Bottle cap quality nondestructive testing system based on nondestructive testing technology

Abstract

The invention discloses a bottle cap quality nondestructive testing system based on a nondestructive testing technology, and particularly relates to the technical field of bottle cap quality analysis, which comprises a bottle cap quality estimation module, a detection resource distribution module, a bottle cap quality detection module, a bottle cap abnormity analysis module and an interactive feedback module, the abnormal quality risk index of the bottle cap is obtained by analyzing the light-transmitting image of the bottle cap, and the detection resources are allocated based on the quality risk index of the monitoring area, so that the monitoring resources of the bottle cap can be effectively allocated, the resource utilization efficiency is improved, and the labor cost is saved; a bottle cap quality detection module is used for carrying out quality detection on a bottle cap, obtaining an appearance abnormity index, a weight abnormity index and a sealing performance abnormity index of the bottle cap, and transmitting the obtained results to a bottle cap abnormity analysis module to obtain a comprehensive quality evaluation coefficient of a monitoring area; and the abnormal occurrence position of the bottle cap is positioned based on the comprehensive quality evaluation coefficient, so that managers are helped to troubleshoot quality abnormality reasons.

Description

Technical Field

[0001] This invention relates to the field of bottle cap quality analysis technology, and more specifically, to a non-destructive testing system for bottle cap quality based on non-destructive testing technology. Background Technology

[0002] As an important component of packaging containers, bottle caps directly affect the safety and quality of products. Non-destructive testing can detect defects and problems inside bottle caps, thereby preventing safety issues during transportation, storage, and use. Non-destructive testing of bottle cap quality can protect consumers' interests, prevent product safety issues caused by bottle cap quality problems, and improve consumers' trust and satisfaction with the products.

[0003] However, non-destructive testing of bottle cap quality also has some shortcomings:

[0004] Slower testing speed: Compared to manual testing, non-destructive testing (NDT) may be slower, requiring more time and equipment resources; Higher cost of NDT equipment: NDT requires advanced testing equipment and materials, which are expensive and may be unaffordable for some small businesses; Higher technical requirements: NDT requires professional technicians to operate and maintain it, which is technically demanding and makes personnel training and management more difficult.

[0005] In summary, the importance of non-destructive testing of bottle caps lies in ensuring product quality, improving production efficiency, reducing production costs, and protecting consumer interests. However, it also has some shortcomings, requiring companies to select and optimize it based on their own circumstances and needs in actual production. Summary of the Invention

[0006] To overcome the aforementioned deficiencies of the prior art, the present invention provides a non-destructive testing system for bottle cap quality based on non-destructive testing technology. By acquiring a light-transmitted image of the bottle cap and analyzing it, a quality risk index of bottle cap anomalies is obtained. Testing resources are allocated based on the quality risk index of the monitoring area to solve the problems mentioned in the background art.

[0007] Technical solution

[0008] To achieve the above objectives, the present invention provides the following technical solution: a non-destructive testing system for bottle cap quality based on non-destructive testing technology, comprising: a bottle cap quality prediction module, a testing resource allocation module, a bottle cap quality testing module, a bottle cap anomaly analysis module, and an interactive feedback module.

[0009] The bottle cap quality prediction module is used to divide the inspection object into several monitoring areas according to time and production line, acquire the light transmittance data of all bottle caps in each monitoring area, compress the light transmittance data of each bottle cap to several pixels, and the pixels of the monitoring area constitute the inspection image of the monitoring area. The inspection image of the monitoring area is analyzed, and the quality risk index of the monitoring area is estimated based on the pixel value fluctuation coefficient and the difference between the pixel value and the standard pixel value.

[0010] The detection resource allocation module allocates detection resources based on the quality risk index obtained by the bottle cap quality prediction module, so that the detection resources allocated to the monitoring area with the high quality risk index are higher than those allocated to the monitoring area with the low quality risk index.

[0011] The bottle cap quality inspection module is used to inspect the quality of bottle caps, obtain the appearance abnormality index, weight abnormality index and sealing performance abnormality index of the bottle caps, and transmit the obtained results to the bottle cap abnormality analysis module.

[0012] The bottle cap anomaly analysis module identifies and marks the anomaly monitoring areas based on the appearance anomaly index, weight anomaly index, and sealing performance anomaly index of the bottle caps in the monitoring area.

[0013] Interactive feedback module: Used to display the results of the bottle cap quality prediction module, the detection resource allocation module, the bottle cap quality detection module, and the bottle cap anomaly analysis module to managers. When the comprehensive quality assessment coefficient of the monitored area exceeds the threshold, it indicates that the production line and time point corresponding to the monitored area are abnormal, prompting managers to investigate the abnormality of the production line and time point.

[0014] Preferably, the bottle cap quality prediction module includes a region division unit, a light transmittance data acquisition unit, and a light transmittance data analysis unit. The region division unit is used to divide the monitored objects into several monitoring regions according to production time and production line, and assign them numbers. The light transmittance data acquisition unit is used to acquire light transmittance images of bottle caps in the monitoring regions. The bottle caps are passed sequentially through light transmittance image acquisition points, each including a light source and an image acquisition device. The light intensity and wavelength range of the light transmittance image acquisition points are set, the center point of the bottle cap is aligned with the light source, and the light transmittance images of the bottle caps are acquired through the image acquisition device. The acquired light transmittance images are then numbered according to their positions. The light transmittance data analysis unit obtains the quality risk index of the monitoring region by analyzing the light transmittance data of the bottle caps.

[0015] Preferably, the method by which the light transmittance data analysis unit obtains the quality risk index of the monitoring area is as follows:

[0016] Obtain the average pixel value of the monitored area. r represents the average pixel value of the i-th monitoring area. jG represents the red pixel value of the j-th bottle cap in the monitoring area. j b represents the green pixel value of the j-th bottle cap in the monitoring area. j The blue pixel value of the j-th bottle cap in the monitoring area is represented by n, where n represents the number of bottle caps in the monitoring area.

[0017] Obtain the pixel fluctuation coefficient of the monitored area: σ i Let x represent the light transmittance fluctuation coefficient of the i-th monitoring area. ij This represents the pixel value of the j-th bottle cap in the i-th monitoring area;

[0018] Based on the pixel fluctuation coefficient of the monitored area and the preset pixels, using the formula Obtain the quality risk index of the monitoring area, where Fzi represents the quality risk index of the i-th monitoring area, xb represents the preset pixel value of the bottle cap, ψ represents the intensity of the transmitted light source, γ represents the wavelength influence factor of the transmitted light source, and kz i This indicates the weighted influence factor of the monitoring area, and Where Jmi represents the number of bottle caps in the i-th monitoring area.

[0019] Preferably, the bottle cap quality inspection module includes an appearance inspection unit, a weight analysis unit, and a sealing performance inspection unit. The appearance inspection unit is used to obtain the appearance anomaly index of the bottle cap. Based on the translucent image and the natural light image of the bottle cap, it inputs them into the appearance anomaly recognition model and outputs the appearance anomaly index of the bottle cap. Based on the translucent image and the natural light image, it obtains the appearance feature vector group of the bottle cap. Based on a dual-channel deep learning model, it obtains the translucent feature anomaly index and the natural light feature anomaly index of the bottle cap, respectively. After joint analysis, it obtains the comprehensive appearance anomaly index of the bottle cap. The weight analysis unit is used to obtain the weight anomaly index of the bottle cap. The sealing performance inspection unit is used to obtain the sealing performance parameters of the bottle cap and obtains the sealing performance anomaly index of the bottle cap through stratified sampling.

[0020] Preferably, a dual-channel deep learning model is built to obtain the appearance anomaly index of the bottle cap, including the following steps:

[0021] Data labeling and partitioning: Appearance feature vectors are labeled in the translucent image and the natural light image respectively. The appearance feature vectors of the bottle cap number, the translucent image, the natural light image, and the label are taken as a sample. Each sample includes data from two channels. The first channel sample includes the appearance feature vectors of the bottle cap number, the translucent image, and the label. The second channel sample includes the appearance feature vectors of the bottle cap number, the natural light image, and the label. The labeled samples are divided into training set and test set according to the ratio.

[0022] Deep learning model training: Initialize the parameters of the deep learning model, set the loss function, input the training set into the deep learning model, adjust the parameters based on the training results of each training session, and train until the loss function is minimized to obtain the appearance anomaly recognition model; the loss function is cross-entropy loss, and the gradient descent optimization algorithm is used to update the weights of the deep learning model;

[0023] Validation of the appearance anomaly recognition model: The accuracy and precision of the appearance anomaly recognition model are validated using a test set.

[0024] Preferably, the method for obtaining the appearance anomaly index of the bottle cap is as follows: Obtain the light transmission feature vector set of the bottle cap, denote the i-th light transmission feature vector as wti, and calculate the light transmission feature anomaly index W1 of the bottle cap using the formula W1 = ∑wti * kti, where kti represents the weight coefficient corresponding to the i-th light transmission feature vector; obtain the natural light feature vector set of the bottle cap, denote the i-th natural light feature vector as wzi, and calculate the natural light feature anomaly index W2 of the bottle cap using the formula W2 = ∑wzi * kzi, where kzi represents the weight coefficient corresponding to the i-th natural light feature vector; and calculate the natural light feature anomaly index W2 of the bottle cap using the formula... The joint analysis yielded the bottle cap appearance anomaly index Wy, where f1 represents the weighting coefficient of the light transmission anomaly index, f2 represents the weighting coefficient of the natural light anomaly index, and 0. <f1<1,0<f2<1,f1+f2=1.0。

[0025] Preferably, the weight analysis unit is used to obtain the weight anomaly index of the bottle caps, obtain the number of bottle caps in the monitoring area, obtain the weight of the bottle caps in each monitoring area, and use a formula... The weight anomaly index Zy of the bottle cap was calculated, where zl 预 This represents the preset weight of a single bottle cap, m1 represents the number of bottle caps, ZL represents the total weight of the bottle caps, and cb represents the bottle cap deformation influence parameter, with a value of [0.5-1.0], which is set based on the size deformation of the bottle cap.

[0026] Preferably, the sealing performance detection unit is used to obtain the sealing performance abnormality index of the bottle cap, obtain the sealing performance parameters of the bottle cap through sealing performance testing, and set the preset sealing performance parameters Mf. 预 The sealing performance parameters obtained from the test are input into the sealing performance evaluation model to obtain the sealing performance anomaly index My of the bottle cap. The sealing performance evaluation model satisfies the formula... Where Mf represents the average value of the sealing performance parameters.

[0027] Preferably, the bottle cap anomaly analysis module identifies and marks anomaly monitoring areas based on the bottle cap quality of the monitoring area. It includes a comprehensive quality assessment unit and an anomaly location unit. The comprehensive quality assessment unit, based on the results obtained from the appearance inspection unit, weight analysis unit, and sealing performance inspection unit, performs a comprehensive analysis to obtain a comprehensive quality assessment coefficient ZP for the monitoring area. The formula for the comprehensive evaluation coefficient is as follows: Where w1 represents the weighting coefficient for appearance abnormalities, w2 represents the weighting coefficient for weight abnormalities, and w3 represents the weighting coefficient for sealing performance abnormalities, and w1, w2, and w3 are all greater than 0, and w1 + w2 + w3 = 1.0, ZP i The comprehensive quality assessment coefficient of the i-th monitoring area is represented by . The anomaly location unit compares the obtained comprehensive quality assessment coefficient with the preset comprehensive quality assessment coefficient of the monitoring area based on the comprehensive quality assessment coefficient of the monitoring area, and marks the monitoring area that exceeds the preset value as an abnormal monitoring area, thereby completing the anomaly location of the bottle cap quality.

[0028] The technical effects and advantages of this invention are as follows:

[0029] (1) This invention can quickly and accurately detect the quality of bottle caps through non-destructive testing, avoiding the cumbersome procedures and time waste of manual testing, and improving production efficiency; non-destructive testing can avoid product scrapping and rework, thereby reducing production costs and resource waste;

[0030] (2) This invention obtains the light transmission image of the bottle cap, analyzes it to obtain the quality risk index of the bottle cap abnormality, and allocates detection resources based on the quality risk index of the monitoring area. This can effectively allocate the monitoring resources of the bottle cap, improve resource utilization efficiency, and save manpower costs.

[0031] (3) The present invention uses a bottle cap quality detection module to perform quality detection on bottle caps, obtain the appearance abnormality index, weight abnormality index and sealing performance abnormality index of bottle caps, transmit the obtained results to the bottle cap abnormality analysis module to obtain the comprehensive quality assessment coefficient of the monitoring area, and locate the location of bottle cap abnormality based on the comprehensive quality assessment coefficient, so as to help managers investigate the cause of quality abnormality. Attached Figure Description

[0032] Figure 1 This is a block diagram of the overall structure of the system of the present invention.

[0033] Figure 2 This is a structural block diagram of the bottle cap quality estimation module of the present invention.

[0034] Figure 3 This is a structural block diagram of the bottle cap quality inspection module of the present invention. Detailed Implementation

[0035] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0036] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the scope of this application and its application or use.

[0037] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0038] Computer systems / servers can be described in the general context of computer system executable instructions (such as program modules) executed by the computer system. Typically, program modules can include routines, programs, object programs, components, logic, data structures, etc., which perform specific tasks or implement specific abstract data types. Computer systems / servers can be implemented in distributed cloud computing environments, where tasks are performed by remote processing devices linked through a communication network. In distributed cloud computing environments, program modules can reside on local or remote computing system storage media, including storage devices.

[0039] The embodiments of the present invention provide, as follows Figure 1 The system described is a non-destructive testing system for bottle cap quality, comprising: a bottle cap quality prediction module, a testing resource allocation module, a bottle cap quality testing module, a bottle cap anomaly analysis module, and an interactive feedback module. The bottle cap quality prediction module acquires the quality risk index of the monitoring area and transmits the prediction result to the testing resource allocation module. The testing resource allocation module receives the result from the bottle cap quality assessment module and allocates testing resources for the bottle caps. The bottle cap quality testing module tests the quality of the bottle caps based on the allocated testing resources and transmits the test result to the bottle cap anomaly analysis module. The bottle cap anomaly analysis module identifies and marks the anomaly monitoring areas based on the result acquired by the bottle cap quality testing module. The interactive feedback module displays the results of the bottle cap quality prediction module, testing resource allocation module, bottle cap quality testing module, and bottle cap anomaly analysis module to management personnel, allowing management personnel to proactively adjust the bottle cap testing resources.

[0040] The bottle cap quality prediction module is used to divide the inspection object into several monitoring areas according to time and production line, acquire the light transmittance data of all bottle caps in each monitoring area, compress the light transmittance data of each bottle cap to several pixels, and the pixels of the monitoring area constitute the inspection image of the monitoring area. The inspection image of the monitoring area is analyzed, and the quality risk index of the monitoring area is estimated based on the pixel value fluctuation coefficient and the difference between the pixel value and the standard pixel value.

[0041] This section explains how to capture images of bottle caps under natural light and under translucent conditions. A translucent image is one that allows light to pass through while retaining the image's tonal variations and details. For plastic bottle caps, the plastic material allows light to pass through while preserving the cap's details. The characteristic of a translucent image is its ability to display tonal variations and details while allowing light to pass through; compared to photographs of bottle caps under natural light, translucent images clearly show the internal details and structure of the bottle cap.

[0042] The detection resource allocation module allocates detection resources based on the quality risk index obtained by the bottle cap quality prediction module, so that the detection resources allocated to the monitoring area with the high quality risk index are higher than those allocated to the monitoring area with the low quality risk index.

[0043] This means that testing resources include, but are not limited to, the labor costs, time costs, equipment accuracy, and the number of samples to be tested. The more testing resources available, the better the testing results.

[0044] This demonstrates that by allocating testing resources, resources can be saved, managers can quickly locate the monitoring area where anomalies occur, and abnormal bottle caps can be rapidly investigated, thereby improving the efficiency of bottle cap quality testing.

[0045] The bottle cap quality inspection module is used to inspect the quality of bottle caps, obtain the appearance abnormality index, weight abnormality index and sealing performance abnormality index of the bottle caps, and transmit the obtained results to the bottle cap abnormality analysis module.

[0046] The bottle cap anomaly analysis module identifies and marks the anomaly monitoring areas based on the appearance anomaly index, weight anomaly index, and sealing performance anomaly index of the bottle caps in the monitoring area.

[0047] Interactive feedback module: Used to display the results of the bottle cap quality prediction module, the detection resource allocation module, the bottle cap quality detection module, and the bottle cap anomaly analysis module to managers. When the comprehensive quality assessment coefficient of the monitored area exceeds the threshold, it indicates that the production line and time point corresponding to the monitored area are abnormal, prompting managers to investigate the abnormality of the production line and time point.

[0048] The interactive feedback module allows managers to obtain information on the quality distribution of bottle caps, which helps them understand the quality of the caps and take corresponding measures. These measures include: preventing abnormal caps from entering the next production step based on the obtained quality data; and enabling managers to obtain production data for the abnormal monitoring area based on the corresponding production line and time point. Analyzing this data helps managers identify the cause of the abnormality and facilitates troubleshooting.

[0049] like Figure 2 As shown, the bottle cap quality prediction module includes a region division unit, a light transmittance data acquisition unit, and a light transmittance data analysis unit. The region division unit is used to divide the monitored objects into several monitoring regions according to production time and production line, and assign them numbers. The light transmittance data acquisition unit is used to acquire light transmittance images of bottle caps in the monitoring regions. The bottle caps are passed sequentially through light transmittance image acquisition points, each of which includes a light source and an image acquisition device. The light intensity and wavelength range of the light transmittance image acquisition points are set, and the center point of the bottle cap is aligned with the light source. The light transmittance images of the bottle caps are acquired through the image acquisition device, and the acquired light transmittance images are numbered according to their positions. The light transmittance data analysis unit obtains the quality risk index of the monitoring region by analyzing the light transmittance data of the bottle caps.

[0050] Furthermore, in actual testing, because the materials and thicknesses of bottle caps are different, it is necessary to adjust the wavelength and intensity of the light source to obtain a translucent image that can show the quality of the bottle cap.

[0051] Furthermore, when the obtained quality risk index is insufficient to distinguish the quality risk of the monitoring area, adjustments are made by compressing the light transmittance data to 1-m. 2 Each pixel improves analysis accuracy;

[0052] Furthermore, different quality defects have different light transmission requirements. For example, the light transmission requirements for detecting cracks and thickness of bottle caps are different. The wavelength and intensity of the light transmission source are set according to the actual situation to achieve different purposes.

[0053] In this embodiment of the invention, it should be explained that the method by which the light transmittance data analysis unit obtains the quality risk index of the monitored area is as follows:

[0054] Obtain the average pixel value of the monitored area. r represents the average pixel value of the i-th monitoring area. j G represents the red pixel value of the j-th bottle cap in the monitoring area. j b represents the green pixel value of the j-th bottle cap in the monitoring area. jThe blue pixel value of the j-th bottle cap in the monitoring area is represented by n, where n represents the number of bottle caps in the monitoring area.

[0055] Obtain the pixel fluctuation coefficient of the monitored area: σ i Let x represent the light transmittance fluctuation coefficient of the i-th monitoring area. ij This represents the pixel value of the j-th bottle cap in the i-th monitoring area;

[0056] Based on the pixel fluctuation coefficient of the monitored area and the preset pixels, using the formula Obtain the quality risk index of the monitoring area, where Fzi represents the quality risk index of the i-th monitoring area, xb represents the preset pixel value of the bottle cap, ψ represents the intensity of the transmitted light source, γ represents the wavelength influence factor of the transmitted light source, and kz i This indicates the weighted influence factor of the monitoring area, and Where Jmi represents the number of bottle caps in the i-th monitoring area.

[0057] like Figure 3 As shown, the bottle cap quality inspection module includes an appearance inspection unit, a weight analysis unit, and a sealing performance inspection unit. The appearance inspection unit is used to obtain the appearance anomaly index of the bottle cap. Based on the translucent image and the natural light image of the bottle cap, it inputs them into the appearance anomaly recognition model and outputs the appearance anomaly index of the bottle cap. Based on the translucent image and the natural light image, it obtains the appearance feature vector group of the bottle cap. Based on a dual-channel deep learning model, it obtains the translucent feature anomaly index and the natural light feature anomaly index of the bottle cap, respectively. After joint analysis, it obtains the comprehensive appearance anomaly index of the bottle cap. The weight analysis unit is used to obtain the weight anomaly index of the bottle cap. The sealing performance inspection unit is used to obtain the sealing performance parameters of the bottle cap. It obtains the sealing performance anomaly index of the bottle cap through stratified sampling.

[0058] In this embodiment of the invention, it should be explained that building a dual-channel deep learning model to obtain the appearance anomaly index of the bottle cap includes the following steps:

[0059] Data labeling and partitioning: Appearance feature vectors are labeled in the translucent image and the natural light image respectively. The appearance feature vectors of the bottle cap number, the translucent image, the natural light image, and the label are taken as a sample. Each sample includes data from two channels. The first channel sample includes the appearance feature vectors of the bottle cap number, the translucent image, and the label. The second channel sample includes the appearance feature vectors of the bottle cap number, the natural light image, and the label. The labeled samples are divided into training set and test set according to the ratio.

[0060] Deep learning model training: Initialize the parameters of the deep learning model, set the loss function, input the training set into the deep learning model, adjust the parameters based on the training results of each training session, and train until the loss function is minimized to obtain the appearance anomaly recognition model; the loss function is cross-entropy loss, and the gradient descent optimization algorithm is used to update the weights of the deep learning model;

[0061] Validation of the appearance anomaly recognition model: The accuracy and precision of the appearance anomaly recognition model are validated using a test set.

[0062] In this embodiment of the invention, it should be explained that the method for obtaining the appearance anomaly index of the bottle cap is as follows: The light transmission feature vector group of the bottle cap is obtained, and the i-th light transmission feature vector of the bottle cap is denoted as wti. The light transmission feature anomaly index W1 of the bottle cap is calculated using the formula W1 = ∑wti * kti, where kti represents the weight coefficient corresponding to the i-th light transmission feature vector; The natural light feature vector group of the bottle cap is obtained, and the i-th natural light feature vector of the bottle cap is denoted as wzi. The natural light feature anomaly index W2 of the bottle cap is calculated using the formula W2 = ∑wzi * kzi, where kzi represents the weight coefficient corresponding to the i-th natural light feature vector; The formula... The joint analysis yielded the bottle cap appearance anomaly index Wy, where f1 represents the weighting coefficient of the light transmission anomaly index, f2 represents the weighting coefficient of the natural light anomaly index, and 0. <f1<1,0<f2<1,f1+f2=1.0。

[0063] In this embodiment of the invention, it should be explained that the weight analysis unit is used to obtain the weight anomaly index of the bottle caps, obtain the number of bottle caps in the monitoring area, and obtain the weight of the bottle caps in each monitoring area, using the formula... The weight anomaly index Zy of the bottle cap was calculated, where zl 预 This represents the preset weight of a single bottle cap, m1 represents the number of bottle caps, ZL represents the total weight of the bottle caps, and cb represents the bottle cap deformation influence parameter, with a value of [0.5-1.0], which is set based on the size deformation of the bottle cap.

[0064] In this embodiment of the invention, it should be explained that the sealing performance detection unit is used to obtain the sealing performance abnormality index of the bottle cap, obtain the sealing performance parameters of the bottle cap through sealing performance testing, and set the preset sealing performance parameters Mf. 预 The sealing performance parameters obtained from the test are input into the sealing performance evaluation model to obtain the sealing performance anomaly index My of the bottle cap. The sealing performance evaluation model satisfies the formula... in This represents the average value of the sealing performance parameters.

[0065] In this embodiment of the invention, it should be explained that the sealing performance parameters are obtained as follows: the sealed bottle is placed in the sealed container, the audio probe is placed close to the bottle opening, an audio signal is emitted and the reflected signal is received, and the sealing performance parameters of the bottle cap are determined based on the degree of signal attenuation. This includes the following steps:

[0066] Step S01: Obtain the object for sealing performance testing, rotate and twist the bottle cap and the corresponding bottle body to the preset state to obtain a sealed bottle, and put the sealed bottle into a sealed container;

[0067] Step S02: Place the audio probe close to the bottle opening to ensure good contact, while avoiding excessive force that could deform or break the bottle. Turn on the power of the plastic bottle sealing tester, adjust the parameters of the tester, and select the appropriate audio signal frequency and intensity, as well as the appropriate test time.

[0068] Step S03: Analyze the collected audio signal to obtain the sealing performance of the bottle cap. After the test is completed, remove the plastic bottle.

[0069] In this embodiment of the invention, it should be explained that when performing the audio detection method, it is necessary to select an appropriate audio signal frequency and intensity, as well as an appropriate test time, to ensure the accuracy and reliability of the detection results; at the same time, it is also necessary to keep the bottle, audio probe and tester clean and dry to avoid affecting the detection results.

[0070] Based on the test results, the sealing performance parameters of the bottle cap were obtained, and then calculated using the formula... Obtain the sealing performance parameters Mf and yp of the bottle cap. a This indicates at time point t a The audio signal, yp b Indicates a point in time tb The audio signal, where η represents the intensity of the audio signal and n represents the number of bottle caps tested.

[0071] In this embodiment of the invention, it should be explained that the bottle cap anomaly analysis module, based on the bottle cap quality of the monitoring area, obtains and marks the anomaly monitoring area, including a comprehensive quality assessment unit and an anomaly location unit. The comprehensive quality assessment unit, based on the results obtained from the appearance inspection unit, weight analysis unit, and sealing inspection unit, performs comprehensive analysis to obtain the comprehensive quality assessment coefficient ZP of the monitoring area. The formula for the comprehensive evaluation coefficient is as follows: Where w1 represents the weighting coefficient for appearance abnormalities, w2 represents the weighting coefficient for weight abnormalities, and w3 represents the weighting coefficient for sealing performance abnormalities, and w1, w2, and w3 are all greater than 0, and w1 + w2 + w3 = 1.0, ZP iThe comprehensive quality assessment coefficient of the i-th monitoring area is represented; the anomaly location unit obtains the anomaly monitoring area based on the comprehensive quality assessment coefficient of the monitoring area, and completes the anomaly location of the bottle cap quality.

[0072] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A non-destructive testing system for bottle cap quality based on non-destructive testing technology, characterized in that, include: The bottle cap quality prediction module is used to divide the inspection object into several monitoring areas according to time and production line, acquire the light transmittance data of all bottle caps in each monitoring area, compress the light transmittance data of each bottle cap to several pixels, and the pixels of the monitoring area constitute the inspection image of the monitoring area. The inspection image of the monitoring area is analyzed, and the quality risk index of the monitoring area is estimated based on the pixel value fluctuation coefficient and the difference between the pixel value and the standard pixel value. The detection resource allocation module allocates detection resources based on the quality risk index obtained by the bottle cap quality prediction module, so that the detection resources allocated to the monitoring area with the high quality risk index are higher than those allocated to the monitoring area with the low quality risk index. The bottle cap quality inspection module is used to inspect the quality of bottle caps, obtain the appearance abnormality index, weight abnormality index and sealing performance abnormality index of the bottle caps, and transmit the obtained results to the bottle cap abnormality analysis module. The bottle cap anomaly analysis module identifies and marks the anomaly monitoring areas based on the appearance anomaly index, weight anomaly index, and sealing performance anomaly index of the bottle caps in the monitoring area. Interactive feedback module: Used to display the results of the bottle cap quality prediction module, the detection resource allocation module, the bottle cap quality detection module, and the bottle cap anomaly analysis module to managers. When the comprehensive quality assessment coefficient of the monitored area exceeds the threshold, it indicates that the production line and time point corresponding to the monitored area are abnormal, prompting managers to investigate the abnormality of the production line and time point.

2. The bottle cap quality non-destructive testing system based on non-destructive testing technology according to claim 1, characterized in that, The bottle cap quality prediction module includes a region division unit, a light transmittance data acquisition unit, and a light transmittance data analysis unit. The region division unit divides the monitored objects into several monitoring regions according to production time and production line, and assigns them numbers. The light transmittance data acquisition unit acquires light transmittance images of the bottle caps in the monitoring regions. The bottle caps are passed sequentially through light transmittance image acquisition points, each including a light source and an image acquisition device. The light intensity and wavelength range of the light transmittance image acquisition points are set, and the center point of the bottle cap is aligned with the light source. The image acquisition device acquires the light transmittance images of the bottle caps and assigns numbers according to their positions. The light transmittance data analysis unit analyzes the light transmittance data of the bottle caps to obtain the quality risk index of the monitoring region.

3. The bottle cap quality non-destructive testing system based on non-destructive testing technology according to claim 2, characterized in that, The method by which the light transmittance data analysis unit obtains the quality risk index of the monitored area is as follows: Obtain the average pixel value of the monitored area r represents the average pixel value of the i-th monitoring area. j G represents the red pixel value of the j-th bottle cap in the monitoring area. j b represents the green pixel value of the j-th bottle cap in the monitoring area. j The blue pixel value of the j-th bottle cap in the monitoring area is represented by n, where n represents the number of bottle caps in the monitoring area. Obtain the pixel fluctuation coefficient of the monitoring area σ i Let x represent the light transmittance fluctuation coefficient of the i-th monitoring area. ij This represents the pixel value of the j-th bottle cap in the i-th monitoring area; Based on the pixel fluctuation coefficient of the monitored area and the preset pixels, using the formula Obtain the quality risk index of the monitoring area, where Fzi represents the quality risk index of the i-th monitoring area, xb represents the preset pixel value of the bottle cap, ψ represents the intensity of the transmitted light source, γ represents the wavelength influence factor of the transmitted light source, and kz i This indicates the weighted influence factor of the monitored area, and Where Jmi represents the number of bottle caps in the i-th monitoring area.

4. The bottle cap quality non-destructive testing system based on non-destructive testing technology according to claim 3, characterized in that, The bottle cap quality inspection module includes an appearance inspection unit, a weight analysis unit, and a sealing performance inspection unit. The appearance inspection unit is used to obtain the appearance anomaly index of the bottle cap. Based on the translucent image and the natural light image of the bottle cap, it inputs them into the appearance anomaly recognition model and outputs the appearance anomaly index of the bottle cap. It obtains the appearance feature vector group of the bottle cap based on the translucent image and the natural light image, and obtains the translucent feature anomaly index and the natural light feature anomaly index of the bottle cap based on a dual-channel deep learning model. After joint analysis, a comprehensive appearance anomaly index of the bottle cap is obtained. The weight analysis unit is used to obtain the weight anomaly index of the bottle cap. The sealing performance inspection unit is used to obtain the sealing performance parameters of the bottle cap and obtains the sealing performance anomaly index of the bottle cap through stratified sampling.

5. A bottle cap quality non-destructive testing system based on non-destructive testing technology according to claim 4, characterized in that, To build a dual-channel deep learning model and obtain the appearance anomaly index of bottle caps, the following steps are included: Data labeling and partitioning: Appearance feature vectors are labeled in the translucent image and the natural light image respectively. The appearance feature vectors of the bottle cap number, the translucent image, the natural light image, and the label are taken as a sample. Each sample includes data from two channels. The first channel sample includes the appearance feature vectors of the bottle cap number, the translucent image, and the label. The second channel sample includes the appearance feature vectors of the bottle cap number, the natural light image, and the label. The labeled samples are divided into training set and test set according to the ratio. Deep learning model training: Initialize the parameters of the deep learning model, set the loss function, input the training set into the deep learning model, adjust the parameters based on the training results of each training session, train until the loss function is minimized, and obtain the appearance anomaly recognition model; Validation of the appearance anomaly recognition model: The accuracy and precision of the appearance anomaly recognition model are validated using a test set.

6. The bottle cap quality non-destructive testing system based on non-destructive testing technology according to claim 5, characterized in that, The method for obtaining the appearance anomaly index of the bottle cap is as follows: Obtain the light transmission feature vector set of the bottle cap, denote the i-th light transmission feature vector as wti, and calculate the light transmission feature anomaly index W1 of the bottle cap using the formula W1 = ∑wti * kti, where kti represents the weight coefficient corresponding to the i-th light transmission feature vector; Obtain the natural light feature vector set of the bottle cap, denote the i-th natural light feature vector as wzi, and calculate the natural light feature anomaly index W2 of the bottle cap using the formula W2 = ∑wzi * kzi, where kzi represents the weight coefficient corresponding to the i-th natural light feature vector; [Further details about the formula would follow here.] The joint analysis yielded the bottle cap appearance anomaly index Wy, where f1 represents the weighting coefficient of the light transmission anomaly index, f2 represents the weighting coefficient of the natural light anomaly index, and 0. <f1<1,0<f2<1,f1+f2=1.0。 7. A non-destructive testing system for bottle cap quality based on non-destructive testing technology according to claim 6, characterized in that, The weight analysis unit is used to obtain the weight anomaly index of the bottle caps, the number of bottle caps in the monitoring area, and the weight of the bottle caps in each monitoring area, using a formula... The weight anomaly index Zy of the bottle cap was calculated, where zl 预 This represents the preset weight of a single bottle cap, m1 represents the number of bottle caps, ZL represents the total weight of the bottle caps, and cb represents the bottle cap deformation influence parameter, with a value of [0.5-1.0], which is set based on the size deformation of the bottle cap.

8. A non-destructive testing system for bottle cap quality based on non-destructive testing technology according to claim 7, characterized in that, The sealing performance detection unit is used to obtain the sealing performance anomaly index of the bottle cap. It acquires the sealing performance parameters of the bottle cap through sealing performance testing, and inputs the preset sealing performance parameters and the obtained sealing performance parameters into the sealing performance evaluation model to obtain the sealing performance anomaly index My of the bottle cap. The sealing performance evaluation model satisfies the formula... in Mf represents the average value of the sealing performance parameters. 预 This indicates the preset sealing performance parameters of the bottle cap.

9. A non-destructive testing system for bottle cap quality based on non-destructive testing technology according to claim 8, characterized in that, The bottle cap anomaly analysis module obtains and marks the anomaly monitoring area based on the bottle cap quality of the monitoring area. It includes a comprehensive quality assessment unit and an anomaly location unit. The comprehensive quality assessment unit obtains the comprehensive quality assessment coefficient ZP of the monitoring area by comprehensively analyzing the results obtained by the appearance inspection unit, weight analysis unit and sealing inspection unit. The anomaly location unit obtains the anomaly monitoring area based on the comprehensive quality evaluation coefficient of the monitoring area, and completes the anomaly location of the bottle cap quality.

Images