ΔD between the two adjacent parameter groups, and calculate the parameter change rate η = ΔD / ΔT between the two adjacent parameter groups; obtain the parameter change rate between any two adjacent parameter groups in the two sets of parameters in the same region, and calculate the average change rate η between the two sets of parameters in the same region. ave ; Step S306: Obtain the average change rate of the high-quality parameter group and the two sets of parameters in the same area corresponding to each defect attribute in each detection log; arbitrarily select an illumination angle and obtain the illumination angle difference ΔT between the high-quality parameter group of the i-th defect attribute and the selected illumination angle. ’ The defect parameter D is obtained from the high-quality parameter group of the i-th defect attribute. i The expected defect parameter of the i-th defect attribute at the selected illumination angle is calculated to be (D). i ) ex =D i -ΔT ’ ×(η i ) ave , where (η i ) ave Let be the average rate of change of the i-th defect attribute; obtain the expected defect parameters of each defect attribute at the selected illumination angle, and sum them to obtain the defect feature value D at the selected illumination angle. T The defect feature values ​​of each illumination angle are obtained, and the illumination angle with the largest defect feature value is selected as the best shooting angle for the defect type in the defect area. Example 2: With an irradiation angle of 30° and bubble diameter selected as the defect attribute, an irradiation angle of 45° and a bubble diameter of 2.5 mm were obtained in the high-quality parameter group. The average rate of change of the bubble diameter was set to 0.02 mm, and the expected defect parameter (D) was calculated at an irradiation angle of 30°. i ) ex =2.5 - 0.02 × (45 - 30) = 2.2 mm; Step S400: When performing real-time defect detection on ceramic wine bottles, the X-Ray shooting path is obtained based on the occurrence frequency of each defect area; the defect area is identified based on the defect situation identified on the shooting path, and the optimal shooting angle for X-Ray is set according to the defect type in the defect area. Step S400 includes the following steps: Step S401: Obtain the defect areas of the ceramic wine bottle in each inspection log, and count the number of defects occurring in each defect area in each inspection log. Let m be the number of defects occurring in the j-th defect area.j The occurrence frequency f=m of the j-th defect region is calculated. j / M, where M is the total number of detection logs; sort each defect area from high to low according to the frequency of occurrence, and generate the X-Ray imaging path; Step S402: The mobile X-Ray takes pictures of each area according to the generated shooting path to obtain real-time standardized images of each area. The defect parameters of each defect attribute in any real-time standardized image are obtained. If a defect attribute has corresponding defect parameters, the area where it is located is set as the target defect area. Step S403: Obtain the defect feature groups of each defect attribute in the target defect area, compare each defect feature group with the defect feature group set of various defect types in the defect database to obtain the defect type of the target defect area; if there is only one defect type in the target defect area, then set the best shooting angle of the unique defect type as the best shooting angle of the target defect area; if there are several defect types in the target defect area, then set the corresponding best shooting angles as the best shooting angles of the target defect area.

[0020] Step S500: Predict the expected recognition of any target defect area at the optimal shooting angle; analyze the actual defect situation presented by the optimal shooting angle, and make anomaly judgments on the defect recognition situation; Step S500 includes the following steps: Step S501: Arbitrarily select a target defect area. When the X-Ray image is taken of the selected target defect area from the optimal shooting angle of the previous target defect area, a first standardized image is generated. Set the optimal shooting angle of the previous target defect area as the initial shooting angle, and obtain the defect parameters corresponding to each defect attribute of the first standardized image at the initial shooting angle. Set the defect parameter corresponding to the i-th defect attribute as D1. i ; Step S502: Obtain the defect type of the selected target defect area and obtain the optimal shooting angle T for the selected target defect area. best The initial shooting angle is set to T0, and the change in shooting angle is calculated to be ΔT. best =T best -T0, setting the average rate of change of the i-th defect attribute as (η) i ) ave The expected defect parameter D2 of the i-th defect attribute under the optimal shooting angle is calculated. i The expected defect feature value (D) is obtained by summing the expected defect parameters of each defect attribute. T ) ex ; Step S503: Acquire the actual defect feature groups of the selected target defect area at the optimal shooting angle, obtain the defect parameters corresponding to each defect attribute, and sum them to obtain the actual defect feature value (D). T ) ac The defect detection deviation rate δ = |(D) of the selected target defect area was calculated. T ) ex -(D T ) ac | / (D T ) ex Preset a deviation rate threshold δ th If δ≥δ th If so, an abnormal alert will be issued regarding the defect identification status of the selected target defect area.

[0021] An X-Ray inspection image data management system, the management system includes an image generation and processing module, a defect identification and classification module, an imaging parameter matching module, an angle adjustment and optimization module, and an anomaly identification and judgment module; The image generation and processing module is used to acquire initial image data generated by X-Ray during any inspection of ceramic wine bottles, preprocess the initial image data to generate standardized images, and summarize the various standardized images generated during the inspection process to generate an inspection log. The defect identification and segmentation module is used to perform defect detection on each standardized image in any inspection log to obtain the defect area of ​​the ceramic wine bottle in any inspection log; and to analyze the defect situation in any defect area to determine the defect type in any defect area. The shooting parameter matching module is used to extract the shooting angle of X-Ray corresponding to any standardized image, and generate matching parameters based on the defect situation presented in the extracted standardized image; it analyzes the difference of defect area in different standardized images of the same defect type to obtain the optimal shooting angle of any defect type in different defect areas. The angle adjustment and optimization module is used to obtain the X-Ray shooting path based on the occurrence frequency of each defect area when performing real-time defect detection on ceramic wine bottles; to identify the defect area based on the defect situation identified on the shooting path; and to set the optimal shooting angle for X-Ray according to the defect type in the defect area. The anomaly detection module is used to predict the expected recognition of any target defect area under the optimal shooting angle; analyze the actual defect situation presented by the optimal shooting angle, and make anomaly judgments on the defect recognition situation.

[0022] The defect identification and classification module includes an image defect identification unit and a defect type classification unit. The image defect recognition unit is used to perform defect detection on each standardized image in any inspection log to obtain the defect area of ​​the ceramic wine bottle in any inspection log; the defect type classification unit is used to analyze the defect situation in any defect area and determine the defect type in any defect area.

[0023] The shooting parameter matching module includes a shooting angle extraction unit and a shooting angle matching unit; The shooting angle extraction unit is used to extract the shooting angle of X-Ray corresponding to any standardized image, and generate matching parameters for the defect situation based on the defect situation presented in the extracted standardized image; the shooting angle matching unit is used to analyze the difference of defect area in different standardized images of the same defect type, and obtain the optimal shooting angle of any defect type in different defect areas.

[0024] The angle adjustment and optimization module includes an actual defect detection unit and an optimal angle setting unit. The actual defect detection unit is used to obtain the X-Ray shooting path based on the occurrence frequency of each defect area when performing real-time defect detection on ceramic wine bottles; the optimal angle setting unit is used to identify defect areas based on the defect situation identified on the shooting path, and set the optimal shooting angle for X-Ray according to the defect type in the defect area.

[0025] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for managing X-ray inspection image data of ceramic wine bottles, characterized in that: The management method includes the following steps: Step S100: Acquire the initial image data generated by X-Ray during any inspection of the ceramic wine bottle, preprocess the initial image data to generate standardized images; summarize the various standardized images generated during the inspection process to generate an inspection log of the inspection process; Step S200: Perform defect detection on each standardized image in any inspection log to obtain the defect area of ​​the ceramic wine bottle in any inspection log; analyze the defect situation in any defect area to determine the defect type in any defect area; Step S300: Extract the shooting angle of X-Ray corresponding to any standardized image, and generate matching parameters for the defect situation based on the defect situation presented by the extracted standardized image; analyze the difference of defect area in different standardized images of the same defect type, and obtain the best shooting angle of any defect type in different defect areas. Step S400: When performing real-time defect detection on ceramic wine bottles, the X-Ray shooting path is obtained based on the occurrence frequency of each defect area; the defect area is identified based on the defect situation identified on the shooting path, and the optimal shooting angle for X-Ray is set according to the defect type in the defect area. Step S500: Predict the expected recognition of any defect area at the optimal shooting angle; analyze the actual defect situation presented by the optimal shooting angle, and make anomaly judgments on the defect recognition situation; Step S300 includes the following steps: Step S301: Randomly select a detection log, randomly select a standardized image from the selected detection log, obtain the illumination angle corresponding to the selected standardized image, and extract the defect region in the selected standardized image to obtain several defect types for each defect region; arbitrarily select a defect type, obtain the actual defect feature group of the selected defect type in the selected standardized image, and summarize the illumination angle, the selected defect type, the defect region where the selected defect type is located, and the actual defect feature group corresponding to the selected defect type to generate a matching parameter group of the selected defect type in the corresponding defect region; Step S302: Randomly select a defect type, obtain each matching parameter group of the selected defect type in the selection detection log, if two matching parameter groups are the same in the defect area, set the two matching parameter groups as the same area parameter group, and divide the selected defect type into the same area parameter group set of each defect area in the selection detection log. Step S303: Obtain the defect area of ​​the selected defect type in different detection logs. If there is an overlapping area between the defect areas in different detection logs, the area of ​​the overlapping area is S. over The defect areas from the two detection logs are combined to obtain a merged region, the area of ​​which is S. merge The degree of overlap between the defect areas in the two detection logs was calculated as C=S. over / S merge A preset overlap threshold C is used. th If C≥C th If so, the defect areas of the two detection logs will be set to the same defect area, and the matching parameter group of the same defect area in the two detection logs will be set to the same area parameter group. Step S304: Obtain the parameter groups of each same area in the selected detection log for the selected defect type. Randomly select the i-th defect attribute from the actual defect feature group of the same area parameter group to obtain the defect parameter of the i-th defect attribute in each same area parameter group. Obtain the illumination angle of each same area parameter group respectively, and sort the same area parameter groups according to the illumination angle from smallest to largest. Select the same area parameter group with the largest defect parameter corresponding to the i-th defect attribute as the high-quality parameter group, and divide all same area parameter groups into two same area parameter groups according to the high-quality parameter group. Step S305: Randomly select two adjacent parameter groups from any set of parameters in the same region, obtain the difference in illumination angle ΔT and the difference in defect parameters ΔD between the two adjacent parameter groups, and calculate the parameter change rate η = ΔD / ΔT between the two adjacent parameter groups; obtain the parameter change rate between any two adjacent parameter groups in the two sets of parameters in the same region, and calculate the average change rate η between the two sets of parameters in the same region. ave ; Step S306: Obtain the average change rate of the high-quality parameter group and the two sets of parameters in the same area corresponding to each defect attribute in each detection log; arbitrarily select an illumination angle and obtain the illumination angle difference ΔT between the high-quality parameter group of the i-th defect attribute and the selected illumination angle. ’ The defect parameter D is obtained from the high-quality parameter group of the i-th defect attribute. i The expected defect parameter of the i-th defect attribute at the selected illumination angle is calculated to be (D). i ) ex =D i -ΔT ’ ×(η i ) ave , where (η i ) ave Let be the average rate of change of the i-th defect attribute; obtain the expected defect parameters of each defect attribute at the selected illumination angle, and sum them to obtain the defect feature value D at the selected illumination angle. T The defect feature values ​​of each illumination angle are obtained, and the illumination angle with the largest defect feature value is selected as the best shooting angle for the defect type in the defect area.

2. The method for managing X-ray inspection image data of ceramic wine bottles according to claim 1, characterized in that: Step S100 includes the following steps: Step S101: Set a preset illumination angle for X-Ray relative to the ceramic bottle, pass X-Ray through the ceramic bottle at the preset illumination angle, and simultaneously start the flat panel detector to capture the image of X-Ray passing through the ceramic bottle to obtain the initial image data corresponding to the preset illumination angle. Step S102: Perform Gamma transformation, USM enhancement, and contrast adjustment on the initial image data in sequence to generate the corresponding standardized image; Step S103: Continuously adjust the X-Ray illumination angle to obtain standardized images at each illumination angle. Summarize the standardized images at all illumination angles to generate a detection log for one detection process.

3. The method for managing X-ray inspection image data of ceramic wine bottles according to claim 2, characterized in that: Step S200 includes the following steps: Step S201: Pre-build a defect database, which stores several defect types. For any defect type, there is a corresponding set of defect feature groups. The defect feature group consists of defect attributes and defect parameters. Summarize the defect attributes of all defect types to obtain a set of defect attributes. Step S202: Randomly select a detection log and randomly extract a standardized image from the selected detection log. Randomly segment the extracted standardized image to obtain several candidate regions. Randomly extract a candidate region and obtain the defect parameters of each defect attribute in the extracted candidate region. If a defect attribute has corresponding defect parameters, then the extracted candidate region is set as an abnormal region. Step S203: Obtain each abnormal region from the extracted standardized image. If there is a common boundary between two abnormal regions, compare the defect attributes of the two abnormal regions to obtain the number of identical defect attributes between the two abnormal regions as A. sim The correlation between the two anomalous regions was calculated as G=A. sim / Min(A1,A2), where Min() is the minimum value function, A1 is the number of defective attributes in one of the abnormal regions, and A2 is the number of defective attributes in another abnormal region; a preset correlation threshold G is used. th If G≥G th Then, the two abnormal regions are merged into a new defect region, and several defect regions are regenerated from the extracted standardized image. Step S204: Randomly select a defect area, and generate actual defect feature groups from the defect attributes of the selected defect area that contain defect parameters; randomly select a defect type from the defect database, and compare each actual defect feature group of the selected defect area with the defect feature group set of the selected defect type. If there is a defect feature group in the defect feature group set whose defect attribute is the same as the defect attribute of the actual defect feature group, and the defect parameter contains the defect parameter of the actual defect feature group, then set the actual defect feature group as a target defect feature group of the selected defect type. Step S205: Count the number of target defect feature groups for each defect type in the selected defect region, and set the number of target defect feature groups for the i-th defect type as b. i A preset threshold b for the number of feature groups is given. th If b i ≥b th Then, the i-th defect type is set as a defect type for the selected defect area, resulting in several defect types for the selected defect area.

4. The method for managing X-ray inspection image data of ceramic wine bottles according to claim 3, characterized in that: Step S400 includes the following steps: Step S401: Obtain the defect areas of the ceramic wine bottle in each inspection log, and count the number of defects occurring in each defect area in each inspection log. Let m be the number of defects occurring in the j-th defect area. j The occurrence frequency f=m of the j-th defect region is calculated. j / M, where M is the total number of detection logs; sort each defect area from high to low according to the frequency of occurrence, and generate the X-Ray imaging path; Step S402: The mobile X-Ray takes pictures of each area according to the generated shooting path to obtain real-time standardized images of each area. The defect parameters of each defect attribute in any real-time standardized image are obtained. If a defect attribute has corresponding defect parameters, the area where it is located is set as the target defect area. Step S403: Obtain the defect feature groups of each defect attribute in the target defect area, compare each defect feature group with the defect feature group set of various defect types in the defect database to obtain the defect type of the target defect area; if there is only one defect type in the target defect area, then set the best shooting angle of the unique defect type as the best shooting angle of the target defect area; if there are several defect types in the target defect area, then set the corresponding best shooting angles as the best shooting angles of the target defect area.

5. The method for managing X-ray inspection image data of ceramic wine bottles according to claim 4, characterized in that: Step S500 includes the following steps: Step S501: Arbitrarily select a target defect area. When the X-Ray image is taken of the selected target defect area from the optimal shooting angle of the previous target defect area, a first standardized image is generated. Set the optimal shooting angle of the previous target defect area as the initial shooting angle, and obtain the defect parameters corresponding to each defect attribute of the first standardized image at the initial shooting angle. Set the defect parameter corresponding to the i-th defect attribute as D1. i ; Step S502: Obtain the defect type of the selected target defect area and obtain the optimal shooting angle T for the selected target defect area. best The initial shooting angle is set to T0, and the change in shooting angle is calculated to be ΔT. best =T best -T0, setting the average rate of change of the i-th defect attribute as (η) i ) ave The expected defect parameter D2 of the i-th defect attribute under the optimal shooting angle is calculated. i The expected defect feature value (D) is obtained by summing the expected defect parameters of each defect attribute. T ) ex ; Step S503: Acquire the actual defect feature groups of the selected target defect area at the optimal shooting angle, obtain the defect parameters corresponding to each defect attribute, and sum them to obtain the actual defect feature value (D). T ) ac The defect detection deviation rate δ = |(D) of the selected target defect area was calculated. T ) ex -(D T ) ac | / (D T ) ex Preset a deviation rate threshold δ th If δ≥δ th If so, an abnormal alert will be issued regarding the defect identification status of the selected target defect area.

6. An X-Ray detection image data management system, used to execute the X-Ray detection image data management method for a ceramic wine bottle according to any one of claims 1-5, characterized in that: The management system includes an image generation and processing module, a defect identification and classification module, a shooting parameter matching module, an angle adjustment and optimization module, and an anomaly identification and judgment module. The image generation and processing module is used to acquire the initial image data generated by X-Ray during any detection process of the ceramic wine bottle, and to preprocess the initial image data to generate a standardized image. The standardized images generated during the detection process are summarized to generate a detection log. The defect identification and segmentation module is used to perform defect detection on each standardized image in any detection log to obtain the defect area of ​​the ceramic wine bottle in any detection log; and to analyze the defect situation in any defect area to determine the defect type in any defect area. The shooting parameter matching module is used to extract the shooting angle of X-Ray corresponding to any standardized image, and generate matching parameters for the defect situation based on the defect situation presented in the extracted standardized image; and analyze the difference of defect area in different standardized images of the same defect type to obtain the best shooting angle of any defect type in different defect areas. The angle adjustment and optimization module is used to obtain the X-Ray shooting path based on the occurrence frequency of each defect area when performing real-time defect detection on ceramic wine bottles; to identify the defect area based on the defect situation identified on the shooting path; and to set the optimal shooting angle for X-Ray according to the defect type in the defect area. The anomaly detection module is used to predict the expected detection status of any target defect area at the optimal shooting angle; analyze the actual defect status presented at the optimal shooting angle; and make anomaly detection judgments on the defect detection status.

7. The X-Ray detection image data management system according to claim 6, characterized in that: The defect identification and classification module includes an image defect identification unit and a defect type classification unit; The image defect recognition unit is used to perform defect detection on each standardized image in any detection log to obtain the defect area of ​​the ceramic wine bottle in any detection log; the defect type classification unit is used to analyze the defect situation in any defect area to determine the defect type in any defect area.

8. The X-Ray detection image data management system according to claim 6, characterized in that: The shooting parameter matching module includes a shooting angle extraction unit and a shooting angle matching unit; The shooting angle extraction unit is used to extract the shooting angle of the X-Ray corresponding to any standardized image, and generate matching parameters for the defect situation based on the defect situation presented in the extracted standardized image; the shooting angle matching unit is used to analyze the difference of defect area in different standardized images of the same defect type, and obtain the optimal shooting angle of any defect type in different defect areas.

9. An X-Ray detection image data management system according to claim 6, characterized in that: The angle adjustment and optimization module includes an actual defect detection unit and an optimal angle setting unit; The actual defect detection unit is used to obtain the X-Ray shooting path based on the occurrence frequency of each defect area when performing real-time defect detection on ceramic wine bottles; the optimal angle setting unit is used to identify defect areas based on the defect situation identified on the shooting path, and set the optimal shooting angle for X-Ray according to the defect type in the defect area.

Images