Method and system for detecting and feeding finished glass wine bottles

Abstract

The application provides a kind of finished glass wine bottle detection feeding method, comprising the following steps: S1: the step of preparing feeding, including wine bottle transfer, the step of carrying out automatic feeding and the step of carrying out automatic correction and guiding;S2: the step of conveying and preliminary arrangement, including starting conveyor belt and stabilizing speed, then guiding wine bottle, then preliminary arranging wine bottle;S3: the step of visual and laser detection;S4: the step of data processing and analysis;S5: the step of sorting;S6: the step of feedback and adjustment.The method can significantly improve the accuracy and efficiency of detection, optimize the production process, improve the overall quality control level and help reduce the rate of defective products and production costs.

Description

TECHNICAL FIELD

[0001] The present application relates to a kind of finished glass wine bottle detection feeding method and system. BACKGROUND

[0002] Finished glass wine bottle detection feeding refers to the quality detection and automatic feeding process of glass wine bottle manufactured on production line.This field currently does not have the efficient way of using advanced sensors and machine vision technology to detect the appearance, size, sealing of wine bottle and other parameters to ensure that each wine bottle meets the quality standard, therefore the present application is proposed to solve this problem. SUMMARY

[0003] The present application aims at the deficiencies of prior art, and provides a kind of finished glass wine bottle detection feeding method, which solves the above problems well.

[0004] To achieve the above requirements, the technical scheme adopted by the present application is as follows: a kind of finished glass wine bottle detection feeding method is provided, which includes the following steps:

[0005] S1: the step of preparing feeding, including wine bottle transfer, automatic feeding, automatic correction and guiding steps;

[0006] S2: the step of conveying and preliminary arrangement, including starting the conveyor belt and stabilizing the speed, then guiding the wine bottle, and then preliminarily arranging the wine bottle;

[0007] S3: the step of visual and laser detection;

[0008] S4: the step of data processing and analysis;

[0009] S5: the step of sorting;

[0010] S6: the step of feedback and adjustment.

[0011] The finished glass wine bottle detection feeding method has the following advantages:

[0012] The method can significantly improve the accuracy and efficiency of detection, optimize the production process, improve the overall quality control level and help reduce the rate of defective products and production cost. BRIEF DESCRIPTION OF DRAWINGS

[0013] The drawings described herein are used to provide further understanding of the present application, and form a part of the present application, in which the same reference numerals are used to represent the same or similar parts, the schematic embodiments of the present application and their descriptions are used to explain the present application, and do not constitute undue limitation on the present application. In the drawings:

[0014] Figure 1 A schematic flowchart of a method for inspecting and loading finished glass wine bottles according to an embodiment of this application is shown. Detailed Implementation

[0015] To make the objectives, technical solutions and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and specific embodiments.

[0016] In the following description, references to "an embodiment," "an embodiment," "an example," "example," etc., indicate that the described embodiment or example includes a particular feature, structure, characteristic, property, element, or limitation, but not every embodiment or example necessarily includes that particular feature, structure, characteristic, property, element, or limitation. Furthermore, the repeated use of the phrase "an embodiment according to this application," while referring to the same embodiment, does not necessarily refer to the same embodiment.

[0017] For simplicity, certain technical features known to those skilled in the art are omitted in the following description.

[0018] According to one embodiment of this application, a method for inspecting and loading finished glass wine bottles is provided, comprising the following steps:

[0019] 1. Material preparation

[0020] Material preparation is the starting point of the entire inspection and loading process, ensuring that glass bottles smoothly enter the automated loading system from the storage area or production line. This step involves the following sub-steps:

[0021] (1) Bottle Transfer: First, finished glass bottles are transferred from the production line or storage area to the input of the automated feeding system. At the end of the production line, there is a temporary storage area for the finished glass bottles. These bottles are then transported to the input area of ​​the feeding system manually or by automated equipment (such as forklifts or automated guided vehicles, AGVs).

[0022] (2) Vibratory feeder or automatic feeding equipment: To ensure that the bottles are arranged in an orderly manner on the conveyor belt, a vibratory feeder or other automatic feeding equipment is used. The vibratory feeder moves the bottles by vibration and gradually arranges them into a single row. For large-scale production lines, a multi-layer vibratory feeder system is used to improve feeding speed and efficiency. The design of the vibratory feeder must ensure that the bottles will not break due to excessive vibration.

[0023] (3) Automatic correction and guidance: Before the bottles enter the conveyor belt, an automatic correction and guidance device is required to ensure that each bottle enters the conveyor belt in the correct direction and posture. The correction device includes mechanical baffles, conveyor belt guiding system, etc., to ensure that the bottles do not tip over or become misaligned when entering the detection system.

[0024] (4) Safety Measures: During the loading preparation stage, safety measures must also be considered to prevent wine bottles from breaking due to collisions or other reasons during the transfer process. This includes installing buffer devices at key locations and using soft materials to wrap the edges of the vibrating disc.

[0025] Through these preparations, it is ensured that the wine bottles are arranged neatly and in the correct direction and position before entering the conveyor belt, avoiding errors or failures in the subsequent detection process.

[0026] 2. Conveying and Preliminary Arrangement

[0027] The conveying and preliminary arrangement stage is an important link before the wine bottles enter the detection station. This process ensures that the wine bottles enter the detection system at a stable speed and in the correct posture for accurate detection. The specific description is as follows:

[0028] (1) Start the conveyor belt: After the loading preparation is completed, start the conveyor belt system. The design of the conveyor belt needs to ensure its smooth operation and adjustable speed. The surface of the conveyor belt will have anti-slip design to prevent wine bottles from falling or sliding during movement.

[0029] (2) Stable speed: The speed of the conveyor belt needs to be adjusted according to the rhythm of the production line. Too fast speed will cause wine bottles to fall or collide with each other, and too slow speed will reduce production efficiency. Frequency converter is used to control the speed of the conveyor belt, which can be flexibly adjusted according to actual needs.

[0030] (3) Guide device: During the conveying process, guide devices are set to ensure that the wine bottles always remain in the central position of the conveyor belt. The guide device is a fixed mechanical baffle and also a dynamic adjustment guide system. The dynamic guide system detects the position of the wine bottle through sensors and automatically adjusts the angle and position of the guide device to keep the wine bottle on the correct track.

[0031] (4) Preliminary arrangement: If the wine bottles are not completely arranged neatly during the loading stage, additional arrangement devices are set on the conveyor belt. For example, by setting multiple transverse baffles, the wine bottles are separated one by one to avoid squeezing or overlapping on the conveyor belt. The arrangement device needs to be customized according to the size and shape of the wine bottle to ensure that it can adapt to different specifications of the wine bottle.

[0032] (5) Buffer zone: A buffer zone is set before the detection station to adjust the flow of wine bottles on the conveyor belt. The buffer zone is a short-distance deceleration conveyor belt or a specially designed buffer box to ensure that the wine bottles maintain appropriate spacing and speed before entering the detection station.

[0033] (6) Sensor detection: Sensors are installed on the conveyor belt to monitor the position and state of the wine bottles in real time. If a wine bottle is detected to be tilted or congested, the system will automatically issue an alarm and stop the conveyor belt for the operator to check and handle.

[0034] Through these measures, it ensures that the wine bottles remain stable and in the correct orientation during transportation, laying the foundation for subsequent detection steps.

[0035] 3. Visual and Laser Detection

[0036] Visual and laser detection is the most critical step in the entire detection and feeding process. Through precise detection equipment and systems, it ensures that each glass wine bottle meets quality standards. The specific description is as follows:

[0037] (1) Visual detection system: The visual detection system is composed of industrial cameras, lighting equipment and image processing software. When the wine bottle enters the detection station, the industrial camera will take high-resolution images of the wine bottle. In order to ensure the clarity of the image, the lighting equipment needs to provide uniform and shadow-free light, which is selected according to the shape and material of the wine bottle.

[0038] (2) Appearance defect detection: The image processing software will analyze the images of the wine bottle taken, detecting the appearance defects of the wine bottle. Appearance defects include cracks, scratches, bubbles and label errors. Through pre-set defect standards, the system quickly identifies and marks unqualified wine bottles.

[0039] (3) Laser measurement equipment: Laser measurement equipment is used to detect the size and shape of the wine bottle. Through laser scanning technology, the system obtains three-dimensional contour data of the wine bottle and compares it with the standard model. Laser measurement equipment includes laser emitter, receiver and data processing unit, which can accurately measure the height, diameter and thickness parameters of the wine bottle.

[0040] (4) Data processing and analysis: The detection system will send the data of visual detection and laser measurement to the central control system. The control system analyzes and processes the data through pre-set algorithms to determine whether the wine bottle meets the quality standards. The data analysis process includes image processing, feature extraction and model matching.

[0041] (5) Real-time feedback: The detection system will provide real-time feedback on the detection results and display the detection status of each wine bottle on the system interface. If an unqualified wine bottle is detected, the system will automatically record the defect type and location for subsequent processing. Operators can view detailed detection reports through the system interface to timely identify and solve production problems.

[0042] (6) Multi-angle detection: In order to improve the comprehensiveness of detection, the visual detection system will take images of the wine bottle from multiple angles. For example, multiple industrial cameras are installed at different positions on the conveyor belt to take images of the top, side and bottom of the wine bottle. Through multi-angle image analysis, every detail is fully detected.

[0043] (7) Fault Handling: During the detection process, if a device malfunctions or the detection system encounters an anomaly, the system will automatically issue an alarm and stop detection. The operator needs to check and eliminate the fault in time to ensure the normal operation of the detection system. Faults include camera out of focus, laser emitter failure.

[0044] Through these meticulous detection steps, it is ensured that each wine bottle is subjected to comprehensive quality inspection and meets production standards.

[0045] 4. Data Processing and Analysis

[0046] The data processing and analysis stage is the core of the detection system. Through analysis and processing of the collected data, it is determined whether the quality of the wine bottle meets the standard. The specific description is as follows:

[0047] (1) Data Collection: When the wine bottle passes through the visual detection system and laser measurement equipment, the system will collect images and three-dimensional contour data in real time. The collected data includes the appearance image, size parameter, shape information, etc. of the wine bottle. These data will be sent to the central control system through high-speed data transmission interface (such as Ethernet or USB).

[0048] (2) Data Preprocessing: The raw data collected needs to be preprocessed to improve the accuracy of analysis. The preprocessing of image data includes denoising, edge detection, image enhancement. The preprocessing of laser measurement data includes coordinate conversion, error correction. The preprocessed data is more suitable for subsequent analysis and processing.

[0049] (3) Image Processing: For image data collected by the visual detection system, image processing algorithms will be analyzed in detail. Common image processing techniques include template matching, feature extraction, image segmentation, etc. The system will compare the processed image with the preset standard image to identify appearance defects and label errors.

[0050] (4) Size and Shape Analysis: For data collected by laser measurement equipment, the system will perform size and shape analysis. Through three-dimensional modeling technology, the system reconstructs the three-dimensional contour of the wine bottle and compares it with the standard model. The parameters analyzed include the height, diameter, wall thickness, mouth shape of the wine bottle. Any parameter that exceeds the tolerance range will be marked as unqualified.

[0051] (5) Defect Classification: The detection system will classify the identified defects according to the preset defect standards. For example, cracks, scratches, bubbles, etc. are classified according to severity. Size and shape defects are classified according to the degree of exceeding the tolerance range. The results of classification are helpful for subsequent processing and improvement.

[0052] (6) Qualification determination: The system will analyze the results of image processing and dimension analysis to determine whether the wine bottle is qualified. The qualification determination algorithm needs to consider multiple factors, including the number, type, location, and severity of defects. Only when all parameters meet the standards will the wine bottle be determined as qualified.

[0053] (7) Data storage: The detection system will store the detection data of each wine bottle in the database. The stored data includes raw images, processed images, dimension parameters, analysis results, etc. Through data storage, the system realizes long-term preservation and traceability of detection data.

[0054] (8) Statistical analysis: Through statistical analysis of stored data, the system finds the rules and problems in production. For example, by counting the number and type of unqualified wine bottles, it analyzes the weak links in the production process and takes improvement measures. Statistical analysis also generates various reports for production management and quality control.

[0055] (9) Real-time monitoring: The detection system will display the detection status of each wine bottle in real time on the control interface. The operator can view detailed detection reports through the interface to understand the current production status. If the system detects abnormal conditions, it will automatically issue an alarm to remind the operator to check and handle.

[0056] Through these detailed data processing and analysis steps, the quality of each wine bottle is thoroughly checked and evaluated, and the produced wine bottles meet the quality standards.

[0057] 5. Sorting

[0058] The sorting stage is the process of classifying qualified and unqualified wine bottles according to the detection results. Through automated sorting devices, qualified wine bottles enter the next production link, and unqualified wine bottles are removed from the production line. The specific description is as follows:

[0059] (1) Type of sorting device: Common sorting devices include pneumatic sorters, electric sorters, and mechanical arms. Pneumatic sorters push wine bottles to designated positions through air flow, electric sorters move wine bottles to different conveyors through electric mechanisms, and mechanical arms control the grabbing and placing of wine bottles through programming.

[0060] (2) Sensor detection: Before the wine bottle enters the sorting device, the system will again confirm the position and state of the wine bottle through the sensor. The sensor detects whether the wine bottle is correctly arranged, whether it is tilted or moved. If abnormal conditions are detected, the system will issue an alarm and pause the sorting operation to avoid mis-sorting.

[0061] (3) Classification Criteria: The sorting system will classify the wine bottles into two categories: pass and fail based on the detection results. The pass bottles will be sent to the next production or packaging process, while the fail bottles will be removed from the production line and placed in a special storage area for disposal or scrap. The classification criteria need to be set according to production requirements to ensure accurate classification.

[0062] (4) Automatic Sorting Operation: The sorting device will automatically classify the wine bottles through the control system's instructions. The pneumatic sorter will push the fail bottles away from the conveyor belt through air flow, the electric sorter will move the bottles to different conveyor belts or storage areas through electric mechanisms, and the robotic arm will grab the fail bottles and place them in designated locations.

[0063] (5) Classified Storage: The pass and fail bottles will be placed in different conveyor belts or storage areas. The pass bottles will be directly sent to the next production or packaging line, while the fail bottles will be placed in a special storage area for further processing. The classified storage area needs to be clearly marked to avoid confusion.

[0064] (6) Record and Traceability: The sorting system will record the classification results of each wine bottle, including detection data and sorting time. Through record and traceability, production management personnel can understand the detection and sorting process of each wine bottle, analyze the reasons for the fail bottles, and take appropriate improvement measures.

[0065] (7) Abnormal Handling: If system failure or operation abnormalities are detected during sorting, the system will automatically issue an alarm and pause the sorting operation. The operator needs to check and handle the failure in a timely manner to ensure the normal operation of the sorting system. Abnormal situations include insufficient air pressure for the pneumatic sorter, motor failure for the electric sorter, and position deviation of the robotic arm.

[0066] (8) Efficiency Optimization: The sorting system needs to be optimized according to the production line's rhythm to ensure that the sorting speed matches the production speed. By adjusting the speed and parameters of the sorting device, the sorting efficiency can be improved, and the production line's downtime can be reduced. Efficiency optimization also includes regular maintenance and maintenance of the sorting device to avoid equipment failure affecting production.

[0067] Through these detailed sorting steps, the pass and fail wine bottles are accurately classified, ensuring smooth operation of the production line and product quality.

[0068] 6. Feedback and Adjustment

[0069] Feedback and adjustment are important steps in the detection and feeding process. Through analysis and summary of detection and sorting results, timely adjustment and optimization of production processes are carried out to ensure production stability and high-quality products. The specific description is as follows:

[0070] (1) Detection result feedback: The detection system records the detection data and sorting results of each wine bottle in real time. These data include appearance defects, size parameters, shape information, and pass or fail. The system will generate a detection report by summarizing these data and display it on the control interface in real time for operators and production managers to view.

[0071] (2) Abnormal situation alarm: If the detection system finds abnormal situations, such as an abnormal increase in the number of unqualified wine bottles or frequent occurrence of certain defects, the system will automatically send an alarm signal. The alarm signal includes sound and light alarms and interface prompts to remind operators to check the production line in time and find and eliminate abnormal reasons.

[0072] (3) Data analysis and statistics: Through statistical analysis of detection data, problems and weak links in the production process are found. Analysis methods include trend analysis, defect classification statistics, and process parameter comparison. Through analysis, the root cause of defects is found, and corresponding improvement measures are developed.

[0073] (4) Process adjustment: According to the results of data analysis, the production process is adjusted. For example, if it is found that a certain process leads to a large number of appearance defects, adjust the operating parameters or equipment settings of the process; if it is found that size deviation frequently occurs, recalibrate the relevant equipment. Process adjustment needs to be carried out when production is suspended to avoid affecting normal production.

[0074] (5) Equipment maintenance and calibration: In order to ensure the accuracy and stability of the detection system, regular maintenance and calibration of the equipment are necessary. Maintenance includes cleaning the lens, checking the sensor, and calibrating the laser measurement equipment. Calibration work needs to be carried out according to the requirements of the equipment manufacturer to ensure the accuracy of the detection results.

[0075] (6) Operator training: In order to improve the efficiency and accuracy of the detection system, regular training of operators is also necessary. Training content includes detection system operation method, fault handling, data analysis skills, etc. Through training, the skill level of operators is improved to ensure the stable operation of the system.

[0076] (7) Implementation of improvement measures: According to the results of data analysis, improvement measures are developed and implemented. Improvement measures include equipment upgrade, process optimization, and operation specification adjustment. After implementing improvement measures, the improvement effect needs to be verified through the detection system, and corresponding records and evaluations are made.

[0077] (8) Continuous improvement: Quality management is a continuous improvement process. Through continuous feedback and adjustment, production processes and product quality will gradually improve. Regularly hold quality analysis meetings to summarize experiences and lessons, share improvement results, form a virtuous cycle, and promote the continuous optimization of the entire production line.

[0078] Through these meticulous feedback and adjustment steps, the stability of the production process and the continuous improvement of product quality are ensured, and the goal of efficient and high-quality production is achieved.

[0079] According to one embodiment of the present application, the specific method of step 4 for data processing and analysis in the finished glass wine bottle detection and feeding method is as follows:

[0080] 1. Data preprocessing is performed in the following way:

[0081] The following formula is used to remove noise:

[0082]

[0083]

[0084] P k|k =(I-K k H)P k|k-1 ;

[0085] where, is the state estimation at time k, z k is the measurement value, H is the measurement matrix, K k is the Kalman gain, P k|k is the estimation error covariance, and R is the measurement noise covariance matrix.

[0086] The following formula is used for outlier detection:

[0087] D 2 =(x-μ) T ∑ -1 (x-μ);

[0088] where x is the data point, μ is the mean vector, and ∑ is the covariance matrix.

[0089] 2. The method of image processing and size analysis is as follows:

[0090] (1) Feature extraction: Independent component analysis is used to extract independent features of the image. Independent component analysis separates mixed signals by maximizing independence, suitable for blind source separation and image feature extraction. X = AS where X is the observation signal matrix, A is the mixing matrix, and S is the independent component matrix. By solving the matrices A and S, the independent features of the image can be obtained.

[0091] (2) Model matching: Generalized maximum likelihood estimation is applied for model matching. Generalized maximum likelihood estimation estimates parameters by maximizing the likelihood function of data.

[0092]

[0093] where L(θ|x) is the likelihood function, θ is the parameter to be estimated, and x is the observed data.

[0094] 3. In defect classification, the following methods are used:

[0095] In the defect classification stage, statistical methods can be applied to improve the accuracy of classification.

[0096] Classification algorithm: Support Vector Machines (SVM) are used to classify different types of defects. SVM separates data points of different classes by constructing a hyperplane:

[0097]

[0098] subject to y i (w T φ(x i )+b)≥1-ξ i , ξ i ≥0;

[0099] where w is the weight vector, b is the bias, ξ i is the slack variable, C is the penalty parameter, φ(x i ) is the feature mapping function, and y i is the class label.

[0100] 4. Statistics in Acceptance Determination

[0101] In the acceptance determination stage, statistical methods can be applied to optimize the determination criteria.

[0102] (1) Multivariate Statistical Process Control: Hotelling T^2 control chart is used to monitor the stability of multiple quality characteristics. Hotelling T^2 control chart provides more comprehensive process monitoring by considering the correlation between characteristics.

[0103]

[0104] where is the sample mean vector, μ is the population mean vector, and ∑ is the sample covariance matrix.

[0105] (2) Bayesian Statistical Methods: Bayesian confidence intervals are used to evaluate the reliability of parameter estimates.

[0106]

[0107] where P(θ|x) is the posterior distribution, P(x|θ) is the likelihood function, P(θ) is the prior distribution, and P(x) is the marginal likelihood.

[0108] 5. Statistics in Statistical Analysis and Feedback

[0109] In the statistical analysis and feedback stage, statistical methods can be applied to optimize the production process as follows:

[0110] (1) Multiple regression analysis: Through multiple regression analysis, multiple key factors affecting the quality of wine bottles are found, and a mathematical model is established for prediction and optimization.

[0111] y = β0 + β1x1 + β2x2 + … + β n x n + ∈;

[0112] Where y represents the response variable, x1, x2, …, xn represent the predictor variables, β0, β1, …, βn represent the regression coefficients, and ∈ represents the error term. The regression coefficients are estimated by the least squares method:

[0113]

[0114] (2) Generalized linear model: Use generalized linear models to handle data with different distributions and link functions. GLM can handle non-normal distributed data by extending the linear model.

[0115]

[0116] Where g is the link function, μ i is the expected value of the response variable, x i is the predictor variable vector, and β is the regression coefficient.

[0117] (3) Multifactor variance analysis: Through multifactor variance analysis, the comprehensive effects of different production processes and conditions on the quality of wine bottles are compared.

[0118]

[0119] Where E represents the within-group error matrix, and H represents the between-group error matrix.

[0120] The above examples only represent several embodiments of the present application, and the description is more specific and detailed, but it cannot be understood as limiting the scope of the present application. It should be noted that for ordinary skilled in the art, without departing from the concept of the present application, make several modifications and improvements, which belong to the scope of protection of the present application. Therefore, the protection scope of the present application should be subject to the claims.

Claims

1. A method for inspecting and feeding finished glass wine bottles, characterized in that, Includes the following steps: S1: The steps for preparing the material for loading include transferring the wine bottles, automatically loading the material, and automatically calibrating and guiding the material. S2: The steps of conveying and preliminary arrangement include starting the conveyor belt and stabilizing the speed, then guiding the bottles, and then performing preliminary arrangement of the bottles; S3: Steps for performing visual and laser inspection; S4: Steps for data processing and analysis; S5: The sorting steps; S6: Steps for providing feedback and making adjustments; Step S3 specifically includes: Visual inspection is performed. The visual inspection system consists of an industrial camera, lighting equipment, and image processing software. When a bottle enters the inspection station, the industrial camera captures a high-resolution image of the bottle. To ensure image clarity, the lighting equipment provides uniform and shadow-free illumination. The image processing software analyzes the captured bottle images to detect appearance defects, including cracks, scratches, bubbles, and label errors. Laser measurement equipment is used to detect the size and shape of the bottle. Through laser scanning technology, the three-dimensional contour data of the bottle is acquired and compared with a standard model. The inspection system sends the visual inspection and laser measurement data to the central control system. The control system determines whether the bottle meets the quality standards. The inspection system provides real-time feedback on the inspection results and displays the inspection status of each bottle on the system interface. If a non-conforming bottle is detected, the system automatically records the defect type and location. Step S4 specifically includes: When a wine bottle passes through a vision inspection system and laser measurement equipment, the system acquires images and 3D contour data in real time. The acquired data includes the bottle's appearance, dimensions, and shape information. This data is sent to the central control system via a high-speed data transmission interface. The raw data is preprocessed to improve the accuracy of the analysis. Preprocessing of the laser measurement data includes coordinate transformation and error correction. For the image data acquired by the vision inspection system, image processing algorithms analyze the processed image and compare it with a preset standard image to identify appearance defects and label errors. For the data acquired by the laser measurement equipment, size and shape analysis is performed to reconstruct the bottle's 3D contour, which is then compared with a standard model. The inspection system classifies the identified defects according to preset defect standards. By comprehensively analyzing the results of image processing and size analysis, the system determines whether the bottle is qualified. The inspection system stores the inspection data for each bottle in a database; the stored data includes the raw image, processed image, dimensions, and analysis results. Data preprocessing is performed as follows when conducting data processing and analysis: The following formula is used to remove noise: ; ； ； in, It is the state estimate at time k. These are measured values. It is a measurement matrix. It is Kalman gain. It estimates the error covariance. It is the measurement noise covariance matrix; Outlier detection is performed using the following formula: ; Where x is a data point, μ is the mean vector, and Σ is the covariance matrix; The methods for image processing and size analysis are as follows: Feature extraction uses independent component analysis to extract independent features of the image. Independent component analysis separates mixed signals by maximizing independence and is suitable for blind source separation and image feature extraction. X=AS, where X is the observed signal matrix, A is the mixture matrix, and S is the independent component matrix. By solving matrices A and S, the independent features of the image can be obtained. Model matching is performed using generalized maximum likelihood estimation, which estimates parameters by maximizing the likelihood function of the data. ;in, θ is the likelihood function, θ is the parameter to be estimated, and x is the observed data; The following method is used for defect classification: Support Vector Machines (SVMs) are used to classify different types of defects. SVMs separate data points of different categories by constructing hyperplanes. ； ； Where w is the weight vector and b is the bias. C is the slack variable, C is the penalty parameter, and is the feature mapping function. It is a category label; The specific methods for determining eligibility are as follows: Hotelling T^2 control charts are used to monitor the stability of multiple quality characteristics. By considering the correlation between the characteristics, Hotelling T^2 control charts provide more comprehensive process monitoring. ； in, Σ is the sample mean vector, μ is the population mean vector, and Σ is the sample covariance matrix; Use Bayesian confidence intervals to evaluate the reliability of parameter estimates; ； Where P(θ|x) is the posterior distribution, P(x|θ) is the likelihood function, P(θ) is the prior distribution, and P(x) is the marginal likelihood; The specific methods for conducting statistical analysis and feedback are as follows: Through multiple regression analysis, we identified several key factors affecting the quality of wine bottles and established a mathematical model for prediction and optimization. ； Where y represents the response variable, x1, x2, ..., xn represent the predictor variables, β0, β1, ..., βn represent the regression coefficients, and ϵ represents the error term. The regression coefficients are estimated using the least squares method. ; Using a generalized linear model to handle data with different distributions and link functions, GLM can handle non-normally distributed data by extending the linear model; ； Where g is the link function, It is the expectation of the response variable. It is a vector of predictor variables, and β is the regression coefficient; Multivariate analysis of variance was used to compare the comprehensive impact of different production processes and conditions on the quality of wine bottles; ； Where E represents the within-group error matrix and H represents the between-group error matrix.

Images