Degradable tableware quality inspection method and system based on image recognition
Image recognition technology to identify defects of degradable tableware and evaluate the passing degree, flexibly set the sampling rate, solve the problem of waste of manpower and time costs in the existing technology, and improve the reliability of product quality and the ability to continuously improve the production process.
Patent Information
- Application Number
- CN202411098246.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-12
- Publication Date
- 2024-11-26
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing quality inspection methods for biodegradable tableware are difficult to flexibly set the sampling rate based on the initial inspection data, resulting in waste of manpower and time costs.
Using image recognition-based methods, the size defects, surface damage defects and surface color defects of degradable tableware are identified through deep learning technology, and the sampling rate is set based on the qualification evaluation value of the production batch team, and the sampling rate is adjusted through resistance testing and ingredient detection. rate, determine whether the product is qualified.
It has achieved accurate management of product quality based on actual testing data, saving resources, improving the reliability of product quality, timely discovering and solving potential problems, and promoting continuous improvement of the production process.
Smart Images

Figure CN119026983A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of tableware production, and specifically relates to a degradable tableware quality inspection method and system based on image recognition. Background Art
[0002] As environmental awareness increases, the market demand for degradable tableware increases, but quality control requirements also increase. Traditional manual inspection methods may be inefficient and rely on the operator's experience, so a more intelligent quality inspection solution is needed. The image recognition-based method uses advanced deep learning techniques, such as convolutional neural networks, to learn and identify various types of defects from large amounts of image data. Image fusion technology is used in this process to increase the diversity and complexity of training data. The image recognition-based degradable tableware quality inspection method can improve the accuracy and efficiency of quality inspection by automating defect detection, while reducing the impact of human factors on the inspection results. In addition, through a continuous feedback mechanism, the system can be continuously optimized and improved to ensure stability and consistency in different production batches.
[0003] Most of the existing quality inspection methods for degradable tableware conduct random inspections on degradable tableware at a fixed ratio. It is difficult to flexibly set the random inspection rate based on the initial inspection data of the degradable tableware, resulting in a waste of manpower and time costs. Summary of the invention
[0004] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a degradable tableware quality inspection method and system based on image recognition, which is used to solve the technical problem that it is difficult to flexibly set the sampling rate based on the initial inspection data of the degradable tableware, resulting in a waste of manpower and time costs.
[0005] In order to solve the above problems, the first aspect of the present invention provides a method for quality inspection of degradable tableware based on image recognition, comprising the following steps:
[0006] According to the layout of the production line, the degradable tableware is divided into several production line groups according to the different production lines, and the degradable tableware in the production line group is divided into several production batch groups according to the different product batches;
[0007] Deploy multiple high-resolution cameras on the quality inspection line of the production batch group to capture image data of biodegradable tableware from multiple angles and perform data preprocessing, including denoising, contrast enhancement, and image normalization;
[0008] Use deep learning technology to identify the image data of biodegradable tableware for each production batch group, and detect size defects, surface damage defects, and surface color defects of biodegradable tableware;
[0009] According to the data of size defects, surface damage defects and surface color defects of the biodegradable tableware tested by the production batch group, the qualification of the corresponding production batch group is evaluated, and the sampling rate of the corresponding production batch group is set according to the qualification evaluation values of different production batch groups;
[0010] According to the sampling rate of the production batch group, the degradable tableware of the corresponding production batch group is sampled and subjected to resistance test and ingredient test; according to the results of resistance test and ingredient test, the sampling rate of the products of the production line group corresponding to the production batch group is adjusted, and it is judged whether the degradable tableware of the corresponding production batch group is qualified.
[0011] As a further solution of the present invention: using deep learning technology to identify the image data of the degradable tableware of each production batch group, and detecting the size defects, surface damage defects and surface color defects of the degradable tableware, the following steps are included:
[0012] Obtain historical data of image data of qualified and defective biodegradable tableware from each production line, and pre-process the images, including: denoising and contrast enhancement;
[0013] The defect types of defective degradable tableware are divided into size defects, surface damage defects and surface color defects, the defect types of the image data of the defective degradable tableware are annotated, defect type labels are added, and the defect positions are marked with marking boxes;
[0014] Randomly extract the same number of image data from defect type image data of size defects, surface damage defects and surface color defects, randomly combine the extracted image data of different defect types, perform image fusion on the combined image data, and the fused image data retains the defect type labels of all image data;
[0015] Using the labeled image data and the fused image data, a neural network model is trained to identify defects in degradable tableware.
[0016] The trained neural network model is used to detect size defects, surface damage defects and surface color defects of biodegradable tableware.
[0017] As a further solution of the present invention: according to the size defect, surface damage defect and surface color defect data of the degradable tableware detected by the production batch group, the qualification of the corresponding production batch group is evaluated, including the following steps:
[0018] Obtain the size defect, surface damage defect and surface color defect data of biodegradable tableware tested by the production batch team;
[0019] Count the number of degradable tableware of each defect type;
[0020] The qualification of the corresponding production batch group is evaluated by the following formula:
[0021]
[0022] Among them, W is the evaluation value of the qualification of the production batch group, S is the total number of degradable tableware in the production batch group, S is the total number of defective degradable tableware in the production batch group, A1 is the number of degradable tableware with size defects in the production batch group, B1 is the number of degradable tableware with surface damage defects in the production batch group, and C1 is the number of degradable tableware with surface color defects in the production batch group.
[0023] As a further solution of the present invention: according to the qualification evaluation values of different production batch groups, the sampling rate of the corresponding production batch group is set, including the following steps:
[0024] Obtain the conformity assessment values of different production batch groups;
[0025] The sampling rate of the corresponding production batch group is set by the following formula:
[0026]
[0027] Where f is the sampling rate of the corresponding production batch group.
[0028] As a further solution of the present invention: sampling the degradable tableware of the corresponding production batch group according to the sampling rate of the production batch group, and performing tolerance test and component test on the sampled degradable tableware, including the following steps:
[0029] According to the sampling rate of the production batch group, the biodegradable tableware of the corresponding production batch group is sampled and randomly divided into two groups for tolerance test;
[0030] The two groups of degradable tableware were subjected to mechanical property tests and thermal stability tests respectively, and image data of the degradable tableware during the test were collected;
[0031] The biodegradable tableware that has completed the resistance test is also subject to ingredient testing.
[0032] As a further solution of the present invention: performing mechanical property test and thermal stability test on two groups of degradable tableware respectively, and collecting image data of the degradable tableware during the test, including the following steps:
[0033] The two groups of biodegradable tableware were tested for mechanical properties and thermal stability;
[0034] For the group undergoing mechanical property testing, the pressure applied to the degradable tableware is gradually increased, and image data of the degradable tableware is collected during the process of increasing pressure;
[0035] For the group undergoing the thermal stability test, the temperature of the test environment of the degradable tableware is gradually increased, and image data and weight data of the degradable tableware are collected during the temperature increase process.
[0036] As a further solution of the present invention: according to the results of the tolerance test and the component test, the sampling rate of the products of the production line group corresponding to the production batch group is adjusted, and whether the degradable tableware of the corresponding production batch group is qualified is determined, including the following steps:
[0037] According to the image data of the degradable tableware during the test, the mechanical strength, thermal stability and component qualification rate of the degradable tableware are evaluated;
[0038] Based on the evaluation results of the mechanical strength, thermal stability and component qualification rate of the degradable tableware, the durability of the sampled tableware was evaluated;
[0039] If the durability evaluation result of the sampled tableware of a production batch group is lower than the preset threshold, the sampling rate of the products of the production line group corresponding to the production batch group will be increased, and a secondary sampling will be conducted. If the durability evaluation result of the secondary sampling is lower than the preset threshold, the degradable tableware of the corresponding production batch group is unqualified; otherwise, the degradable tableware of the corresponding production batch group is qualified.
[0040] As a further solution of the present invention: evaluating the mechanical strength, thermal stability and component qualification rate of the degradable tableware according to the image data of the degradable tableware during the test, comprising the following steps:
[0041] In the process of obtaining the mechanical properties test, by collecting image data of the degradable tableware during the process of increasing pressure, the image data in the mechanical properties test process is input into the trained neural network model for identifying defects in the degradable tableware, and the corresponding mechanical properties test pressure value when the neural network model first identifies the size defect of the degradable tableware is collected as the strength pressure value, and the corresponding mechanical properties test pressure value when the neural network model first identifies the surface damage defect of the degradable tableware is collected as the toughness pressure value;
[0042] According to the strength pressure value and toughness pressure value of the sampled degradable tableware, the mechanical strength of the degradable tableware is evaluated by the following formula:
[0043]
[0044] Wherein, K is the evaluation value of the mechanical strength of the degradable tableware, F1 is the strength pressure value of the degradable tableware, and F2 is the toughness pressure value of the degradable tableware;
[0045] Acquire image data and weight data of the degradable tableware during the thermal stability test, and collect the temperature of the thermal stability test when the weight of the degradable tableware drops to 95% of the weight of the degradable tableware itself as the first resistance temperature value;
[0046] The image data during the thermal stability test is input into the neural network model trained for identifying defects in degradable tableware, and the temperature value of the thermal stability test corresponding to when the neural network model first identifies the size defect of the degradable tableware is collected as the second resistance temperature value;
[0047] According to the first resistance temperature value and the second resistance temperature value, the thermal stability of the degradable tableware is evaluated by the following formula:
[0048]
[0049] Wherein, D is the evaluation value of the thermal stability of the degradable tableware, T1 is the first resistance temperature value of the degradable tableware, and T2 is the second resistance temperature value of the degradable tableware;
[0050] Detect the ingredients of degradable tableware and analyze whether the content of non-degradable substances and harmful substances exceeds the preset threshold. If exceeded, the corresponding degradable tableware ingredients are unqualified, otherwise the corresponding degradable tableware ingredients are qualified.
[0051] As a further solution of the present invention: according to the evaluation results of the mechanical strength, thermal stability and component qualification rate of the degradable tableware, the durability of the sampled tableware is evaluated, comprising the following steps:
[0052] Obtain the evaluation results of the mechanical strength, thermal stability and component qualification rate of biodegradable tableware;
[0053] The number of degradable tableware with an assessed value of mechanical strength greater than 0.2 among the degradable tableware sampled;
[0054] The number of degradable tableware with a thermal stability assessment value greater than 1 among the degradable tableware sampled;
[0055] Count the number of qualified products tested for the composition of degradable tableware;
[0056] The durability of the sampled tableware was evaluated using the following formula:
[0057]
[0058] Among them, E is the durability evaluation value of the sampled tableware, N is the total number of sampled degradable tableware, N1 is the number of degradable tableware whose mechanical strength evaluation value is greater than 0.2, N2 is the number of degradable tableware whose thermal stability evaluation value is greater than 1, and N3 is the number of degradable tableware that passes the ingredient test.
[0059] As a further solution of the present invention: if the durability evaluation result of the sampled tableware of the production batch group is lower than the preset threshold, the sampling rate of the products of the production line group corresponding to the production batch group is increased, and a secondary sampling is performed, including the following steps:
[0060] Obtain the durability evaluation value of the sampled tableware of the production batch group;
[0061] If the durability evaluation value of the sampled tableware of a production batch group is less than 0.8, the sampling rate of the products of the production line group corresponding to the production batch group shall be multiplied by 2 and a second sampling shall be conducted.
[0062] As another aspect of the present invention: a degradable tableware quality inspection system based on image recognition, comprising:
[0063] Grouping module: According to the layout of the production line, the degradable tableware is divided into several production line groups according to the different production lines, and the degradable tableware in the production line group is divided into several production batch groups according to the different product batches;
[0064] Image acquisition module: Multiple high-resolution cameras are deployed in the quality inspection line of the production batch group to capture image data of biodegradable tableware from multiple angles and pre-process the data, including denoising, contrast enhancement and image normalization;
[0065] Defect recognition module: Use deep learning technology to identify the image data of degradable tableware of each production batch group, and detect size defects, surface damage defects and surface color defects of degradable tableware;
[0066] Initial inspection qualification evaluation module: Based on the size defect, surface damage defect and surface color defect data of the degradable tableware tested by the production batch group, the qualification of the corresponding production batch group is evaluated, and the sampling rate of the corresponding production batch group is set according to the qualification evaluation values of different production batch groups;
[0067] Sampling inspection module: Sampling inspection is performed on the degradable tableware of the corresponding production batch group according to the sampling inspection rate of the production batch group, and resistance test and composition test are performed on the sampled degradable tableware; according to the results of resistance test and composition test, the sampling inspection rate of the products of the production line group of the corresponding production batch group is adjusted, and it is determined whether the degradable tableware of the corresponding production batch group is qualified.
[0068] Compared with the prior art, the present invention has the following beneficial effects:
[0069] The present invention detects the size defects, surface damage defects and surface color defect data of the degradable tableware according to the production batch group, evaluates the qualification of the corresponding production batch group, and sets the sampling rate of the corresponding production batch group according to the qualification evaluation values of different production batch groups; according to the initial inspection results, by setting the sampling rate according to the actual detection data, the product quality can be managed and controlled more accurately. The qualification evaluation values of different production batch groups can reflect their production quality levels, and then adjust the sampling rate to ensure the production of high-quality products. Setting the sampling rate based on actual data can effectively save resources. The sampling rate of production batch groups with high initial inspection qualification is relatively low, saving manpower and time costs, while the sampling rate for batches with higher quality risks is relatively high, which is convenient for timely detection of problems and adjustments.
[0070] The present invention performs tolerance test and component test on the sampled degradable tableware; according to the results of tolerance test and component test, the sampling rate of the products of the production line group corresponding to the production batch group is adjusted, and it is judged whether the degradable tableware of the corresponding production batch group is qualified. Through tolerance test and component test, the quality and performance of the product can be evaluated more accurately, avoiding missing potential problems due to simple sampling, and improving the reliability of product quality. At the same time, through the results of tolerance test and component test, the product meets the expected use standards and safety requirements; after detailed component analysis and tolerance test of the sampled problematic products, it can be traced back to the specific production batch group and production line, which helps to quickly locate and solve potential problems in the production process and promote continuous improvement. BRIEF DESCRIPTION OF THE DRAWINGS
[0071] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0072] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION
[0073] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than 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.
[0074] See also Figure 1 The first aspect of the present invention provides a method for quality inspection of degradable tableware based on image recognition, comprising the following steps:
[0075] According to the layout of the production line, the degradable tableware is divided into several production line groups according to the different production lines, and the degradable tableware in the production line group is divided into several production batch groups according to the different product batches;
[0076] Deploy multiple high-resolution cameras on the quality inspection line of the production batch group to capture image data of biodegradable tableware from multiple angles and perform data preprocessing, including denoising, contrast enhancement, and image normalization;
[0077] Use deep learning technology to identify the image data of biodegradable tableware for each production batch group, and detect size defects, surface damage defects, and surface color defects of biodegradable tableware;
[0078] According to the data of size defects, surface damage defects and surface color defects of the biodegradable tableware tested by the production batch group, the qualification of the corresponding production batch group is evaluated, and the sampling rate of the corresponding production batch group is set according to the qualification evaluation values of different production batch groups;
[0079] According to the sampling rate of the production batch group, the degradable tableware of the corresponding production batch group is sampled and subjected to resistance test and ingredient test; according to the results of resistance test and ingredient test, the sampling rate of the products of the production line group corresponding to the production batch group is adjusted, and it is judged whether the degradable tableware of the corresponding production batch group is qualified.
[0080] Specifically, in this embodiment, the degradable tableware is divided into several production line groups according to the layout of the production line, and the degradable tableware in the production line group is divided into several production batch groups according to different product batches; the image data of the degradable tableware of each production batch group is identified by deep learning technology, and the size defects, surface damage defects and surface color defects of the degradable tableware are detected; the size defects, surface damage defects and surface color defects of the degradable tableware can be quickly detected, which is convenient for improving product quality, reducing the defective rate, and conducting a quick initial inspection of the degradable tableware.
[0081] According to the data of size defects, surface damage defects and surface color defects of biodegradable tableware tested by the production batch group, the qualification of the corresponding production batch group is evaluated, and the sampling rate of the corresponding production batch group is set according to the qualification evaluation values of different production batch groups; according to the initial inspection results, by setting the sampling rate according to the actual inspection data, the product quality can be managed and controlled more accurately. The qualification evaluation values of different production batch groups can reflect their production quality levels, and then adjust the sampling rate to ensure the production of high-quality products. Setting the sampling rate based on actual data can effectively save resources. The sampling rate of production batch groups with high initial inspection qualification is relatively low, saving manpower and time costs, while the sampling rate for batches with higher quality risks is relatively high, which is convenient for timely detection of problems and adjustments.
[0082] According to the sampling rate of the production batch group, the biodegradable tableware of the corresponding production batch group is sampled, and the biodegradable tableware sampled is subjected to tolerance test and component test; according to the results of tolerance test and component test, the sampling rate of the products of the production line group of the corresponding production batch group is adjusted, and it is judged whether the biodegradable tableware of the corresponding production batch group is qualified. Through tolerance test and component test, the quality and performance of the product can be evaluated more accurately, potential problems can be avoided from being missed due to simple sampling, and the reliability of product quality can be improved. At the same time, through the results of tolerance test and component test, the product meets the expected use standards and safety requirements, and the market competitiveness and customer trust of the product are enhanced; more importantly, after detailed component analysis and tolerance test of the problem products sampled, they can be traced back to the specific production batch group and production line, which helps to quickly locate and solve potential problems in the production process and promote continuous improvement.
[0083] In one embodiment of the present invention, the image data of the degradable tableware of each production batch group is recognized by deep learning technology to detect the size defects, surface damage defects and surface color defects of the degradable tableware, including the following steps:
[0084] Obtain historical data of image data of qualified and defective biodegradable tableware from each production line, and pre-process the images, including: denoising and contrast enhancement;
[0085] The defect types of defective degradable tableware are divided into size defects, surface damage defects and surface color defects, the defect types of the image data of the defective degradable tableware are annotated, defect type labels are added, and the defect positions are marked with marking boxes;
[0086] Randomly extract the same number of image data from defect type image data of size defects, surface damage defects and surface color defects, randomly combine the extracted image data of different defect types, perform image fusion on the combined image data, and the fused image data retains the defect type labels of all image data;
[0087] Using the labeled image data and the fused image data, a neural network model is trained to identify defects in degradable tableware.
[0088] The trained neural network model is used to detect size defects, surface damage defects and surface color defects of biodegradable tableware.
[0089] Specifically, by randomly extracting the same number of image data from defect type image data of size defects, surface damage defects and surface color defects, randomly combining the extracted image data of different defect types, and performing image fusion on the combined image data, the fused image data retains the defect type labels of all image data; the neural network model is trained by the labeled image data and the fused image data to identify the defects of degradable tableware; the image data of different types of defects are fused so that the training model can learn and identify data with multiple defect types coexisting at the same time, thereby improving the comprehensiveness and comprehensiveness of the detection. Image fusion technology can increase the diversity of training data, help avoid overfitting of the model to a specific type of data, and improve the generalization ability and recognition accuracy of the model. Compared with manually labeling a large amount of single-type defect data, image fusion can generate rich training data more quickly, saving cost and time investment. The trained neural network model can automatically identify and classify defects, greatly improving the efficiency and accuracy of defect detection, and reducing the dependence on manual operation and possible subjective errors.
[0090] In one embodiment of the present invention, based on the size defect, surface damage defect and surface color defect data of the degradable tableware detected by the production batch group, the qualification of the corresponding production batch group is evaluated, including the following steps:
[0091] Obtain the size defect, surface damage defect and surface color defect data of biodegradable tableware tested by the production batch team;
[0092] Count the number of degradable tableware of each defect type;
[0093] The qualification of the corresponding production batch group is evaluated by the following formula:
[0094]
[0095] Among them, W is the evaluation value of the qualification of the production batch group, S is the total number of degradable tableware in the production batch group, S is the total number of defective degradable tableware in the production batch group, A1 is the number of degradable tableware with size defects in the production batch group, B1 is the number of degradable tableware with surface damage defects in the production batch group, and C1 is the number of degradable tableware with surface color defects in the production batch group.
[0096] In one embodiment of the present invention, according to the qualification evaluation values of different production batch groups, setting the sampling rate of the corresponding production batch group includes the following steps:
[0097] Obtain the conformity assessment values of different production batch groups;
[0098] The sampling rate of the corresponding production batch group is set by the following formula:
[0099]
[0100] Where f is the sampling rate of the corresponding production batch group.
[0101] In one embodiment of the present invention, sampling is performed on the degradable tableware of the corresponding production batch group according to the sampling rate of the production batch group, and the tolerance test and component test are performed on the sampled degradable tableware, including the following steps:
[0102] According to the sampling rate of the production batch group, the biodegradable tableware of the corresponding production batch group is sampled and randomly divided into two groups for tolerance test;
[0103] The two groups of degradable tableware were subjected to mechanical property tests and thermal stability tests respectively, and image data of the degradable tableware during the test were collected;
[0104] The biodegradable tableware that has completed the resistance test is also subject to ingredient testing.
[0105] In one embodiment of the present invention, two groups of degradable tableware are respectively subjected to a mechanical property test and a thermal stability test, and image data of the degradable tableware during the test is collected, including the following steps:
[0106] The two groups of biodegradable tableware were tested for mechanical properties and thermal stability;
[0107] For the group undergoing mechanical property testing, the pressure applied to the degradable tableware is gradually increased, and image data of the degradable tableware is collected during the process of increasing pressure;
[0108] For the group undergoing the thermal stability test, the temperature of the test environment of the degradable tableware is gradually increased, and image data and weight data of the degradable tableware are collected during the temperature increase process.
[0109] In one embodiment of the present invention, according to the results of the tolerance test and the component test, the sampling rate of the products of the production line group corresponding to the production batch group is adjusted, and whether the degradable tableware of the corresponding production batch group is qualified is determined, including the following steps:
[0110] According to the image data of the degradable tableware during the test, the mechanical strength, thermal stability and component qualification rate of the degradable tableware are evaluated;
[0111] Based on the evaluation results of the mechanical strength, thermal stability and component qualification rate of the degradable tableware, the durability of the sampled tableware was evaluated;
[0112] If the durability evaluation result of the sampled tableware of a production batch group is lower than the preset threshold, the sampling rate of the products of the production line group corresponding to the production batch group will be increased, and a secondary sampling will be conducted. If the durability evaluation result of the secondary sampling is lower than the preset threshold, the degradable tableware of the corresponding production batch group is unqualified; otherwise, the degradable tableware of the corresponding production batch group is qualified.
[0113] In one embodiment of the present invention, the mechanical strength, thermal stability and component qualification rate of the degradable tableware are evaluated according to the image data of the degradable tableware during the test, comprising the following steps:
[0114] In the process of obtaining the mechanical properties test, by collecting image data of the degradable tableware during the process of increasing pressure, the image data in the mechanical properties test process is input into the trained neural network model for identifying defects in the degradable tableware, and the corresponding mechanical properties test pressure value when the neural network model first identifies the size defect of the degradable tableware is collected as the strength pressure value, and the corresponding mechanical properties test pressure value when the neural network model first identifies the surface damage defect of the degradable tableware is collected as the toughness pressure value;
[0115] According to the strength pressure value and toughness pressure value of the sampled degradable tableware, the mechanical strength of the degradable tableware is evaluated by the following formula:
[0116]
[0117] Wherein, K is the evaluation value of the mechanical strength of the degradable tableware, F1 is the strength pressure value of the degradable tableware, and F2 is the toughness pressure value of the degradable tableware;
[0118] Acquire image data and weight data of the degradable tableware during the thermal stability test, and collect the temperature of the thermal stability test when the weight of the degradable tableware drops to 95% of the weight of the degradable tableware itself as the first resistance temperature value;
[0119] The image data during the thermal stability test is input into the neural network model trained for identifying defects in degradable tableware, and the temperature value of the thermal stability test corresponding to when the neural network model first identifies the size defect of the degradable tableware is collected as the second resistance temperature value;
[0120] According to the first resistance temperature value and the second resistance temperature value, the thermal stability of the degradable tableware is evaluated by the following formula:
[0121]
[0122] Wherein, D is the evaluation value of the thermal stability of the degradable tableware, T1 is the first resistance temperature value of the degradable tableware, and T2 is the second resistance temperature value of the degradable tableware;
[0123] Detect the ingredients of degradable tableware and analyze whether the content of non-degradable substances and harmful substances exceeds the preset threshold. If exceeded, the corresponding degradable tableware ingredients are unqualified, otherwise the corresponding degradable tableware ingredients are qualified.
[0124] In one embodiment of the present invention, the durability of sampled tableware is evaluated based on the evaluation results of the mechanical strength, thermal stability and component qualification rate of the degradable tableware, comprising the following steps:
[0125] Obtain the evaluation results of the mechanical strength, thermal stability and component qualification rate of biodegradable tableware;
[0126] The number of degradable tableware with an assessed value of mechanical strength greater than 0.2 among the degradable tableware sampled;
[0127] The number of degradable tableware with a thermal stability assessment value greater than 1 among the degradable tableware sampled;
[0128] Count the number of qualified products tested for the composition of degradable tableware;
[0129] The durability of the sampled tableware was evaluated using the following formula:
[0130]
[0131] Among them, E is the durability evaluation value of the sampled tableware, N is the total number of sampled degradable tableware, N1 is the number of degradable tableware whose mechanical strength evaluation value is greater than 0.2, N2 is the number of degradable tableware whose thermal stability evaluation value is greater than 1, and N3 is the number of degradable tableware that passes the ingredient test.
[0132] In one embodiment of the present invention, if the durability evaluation result of the sampled tableware of the production batch group is lower than the preset threshold, the sampling rate of the products of the production line group corresponding to the production batch group is increased, and a secondary sampling is performed, including the following steps:
[0133] Obtain the durability evaluation value of the sampled tableware of the production batch group;
[0134] If the durability evaluation value of the sampled tableware of a production batch group is less than 0.8, the sampling rate of the products of the production line group corresponding to the production batch group shall be multiplied by 2 and a second sampling shall be conducted.
[0135] In another embodiment of the present invention, a biodegradable tableware quality inspection system based on image recognition is provided, comprising:
[0136] Grouping module: According to the layout of the production line, the degradable tableware is divided into several production line groups according to the different production lines, and the degradable tableware in the production line group is divided into several production batch groups according to the different product batches;
[0137] Image acquisition module: Multiple high-resolution cameras are deployed in the quality inspection line of the production batch group to capture image data of biodegradable tableware from multiple angles and pre-process the data, including denoising, contrast enhancement and image normalization;
[0138] Defect recognition module: Use deep learning technology to identify the image data of degradable tableware of each production batch group, and detect size defects, surface damage defects and surface color defects of degradable tableware;
[0139] Initial inspection qualification evaluation module: Based on the size defect, surface damage defect and surface color defect data of the degradable tableware tested by the production batch group, the qualification of the corresponding production batch group is evaluated, and the sampling rate of the corresponding production batch group is set according to the qualification evaluation values of different production batch groups;
[0140] Sampling inspection module: Sampling inspection is performed on the degradable tableware of the corresponding production batch group according to the sampling inspection rate of the production batch group, and resistance test and composition test are performed on the sampled degradable tableware; according to the results of resistance test and composition test, the sampling inspection rate of the products of the production line group of the corresponding production batch group is adjusted, and it is determined whether the degradable tableware of the corresponding production batch group is qualified.
[0141] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.
Claims
1. A method for quality inspection of degradable tableware based on image recognition, characterized in that: The following steps are involved: According to the layout of the production line, the degradable tableware is divided into several production line groups according to the different production lines, and the degradable tableware in the production line group is divided into several production batch groups according to the different product batches; Deploy multiple high-resolution cameras on the quality inspection line of the production batch group to capture image data of biodegradable tableware from multiple angles and perform data preprocessing, including denoising, contrast enhancement, and image normalization; Use deep learning technology to identify the image data of biodegradable tableware for each production batch group, and detect size defects, surface damage defects, and surface color defects of biodegradable tableware; According to the data of size defects, surface damage defects and surface color defects of the biodegradable tableware tested by the production batch group, the qualification of the corresponding production batch group is evaluated, and the sampling rate of the corresponding production batch group is set according to the qualification evaluation values of different production batch groups; According to the sampling rate of the production batch group, the degradable tableware of the corresponding production batch group is sampled and subjected to resistance test and ingredient test; according to the results of resistance test and ingredient test, the sampling rate of the products of the production line group corresponding to the production batch group is adjusted, and it is judged whether the degradable tableware of the corresponding production batch group is qualified.
2. The method for quality inspection of degradable tableware based on image recognition according to claim 1, characterized in that: The image data of the degradable tableware of each production batch group is recognized by deep learning technology to detect the size defects, surface damage defects and surface color defects of the degradable tableware, including the following steps: Obtain historical data of image data of qualified and defective biodegradable tableware from each production line, and pre-process the images, including: denoising and contrast enhancement; The defect types of defective degradable tableware are divided into size defects, surface damage defects and surface color defects, the defect types of the image data of the defective degradable tableware are annotated, defect type labels are added, and the defect positions are marked with marking boxes; Randomly extract the same number of image data from defect type image data of size defects, surface damage defects and surface color defects, randomly combine the extracted image data of different defect types, perform image fusion on the combined image data, and the fused image data retains the defect type labels of all image data; Using the labeled image data and the fused image data, a neural network model is trained to identify defects in degradable tableware. The trained neural network model is used to detect size defects, surface damage defects and surface color defects of biodegradable tableware.
3. The method for quality inspection of degradable tableware based on image recognition according to claim 1, characterized in that: According to the size defect, surface damage defect and surface color defect data of the biodegradable tableware detected by the production batch group, the qualification of the corresponding production batch group is evaluated, including the following steps: Obtain the size defect, surface damage defect and surface color defect data of biodegradable tableware tested by the production batch team; Count the number of degradable tableware of each defect type; The qualification of the corresponding production batch group is evaluated by the following formula: Among them, W is the evaluation value of the qualification of the production batch group, S is the total number of degradable tableware in the production batch group, S is the total number of defective degradable tableware in the production batch group, A1 is the number of degradable tableware with size defects in the production batch group, B1 is the number of degradable tableware with surface damage defects in the production batch group, and C1 is the number of degradable tableware with surface color defects in the production batch group.
4. The method for quality inspection of degradable tableware based on image recognition according to claim 3, characterized in that: According to the qualification evaluation values of different production batch groups, the sampling rate of the corresponding production batch groups is set, including the following steps: Obtain the conformity assessment values of different production batch groups; The sampling rate of the corresponding production batch group is set by the following formula: Where f is the sampling rate of the corresponding production batch group.
5. The method for quality inspection of degradable tableware based on image recognition according to claim 1, characterized in that: According to the sampling rate of the production batch group, the biodegradable tableware of the corresponding production batch group is sampled, and the biodegradable tableware sampled is subjected to tolerance test and component test, including the following steps: According to the sampling rate of the production batch group, the biodegradable tableware of the corresponding production batch group is sampled and randomly divided into two groups for tolerance test; The two groups of degradable tableware were subjected to mechanical property tests and thermal stability tests respectively, and image data of the degradable tableware during the test were collected; The biodegradable tableware that has completed the resistance test is also subject to ingredient testing.
6. The method for quality inspection of degradable tableware based on image recognition according to claim 5, characterized in that: The two groups of degradable tableware are respectively subjected to mechanical property tests and thermal stability tests, and image data of the degradable tableware during the test are collected, including the following steps: The two groups of biodegradable tableware were tested for mechanical properties and thermal stability; For the group undergoing mechanical property testing, the pressure applied to the degradable tableware is gradually increased, and image data of the degradable tableware is collected during the process of increasing pressure; For the group undergoing the thermal stability test, the temperature of the test environment of the degradable tableware is gradually increased, and image data and weight data of the degradable tableware are collected during the temperature increase process.
7. The method for quality inspection of degradable tableware based on image recognition according to claim 6, characterized in that: According to the results of the tolerance test and the component test, the sampling rate of the products of the production line group corresponding to the production batch group is adjusted, and whether the degradable tableware of the corresponding production batch group is qualified is judged, including the following steps: According to the image data of the degradable tableware during the test, the mechanical strength, thermal stability and component qualification rate of the degradable tableware are evaluated; Based on the evaluation results of the mechanical strength, thermal stability and component qualification rate of the degradable tableware, the durability of the sampled tableware was evaluated; If the durability evaluation result of the sampled tableware of a production batch group is lower than the preset threshold, the sampling rate of the products of the production line group corresponding to the production batch group will be increased, and a secondary sampling will be conducted. If the durability evaluation result of the secondary sampling is lower than the preset threshold, the degradable tableware of the corresponding production batch group is unqualified; otherwise, the degradable tableware of the corresponding production batch group is qualified.
8. The method for quality inspection of degradable tableware based on image recognition according to claim 7, characterized in that: According to the image data of the degradable tableware during the test, the mechanical strength, thermal stability and component qualification rate of the degradable tableware are evaluated, including the following steps: In the process of obtaining the mechanical properties test, by collecting image data of the degradable tableware during the process of increasing pressure, the image data in the mechanical properties test process is input into the trained neural network model for identifying defects in the degradable tableware, and the corresponding mechanical properties test pressure value when the neural network model first identifies the size defect of the degradable tableware is collected as the strength pressure value, and the corresponding mechanical properties test pressure value when the neural network model first identifies the surface damage defect of the degradable tableware is collected as the toughness pressure value; According to the strength pressure value and toughness pressure value of the sampled degradable tableware, the mechanical strength of the degradable tableware is evaluated by the following formula: Wherein, K is the evaluation value of the mechanical strength of the degradable tableware, F1 is the strength pressure value of the degradable tableware, and F2 is the toughness pressure value of the degradable tableware; Acquire image data and weight data of the degradable tableware during the thermal stability test, and collect the temperature of the thermal stability test when the weight of the degradable tableware drops to 95% of the weight of the degradable tableware itself as the first resistance temperature value; The image data during the thermal stability test is input into the neural network model trained for identifying defects in degradable tableware, and the temperature value of the thermal stability test corresponding to when the neural network model first identifies the size defect of the degradable tableware is collected as the second resistance temperature value; According to the first resistance temperature value and the second resistance temperature value, the thermal stability of the degradable tableware is evaluated by the following formula: Wherein, D is the evaluation value of the thermal stability of the degradable tableware, T1 is the first resistance temperature value of the degradable tableware, and T2 is the second resistance temperature value of the degradable tableware; Detect the ingredients of degradable tableware and analyze whether the content of non-degradable substances and harmful substances exceeds the preset threshold. If exceeded, the corresponding degradable tableware ingredients are unqualified, otherwise the corresponding degradable tableware ingredients are qualified.
9. The method for quality inspection of degradable tableware based on image recognition according to claim 8, characterized in that: According to the evaluation results of the mechanical strength, thermal stability and component qualification rate of the degradable tableware, the durability of the sampled tableware is evaluated, including the following steps: Obtain the evaluation results of the mechanical strength, thermal stability and component qualification rate of biodegradable tableware; The number of degradable tableware with an assessed value of mechanical strength greater than 0.2 among the degradable tableware sampled; The number of degradable tableware with a thermal stability assessment value greater than 1 among the degradable tableware sampled; Count the number of qualified products tested for the composition of degradable tableware; The durability of the sampled tableware was evaluated using the following formula: Among them, E is the durability evaluation value of the sampled tableware, N is the total number of sampled degradable tableware, N1 is the number of degradable tableware whose mechanical strength evaluation value is greater than 0.2, N2 is the number of degradable tableware whose thermal stability evaluation value is greater than 1, and N3 is the number of degradable tableware that passes the ingredient test.
10. A biodegradable tableware quality inspection system based on image recognition, characterized in that: include: Grouping module: According to the layout of the production line, the degradable tableware is divided into several production line groups according to the different production lines, and the degradable tableware in the production line group is divided into several production batch groups according to the different product batches; Image acquisition module: Multiple high-resolution cameras are deployed in the quality inspection line of the production batch group to capture image data of biodegradable tableware from multiple angles and pre-process the data, including denoising, contrast enhancement and image normalization; Defect recognition module: Use deep learning technology to identify the image data of degradable tableware of each production batch group, and detect size defects, surface damage defects and surface color defects of degradable tableware; Initial inspection qualification evaluation module: Based on the size defect, surface damage defect and surface color defect data of the degradable tableware tested by the production batch group, the qualification of the corresponding production batch group is evaluated, and the sampling rate of the corresponding production batch group is set according to the qualification evaluation values of different production batch groups; Sampling inspection module: Sampling inspection is performed on the degradable tableware of the corresponding production batch group according to the sampling inspection rate of the production batch group, and resistance test and composition test are performed on the sampled degradable tableware; according to the results of resistance test and composition test, the sampling inspection rate of the products of the production line group of the corresponding production batch group is adjusted, and it is determined whether the degradable tableware of the corresponding production batch group is qualified.
Citation Information
Patent Citations
Copper surface defect detection method and device based on line-scan digital camera
CN114638797A
Disposable tableware surface defect detection system based on deep learning
CN117197090A
Aluminum template quality inspection method based on machine vision
CN117890214A
