Filled product analysis and quality inspection method and system, readable storage medium and device

Abstract

The invention provides a filling product analysis and quality inspection method based on a millimeter wave radar and a deep learning model, and aims to improve the detection precision of tail folding and cover locking of a tank body. Detecting the tail folding of the tank body through a millimeter wave radar, segmenting the data based on a loss value, calculating the voidage of the tail folding of the tank body, and marking the tank body of which the tail folding is not completely packaged; an infrared thermal imaging technology is used for obtaining a lock cover image, the integrity of a lock cover is analyzed through a deep learning model, whether a tank body is packaged perfectly or not is judged, and filling products with incomplete lock cover packaging are marked. In addition, the invention also provides a quality inspection system, a readable storage medium and a device so as to realize automation and intelligence of the method.

Description

Technical Field

[0001] The present invention belongs to the technical field of production control, and relates to a method, system, readable storage medium and device for analyzing quality inspection of filled products. Background Art

[0002] When filling products are mass-produced, the tail of the can body must first be folded, that is, the tail of the can body must be sealed. After the folding is completed, the filling process is carried out. After the filling is completed, the cover body and the head of the can body are locked. The locking operation is the sealing operation after the can body is filled.

[0003] The traditional water injection test method is time-consuming and costly when detecting the folded tail of the can body. When detecting the integrity of the locking operation, it is difficult to accurately identify the state of the locking operation, resulting in an inability to effectively judge the sealing between the cover and the can body. When marking incompletely packaged products, it is difficult to make real-time and accurate judgments. Summary of the invention

[0004] In order to solve the problems existing in the background technology, the technical solution adopted by the present invention is as follows:

[0005] A filling product analysis and quality inspection method, comprising:

[0006] Use millimeter wave radar to detect the folded tail of the tank to be filled, and segment the millimeter wave radar detection data based on the loss value;

[0007] After the filling of the filled product is completed, the can body is capped, and the infrared thermal image of the can body with the cap locked is obtained. The integrity of the corresponding cap locking operation in the infrared thermal image of the can body is analyzed through the deep learning model;

[0008] The segmented millimeter-wave radar detection data and the integrity of the capping operation are integrated and analyzed to obtain the can closure degree of the current filled product;

[0009] Based on the sealing degree of the tank body, determine whether the tank body is sealed intact, and mark the tank body that is not completely sealed.

[0010] Furthermore, the specific method of using the millimeter wave radar to detect the folded tail of the tank to be filled and segmenting the millimeter wave radar detection data based on the loss value is:

[0011] Obtain the echo data of the millimeter-wave radar, including distance information, and record it in a time series manner;

[0012] Calculate the loss value of the radar signal echo. The specific formula is:

[0013]

[0014] Where r is the distance from the radar to the tank body, r0 is the reference value of the distance from the radar to the tank body, and n is the radar wave transmission path loss index;

[0015] A sliding window is created in the echo data of the millimeter-wave radar, the echo data of the millimeter-wave radar is separated into multiple windows by using the sliding window, and an average value of the loss value is calculated in each window;

[0016] A clustering algorithm is used to classify the loss values, and the echo data of each type of millimeter-wave radar is labeled with a loss value to complete the segmentation of the millimeter-wave radar detection data.

[0017] Furthermore, the specific method of obtaining the infrared thermal image of the tank body after the lid is locked and analyzing the integrity of the lid locking operation corresponding to the infrared thermal image of the tank body through the deep learning model is:

[0018] Obtain the infrared thermal image of the tank body after the lock cover is completed, extract the position of the lock cover in the infrared thermal image of the tank body through Harris corner detection, and extract the feature set around the lock cover;

[0019] Calculate the information gain of each feature in the feature set around the lock cover. The specific formula is:

[0020]

[0021] Among them, H(D) is the entropy of the feature set around the lock cover, H(D v ) is the conditional entropy based on feature A, which reduces the influence of unimportant features through penalty terms and selects the top N features, where N is a positive integer;

[0022] Based on the first N features, the deep learning model is used to judge the integrity of the locking operation corresponding to all tank infrared thermal images.

[0023] Furthermore, the specific method of judging the integrity of the locking operation corresponding to all tank infrared thermal images through the deep learning model is:

[0024] Arrange each of the first N features according to the size of the specific value to generate a set of N features;

[0025] Randomly select a feature set M, mark the last feature value and the first feature value in M, randomly select a value between the last feature value and the first feature value to split the feature set M, and mark the split value;

[0026] According to the above method, iterate to the position of the median of the feature set M until all features in the feature set M are labeled, select the average of all segmentation values, set the discrete threshold, and obtain the corresponding discrete values ​​for all feature values ​​in the feature set M;

[0027] The discrete value of each eigenvalue Mi in the feature set M is used as the numerator, and the difference between the maximum eigenvalue and the minimum eigenvalue is used as the denominator, whose value is recorded as Zi, and the integrity value of the lock cover operation is set to 1-Zi.

[0028] Furthermore, the specific method of integrating and analyzing the segmented millimeter-wave radar detection data and the integrity of the cap locking operation to obtain the can closure degree of the current filled product is:

[0029] Calculate the gradient change threshold of each segment of millimeter-wave radar exploration data, multiply each segment of millimeter-wave radar detection data by the corresponding loss value ratio, and obtain the void rate of the tank tail;

[0030] Set the tank body folding tail void rate threshold and the cover locking operation integrity threshold;

[0031] The void ratio of the folded tail of the tank body and the integrity of the locking cover operation are calculated. If both conditions are met, it indicates that the current tank body is intact.

[0032] On the other hand, the present invention provides a filling product analysis and quality inspection system, comprising:

[0033] The data acquisition module acquires the detection data of the tail fold of the tank to be filled by the millimeter-wave radar, and acquires the infrared thermal image of the tank after the cap is locked;

[0034] The data processing module segments the millimeter-wave radar detection data based on the loss value, analyzes the integrity of the can body infrared thermal image corresponding to the locking operation, and also calculates the can body sealing degree of the filled product;

[0035] The control module marks the tanks that are not fully packaged.

[0036] In another aspect, the present invention provides a computer-readable storage medium storing the above system.

[0037] On the other hand, the present invention provides a filling product analysis and quality inspection device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the above-mentioned filling product analysis and quality inspection method through the computer program.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] By using millimeter-wave radar technology to accurately detect the folded tail of the tank and segmenting the detection data based on the loss value, this method can more accurately identify changes in the internal structure of the tank and improve the accuracy of detection.

[0040] Combining infrared thermal imaging technology and deep learning models to analyze the integrity of the tank cover operation can not only quickly determine whether the cover operation is intact, but also reduce the need for manual inspection and improve the efficiency of the entire inspection process.

[0041] The present invention not only takes into account the distortion of the detection data at the folded tail of the tank body, but also combines the integrity of the locking cover operation. By analyzing these two aspects of data, the sealing performance of the tank body can be comprehensively evaluated to ensure the sealing and safety of the product.

[0042] Based on the judgment standard of can closure degree, incompletely packaged products can be effectively identified and marked to prevent unqualified products from entering the market. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 is a flow chart of the method of the present invention;

[0044] Figure 2 It is a system structure diagram of the present invention. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] like Figure 1 As shown, a filling product analysis and quality inspection method includes:

[0047] In the process of producing filled products, millimeter wave radar is used to detect the folding of the can to be filled. Through a specific algorithm, the data detected by the millimeter wave radar is segmented according to the loss value to analyze the data characteristics more accurately. The folding is the encapsulation operation of the can tail. After the segmentation process is completed, the void rate of the folded tail of the can is calculated based on the segmentation results, and the cans with incomplete folding and encapsulation are marked.

[0048] After the can body with the tail folded and fully packaged is filled, the can body is immediately locked. The locking operation is a packaging operation after the can body is filled.

[0049] At this time, infrared thermal imaging technology is used to obtain an infrared thermal image of the tank body with the cap locked, and the image is deeply analyzed with the help of a deep learning model to determine the integrity of the locking operation, such as whether the lock is tight, whether there is any deformation or gap, etc.

[0050] Afterwards, based on the integrity of the locking operation, the filled products with incomplete locking caps will be marked. If the locking cap operation closure does not meet the set standard, the can body is judged to be incompletely sealed, and the corresponding filled products will be marked to ensure the product packaging quality and guarantee the safety and stability of the product during storage, transportation and sales.

[0051] Before the filling process, the millimeter wave radar is used to detect the folded tail of the can to be filled. After the millimeter wave radar collects the data, the radar data is segmented.

[0052] Get the echo data of the millimeter wave radar, including information such as distance, and record it in a time series. This information may be collected through multiple pulse echoes and recorded as a time series. Generally speaking, the radar will convert the echo signal into a distance map when receiving the echo.

[0053] The loss value can be defined in several ways, usually based on the amplitude attenuation of the echo signal, the decrease in signal-to-noise ratio, or the error. The attenuation of the radar signal is usually inversely proportional to the square of the propagation distance. The loss value can be calculated using the following formula:

[0054]

[0055] Where r is the distance from the radar to the tank folded tail, r0 is the reference value of the distance from the radar to the tank folded tail, and n is the radar wave transmission path loss index. An increase in attenuation amplitude usually means a decrease in the quality of radar detection.

[0056] A sliding window is constructed based on the echo data of the millimeter-wave radar, and the echo data is separated into several independent windows with the help of this sliding window. The average value of the loss value is accurately calculated within the range of each window. Advanced clustering algorithms are used to classify the loss values, and corresponding loss value labels are assigned to each type of millimeter-wave radar echo data based on the classification results, so as to achieve effective segmentation of millimeter-wave radar detection data.

[0057] Calculate the gradient change threshold of each segment of millimeter wave radar exploration data. The gradient change of radar data can reflect the difference of objects detected by radar. For example, when radar detects material defects, the waves returned by radar have obvious density changes.

[0058] Multiply each segment of millimeter wave radar detection data by the corresponding loss value ratio, where the loss ratio is the ratio of the loss wave of radar data to the entire radar wave data. The void rate of the tank tail folding can be obtained through the above calculation.

[0059] Set the fold tail hole rate threshold.

[0060] The void rate of the folded tail of the can is calculated. If the folded tail void rate threshold condition is met, it indicates that the current can folded tail packaging is intact. Cans that do not meet the folded tail void rate threshold are marked to avoid entering the subsequent production process.

[0061] After the can body is completely filled with the folded tail sealed, the can body is immediately locked. The infrared thermal imaging technology is used to obtain the infrared thermal image of the can body after the cap is locked, and the image is deeply analyzed with the help of a deep learning model to determine the integrity of the locking operation, such as whether the lock is tight, whether there is deformation or gaps, etc.

[0062] Obtain its infrared thermal image. Use Harris corner detection technology to accurately extract the position information of the cover in the infrared thermal image of the tank, and then extract the feature set around the cover.

[0063] Calculate the information gain of each feature in the feature set around the cover. The specific formula is:

[0064]

[0065] Among them, H(D) is the entropy of the feature set around the cover, H(D v ) is the conditional entropy based on feature A, which reduces the influence of unimportant features through penalty terms.

[0066] It involves the entropy of the feature set and the conditional entropy based on feature A, and introduces a penalty term, the purpose of which is to effectively reduce the interference and influence of unimportant features on the results. Finally, the top N features are selected based on the calculation results, where N is a positive integer, in order to provide accurate and critical data support for the subsequent integrity analysis of the locking cover operation. N can be selected through repeated experiments.

[0067] Based on the top N most influential features, the deep learning model is used to judge the integrity of the locking operation corresponding to the infrared thermal images of all tanks.

[0068] Arrange the first N features in order according to the size of their specific values ​​to generate a set of N features.

[0069] Randomly select a feature set M from it, and clearly mark the last feature value and the first feature value in the set. Then, randomly select values ​​within the interval between the last feature value and the first feature value, use these values ​​to reasonably split the feature set, and mark the split values.

[0070] According to the above method, iterate to the position of the median of the feature set M. If the number of elements in the feature set M is an odd number, do not process the median. After all features in the feature set M are labeled, select the average of all segmentation values, set the discrete threshold, and obtain the corresponding discrete values ​​for all feature values ​​in the feature set M;

[0071] The discrete value of each eigenvalue Mi in the feature set M is used as the numerator, the difference between the maximum eigenvalue and the minimum eigenvalue is used as the denominator, and its value is recorded as Zi, and the value of the integrity is set to 1-Zi.

[0072] Set the integrity threshold for the cover lock operation.

[0073] The integrity of the locking cover operation is calculated, and if the integrity of the locking cover operation meets the conditions, it indicates that the current tank is intact, thereby completing the comprehensive evaluation and judgment of the tank sealing degree.

[0074] If the sealing degree of the can body does not meet the set standard, the can body is judged to be incompletely sealed and the corresponding filled product will be marked to ensure the quality of product packaging and guarantee the safety and stability of the product during storage, transportation and sales.

[0075] On the other hand, Figure 2 As shown, the present invention provides a filling product analysis and quality inspection system, which mainly includes three key modules:

[0076] The core function of the data acquisition module is to accurately obtain the data information detected by the millimeter-wave radar on the folded tail of the can to be filled, and at the same time, efficiently obtain the infrared thermal image of the can after the cap is locked;

[0077] The data processing module can segment the millimeter-wave radar detection data in detail according to the loss value, calculate the void ratio of the folded tail of the tank, deeply analyze the integrity of the corresponding lock cover in the infrared thermal image of the tank, and accurately calculate the sealing degree of the tank of the filled product;

[0078] The control module has an extremely important task, which is to mark those cans that are not completely sealed, so as to ensure that the quality and quality of the entire filling product production line meet high standards. Incompletely sealed cans include cans with incomplete folding and incomplete locking.

[0079] On the other hand, the present invention also provides a computer-readable storage medium, wherein the storage medium stores the above-mentioned filling product analysis and quality inspection system, wherein the above-mentioned filling product analysis and quality inspection method is executed when the filling product analysis and quality inspection system is running. Computer-readable media include permanent and non-permanent, removable and non-removable media, and the storage of the filling product analysis and quality inspection system can be realized by any method or technology.

[0080] On the other hand, the present invention provides a filling product analysis quality inspection device, comprising a memory, a processor, and a filling product analysis quality inspection system stored in the memory and operable on the processor, wherein the processor executes a filling product analysis quality inspection method through the filling product analysis quality inspection system. The filling product analysis quality inspection method provided by the present invention must be operated with the help of the filling product analysis quality inspection system as a software platform, and the filling product analysis quality inspection system must rely on the hardware platform of the filling product analysis quality inspection device provided by the present invention to operate.

[0081] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for analyzing and inspecting the quality of a filled product, characterized in that: Included are: Use millimeter-wave radar to detect the folded tail of the can to be filled, segment the millimeter-wave radar detection data based on the loss value, calculate the void rate of the folded tail of the can based on the segmentation result, and mark the can with incomplete folded tail packaging; After the can body with the tail folded and sealed is filled, the can body is capped, and the infrared thermal image of the can body with the cap locked is obtained. The integrity of the corresponding cap locking operation in the infrared thermal image of the can body is analyzed through the deep learning model; Based on the integrity of the capping operation, cans with incomplete capping will be marked.

2. A method for analyzing and inspecting the quality of a filled product according to claim 1, characterized in that: The specific method of using the millimeter wave radar to detect the folded tail of the tank to be filled and segmenting the millimeter wave radar detection data based on the loss value is as follows: Obtain the echo data of the millimeter-wave radar, including distance information, and record it in a time series manner; Calculate the loss value of the radar signal echo. The specific formula is: Where r is the distance from the radar to the tank body, r0 is the reference value of the distance from the radar to the tank body, and n is the radar wave transmission path loss index; A sliding window is created in the echo data of the millimeter-wave radar, the echo data of the millimeter-wave radar is separated into multiple windows by using the sliding window, and an average value of the loss value is calculated in each window; A clustering algorithm is used to classify the loss values, and the echo data of each type of millimeter-wave radar is labeled with a loss value to complete the segmentation of the millimeter-wave radar detection data.

3. A method for analyzing and inspecting the quality of a filled product according to claim 2, characterized in that: The specific method for calculating the void ratio of the tank tail fold based on the segmentation results is: Calculate the gradient change threshold of each segment of millimeter-wave radar exploration data, multiply each segment of millimeter-wave radar detection data by the corresponding loss value ratio, and obtain the void rate of the tank tail; Set the fold tail void rate threshold; The void ratio of the folded tail of the tank body is calculated. If the folded tail void ratio threshold condition is met, it indicates that the current folded tail packaging of the tank body is intact.

4. A method for analyzing and inspecting the quality of a filled product according to claim 1, characterized in that: The specific method of obtaining the infrared thermal image of the tank body after the lid is locked and analyzing the integrity of the corresponding lid locking operation in the infrared thermal image of the tank body through the deep learning model is: Obtain the infrared thermal image of the can body after the cover is locked, extract the position of the cover in the infrared thermal image of the can body by Harris corner detection, and extract the feature set around the cover; Calculate the information gain of each feature in the feature set around the cover. The specific formula is: Among them, H(D) is the entropy of the feature set around the cover, H(D v ) is the conditional entropy based on feature A, which reduces unimportant features through penalty terms and selects the top N features, where N is a positive integer; Based on the first N features, the deep learning model is used to judge the integrity of the locking operation corresponding to the infrared thermal images of all tanks.

5. A method for analyzing and inspecting the quality of a filled product according to claim 4, characterized in that: The specific method of judging the integrity of the locking operation corresponding to the infrared thermal images of all tanks through the deep learning model is: Arrange each of the first N features according to the size of the specific value to generate a set of N features; Randomly select a feature set M, mark the last feature value and the first feature value in M, randomly select a value between the last feature value and the first feature value to split the feature set M, and mark the split value; According to the above method, iterate to the position of the median of the feature set M. If the number of elements in the feature set M is an odd number, do not process the median until all features in the feature set M are labeled. Select the average of all segmentation values, set the discrete threshold, and obtain the corresponding discrete values ​​for all feature values ​​in the feature set M. The discrete value of each eigenvalue Mi in the feature set M is used as the numerator, and the difference between the maximum eigenvalue and the minimum eigenvalue is used as the denominator, and its value is recorded as Zi, and the integrity value of the lock cover operation is set to 1-Zi; Set the integrity threshold for the cover lock operation; The integrity of the locking cover operation is calculated, and if the integrity of the locking cover operation meets the conditions, it indicates that the current tank body is intact.

6. A filling product analysis and quality inspection system, characterized in that: Included are: The data acquisition module acquires the detection data of the tail fold of the tank to be filled by the millimeter-wave radar, and acquires the infrared thermal image of the tank after the cap is locked; The data processing module segments the millimeter-wave radar detection data based on the loss value, calculates the void ratio of the tank tail fold, and analyzes the integrity of the tank's infrared thermal image corresponding to the locking cover operation; The control module marks the tanks that are not fully packaged.

7. A computer-readable storage medium, characterized in that: A system as claimed in claim 6 is stored.

8. A filling product analysis and quality inspection device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: The processor executes the filled product analysis and quality inspection method described in any one of claims 1 to 5 through the computer program.

Images