An automated color difference detection system for bag production

Abstract

The application is suitable for the color difference detection technical field in packaging bag production, and provides an automatic color difference detection system for packaging bag production, a film running state monitoring module constructs a film running state monitoring model according to a film vibration degree, a film tension degree and a film transverse drift degree, and outputs a film running stability index; a visual detection system health degree monitoring module constructs a visual detection system health degree monitoring model according to a light source power stability, an illumination uniformity and a camera sensor temperature, and outputs a visual system health degree index; and an adaptive conveying speed control module outputs a recommended conveying speed adjustment amount. The application introduces the film running state monitoring module and the visual detection system health degree monitoring module, respectively constructs the film running stability index and the visual system health degree index, realizes comprehensive and quantitative perception and evaluation of the key influencing factors, and overcomes the limitation of isolated problem in the prior art.

Description

Technical Field

[0001] This invention belongs to the field of color difference detection technology in packaging bag production, and particularly relates to an automated color difference detection system for packaging bag production. Background Technology

[0002] In the production of packaging bags, especially color-printed packaging bags, color difference is a key indicator for measuring product quality. Even slight color deviations not only affect the consistency of product appearance but can also damage brand image; therefore, online, real-time, automated color difference detection is crucial. Current machine vision-based inspection systems face multiple complex challenges in high-speed, continuous production environments, making it difficult to guarantee detection accuracy and stability.

[0003] The instability of the thin film's operating state constitutes a core obstacle. Existing systems generally lack a comprehensive quantitative monitoring mechanism for vibration amplitude and tension fluctuation range, making it impossible to construct dynamic evaluation models and quantify the negative impact of these state changes on color information extraction in real time. The degradation problem of the visual inspection system itself is also prominent. Power attenuation or short-term fluctuations of the light source can change the illumination spectral distribution characteristics, causing color measurements to drift in the CIE color space; the decrease in illumination uniformity caused by long-term use of LED light sources results in local brightness differences in the image, leading to significant discrepancies in color evaluation results for different areas of the same thin film; the camera sensor's temperature rises during continuous operation, increasing thermal noise levels and deteriorating the image signal-to-noise ratio, especially under low-light conditions where color difference detection errors are drastically amplified.

[0004] The lack of adaptive dynamic adjustment capabilities in the system exacerbates the aforementioned problems. Traditional systems, to ensure detection reliability under worst-case conditions, are forced to adopt fixed and conservative conveyor speeds, significantly reducing production efficiency. While some systems introduce simple feedback speed control mechanisms, the adjustment logic is rigid, relying solely on single indicators such as historical false detection rates. They fail to integrate multi-dimensional status information, such as thin-film operational stability and vision system health, for collaborative decision-making, making it impossible to achieve a dynamic optimal balance between speed, accuracy, and system reliability. This results in a trade-off between production efficiency and detection quality. Therefore, existing technologies urgently need improvement to address these issues. Summary of the Invention

[0005] The purpose of this invention is to provide an automated color difference detection system for packaging bag production, aiming to solve the above-mentioned problems.

[0006] This invention is implemented as follows: an automated color difference detection system for packaging bag production, comprising: The thin film operation status monitoring module is used to construct a thin film operation status monitoring model based on the degree of thin film vibration, thin film tension and thin film lateral drift, and output the thin film operation stability index. The visual inspection system health monitoring module is used to construct a visual inspection system health monitoring model based on light source power stability, illumination uniformity, and camera sensor temperature, and outputs a visual system health index. The color difference detection confidence assessment module uses ambient temperature as an influencing factor to construct a color difference detection confidence assessment model based on the film operation stability index and the visual system health index, and outputs the real-time detection confidence index. The adaptive transport speed control module is used to construct an adaptive transport speed control model based on the real-time detection confidence index, the film operation stability index and the current film transport speed, and output a recommended transport speed adjustment amount, and periodically adjust the film transport speed according to the recommended transport speed adjustment amount.

[0007] A further technical solution is to subtract the vibration value under ideal conditions from the quantized value of the real-time vibration amplitude, and then divide the result by (the upper limit of vibration allowed by the system minus the vibration value under ideal conditions) to obtain the quotient. Subtracting the quotient from 1 yields the membrane vibration degree index. The difference between the real-time film tension value and the tension value under ideal conditions is taken as the absolute value and then divided by the allowable deviation range of the film tension value to obtain the quotient. The film tension index is obtained by subtracting the quotient from 1. The quotient is obtained by dividing the real-time lateral drift of the thin film by the maximum allowable offset threshold of the system. The lateral drift index of the thin film is obtained by subtracting the quotient from 1.

[0008] Further technical solutions, in the thin film operating status monitoring model: The membrane operation stability index is calculated based on the membrane vibration degree index, membrane tension index, and membrane lateral drift index. The calculation method is as follows: first, take the minimum value among the three as the basic short board value, and then perform exponential weighting on it. Then, based on the coupling penalty coefficient, the square root of the sum of squares of the deviations of the membrane vibration degree index, membrane tension index, and membrane lateral drift index from the ideal value "1" is calculated. The square root is multiplied by the preset coupling penalty coefficient to obtain the penalty term. The penalty term is then subtracted from 1, and the result is multiplied by the aforementioned weighted short board value to obtain the final membrane operation stability index.

[0009] A further technical solution is to subtract the ideal volatility from the light source power volatility, and then divide the result by (the difference between the upper limit of volatility and the ideal volatility). Subtracting this result from 1 yields the light source power stability index. The light uniformity index is obtained by taking the absolute value of the difference between the real-time light uniformity and the ideal uniformity, and then dividing it by (the difference between the upper limit of the allowable light uniformity and the ideal uniformity). The difference between the real-time camera sensor temperature value and the optimal camera sensor temperature value is taken as the absolute value and then divided by (the maximum allowable camera sensor temperature value minus the optimal camera sensor temperature value) to obtain a result. Subtracting this result from 1 yields the camera sensor temperature index.

[0010] Further technical solutions, in the health monitoring model of the visual inspection system: The visual system health index is obtained by multiplying the light source power stability index, the illumination uniformity index, and the camera sensor temperature index by an exponential decay factor. The calculation method of the exponential attenuation factor is as follows: with the natural constant e as the base, the sum of the squares of the deviations of the light source power stability index, the illumination uniformity index and the camera sensor temperature index from the ideal value "1" is taken as a whole, multiplied by a preset synergistic degradation amplification factor, and the negative number is taken, and then exponential calculation is performed to obtain the final attenuation factor.

[0011] A further technical solution involves subtracting the optimal ambient temperature from the real-time ambient temperature value, taking the absolute value, and then dividing by (the maximum allowable ambient temperature value minus the optimal ambient temperature value) to obtain a result. Subtracting this result from 1 yields the ambient temperature index. In the color difference detection confidence assessment model, the calculation method for the real-time detection confidence index is as follows: First, the arithmetic mean of the thin film operation stability index and the vision system health index is calculated. Then, an adjustment factor is calculated based on the ambient temperature index and the preset environmental vulnerability coefficient. This adjustment factor is "1" minus a correction term consisting of the environmental vulnerability coefficient, the deviation of the ambient temperature index from the ideal value "1", and the deviation of the aforementioned arithmetic mean from the ideal value "1". Finally, the aforementioned arithmetic mean is multiplied by the adjustment factor to obtain the final real-time detection confidence index.

[0012] A further technical solution involves subtracting the minimum allowable conveying speed from the current film conveying speed, and then dividing by the allowable conveying speed range to obtain the current conveying speed index. In the adaptive conveying speed control model, the recommended conveying speed adjustment is calculated as follows: The adjustment period is multiplied by the system's allowable speed range as the basic scaling factor; based on the difference between the real-time detection confidence index and the target confidence, the first adjustment component is calculated using the hyperbolic tangent function, which is controlled by a preset adjustment intensity coefficient and a preset confidence deviation sensitivity. The second adjustment component is calculated based on the difference between the thin film operation stability index and the stability speed regulation threshold; the first adjustment component is multiplied by the second adjustment component to obtain the active adjustment term related to the confidence level. At the same time, subtract a reference value obtained by power mapping of the thin film operation stability index from the current transmission speed index, and then multiply by the preset speed inertia damping coefficient to obtain the speed inertia damping term that suppresses excessive speed change. Finally, the active adjustment term is subtracted from the inertial damping term, and multiplied by the aforementioned basic scaling factor to obtain the final recommended transmission speed adjustment.

[0013] Further technical solutions also include a thermal management optimization control module, which is used to construct a heat dissipation fan control model based on the sensor target temperature, the sensor actual temperature, the real-time detection confidence index, and the ambient temperature, and output a cooling fan power adjustment value to optimize the heat dissipation efficiency of the detection system based on the cooling fan power adjustment value.

[0014] A further technical solution involves substituting the actual sensor temperature and the target sensor temperature into the maximum-minimum normalization formula for processing, and obtaining the actual sensor temperature index and the target sensor temperature index respectively; in the heat dissipation fan control model, the calculation method for the cooling fan power regulation value is as follows: The deviation of the ambient temperature index from the ideal value "1" is used as the basic cooling requirement; the portion of the actual temperature index of the sensor that exceeds the target temperature index is divided by the deviation of the target temperature index from the ideal value "1", and this result is multiplied by the real-time detection confidence index as the overheating positive feedback adjustment term. The portion of the sensor's target temperature index that exceeds the actual temperature index is divided by the target temperature index, and this result is multiplied by the real-time detection confidence index and the energy-saving adjustment coefficient to form an energy-saving suppression adjustment term. Add the basic cooling demand item and the overheating positive feedback adjustment item, and then subtract the energy saving suppression adjustment item to obtain the preliminary control value; finally, limit the preliminary control value within the range of [0, 1] to obtain the final cooling fan power control value.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: This application introduces a thin film operation status monitoring module and a visual inspection system health monitoring module, and constructs a thin film operation stability index and a visual system health index respectively, realizing a comprehensive and quantitative perception and evaluation of these key influencing factors, overcoming the limitations of viewing problems in isolation in the prior art.

[0016] The color difference detection confidence assessment module in this application considers ambient temperature as an important influencing factor and integrates it with the thin film operation stability index and the visual system health index to construct a color difference detection confidence assessment model and output a real-time detection confidence index. This multi-dimensional information fusion assessment enables the system to more accurately reflect the reliability of the current detection results and makes up for the shortcomings of existing technologies in considering external interference factors.

[0017] The adaptive conveyor speed control module of this application constructs an adaptive conveyor speed control model based on the real-time detection confidence index, the film operation stability index, and the current film conveyor speed, and outputs a recommended conveyor speed adjustment amount to achieve periodic adaptive adjustment of the film conveyor speed. This intelligent decision-making and dynamic adjustment capability based on multi-dimensional state information enables the system to dynamically optimize production efficiency while ensuring detection accuracy, avoiding the inefficiency of traditional fixed-speed operation and surpassing the limitations of simple feedback speed regulation, achieving a dynamic optimal balance between speed, accuracy, and system reliability. Therefore, the overall technical solution of this application provides a more robust, efficient, and intelligent automated color difference detection solution.

[0018] This application overcomes the problem that traditional heat dissipation control strategies fail to fully consider the precise difference between the actual and target sensor temperatures, changes in the real-time detection confidence index, and the combined effects of ambient temperature fluctuations. Specifically, by incorporating the precise deviation between the actual and target sensor temperatures, heat dissipation adjustment becomes more accurate, avoiding the adjustment errors caused by coarse control based solely on a single temperature threshold. Simultaneously, by introducing the real-time detection confidence index as an adjustment factor, the heat dissipation strategy can be closely linked to the reliability requirements of the current detection task. When the detection confidence is high, the system can appropriately reduce the heat dissipation intensity to save energy, while when the detection confidence is low, it will increase heat dissipation to ensure the sensor is in optimal working condition, thereby improving detection accuracy. Furthermore, the inclusion of ambient temperature allows the heat dissipation system to dynamically adapt to changes in the external environment, avoiding insufficient or excessive heat dissipation due to ambient temperature fluctuations. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of an automated color difference detection system used in packaging bag production. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0021] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0022] like Figure 1 As shown, an automated color difference detection system for packaging bag production, provided in one embodiment of the present invention, includes: The thin film operation status monitoring module is used to construct a thin film operation status monitoring model based on the degree of thin film vibration, thin film tension and thin film lateral drift, and output the thin film operation stability index. The visual inspection system health monitoring module is used to construct a visual inspection system health monitoring model based on light source power stability, illumination uniformity, and camera sensor temperature, and outputs a visual system health index. The color difference detection confidence assessment module uses ambient temperature as an influencing factor to construct a color difference detection confidence assessment model based on the film operation stability index and the visual system health index, and outputs the real-time detection confidence index. The adaptive transport speed control module is used to construct an adaptive transport speed control model based on the real-time detection confidence index, the film operation stability index and the current film transport speed, and output a recommended transport speed adjustment amount, and periodically adjust the film transport speed according to the recommended transport speed adjustment amount.

[0023] In this embodiment, an automated color difference detection system is used. This system is designed to automatically and in real-time detect the film color during the packaging bag production process to ensure product color consistency and identify any deviations from the standard color sample. Its core function is to improve the accuracy, stability, and production efficiency of detection by integrating multiple monitoring and control mechanisms.

[0024] The film operation status monitoring module is responsible for evaluating the mechanical motion stability of the packaging film during transport. It constructs a comprehensive model by collecting data on the film's vibration level, tension, and lateral drift, and outputs a quantified film operation stability index. This index reflects the potential impact of the film's physical state in the detection area on image acquisition. The film operation stability index is a comprehensive quantitative value, typically between 0 and 1, used to represent the stability of the film's operating state. A higher index value indicates more stable film operation and less interference with color difference detection; a lower index value indicates more unstable film operation and a greater negative impact on color difference detection.

[0025] The visual inspection system health monitoring module is used to evaluate the operational status of the visual inspection unit in an automated color difference detection system. It constructs a health model by monitoring key parameters such as light source power stability, illumination uniformity, and camera sensor temperature, and outputs a quantified visual system health index. This index reflects the impact of the visual system's performance on inspection quality. The visual system health index is a comprehensive quantitative value, typically between 0 and 1, used to represent the overall health of the visual inspection system. A higher index value indicates better visual system performance and stronger support for color difference detection; a lower index value indicates decreased visual system performance, which may affect the accuracy of color difference detection.

[0026] Color Difference Detection Confidence Assessment Module: This module is responsible for comprehensively assessing the reliability of the current color difference detection results. It uses ambient temperature as an external influencing factor and combines it with the film operation stability index and the visual system health index to construct a confidence assessment model, ultimately outputting a real-time detection confidence index. The real-time detection confidence index is a comprehensive quantitative value, typically between 0 and 1, used to represent the reliability of the current color difference detection results. A higher index value indicates a more reliable detection result; a lower index value indicates lower reliability, potentially requiring intervention.

[0027] The adaptive conveyor speed control module constructs an adaptive control model based on the real-time detection confidence index, the film operation stability index, and the current film conveyor speed. It outputs a recommended conveyor speed adjustment and periodically adjusts the film conveyor speed accordingly to optimize production efficiency while ensuring detection accuracy. The recommended conveyor speed adjustment indicates the magnitude and direction (increase or decrease) of the suggested adjustment to the current film conveyor speed. A positive value indicates a suggested acceleration, a negative value indicates a suggested deceleration, and a zero value indicates maintaining the current speed.

[0028] In a preferred embodiment of the present invention, the difference between the quantized value of the real-time vibration amplitude and the vibration value under ideal conditions is divided by (the upper limit of vibration allowed by the system minus the vibration value under ideal conditions) to obtain a quotient result. 1 is subtracted from the quotient result to obtain the film vibration degree index. The difference between the real-time film tension value and the tension value under ideal conditions is taken as the absolute value and then divided by the allowable deviation range of the film tension value to obtain the quotient. The film tension index is obtained by subtracting the quotient from 1. The quotient is obtained by dividing the real-time lateral drift of the thin film by the maximum allowable offset threshold of the system. The lateral drift index of the thin film is obtained by subtracting the quotient from 1.

[0029] In this embodiment, the technical approach aims to transform the real-time vibration state of the thin film into a standardized index. The quantified value of the real-time vibration amplitude can be obtained in various ways. For example, a laser displacement sensor can be used to perform non-contact measurement on the thin film surface to obtain its vertical displacement change; alternatively, a high frame rate camera combined with image processing algorithms can be used to quantify the vibration amplitude by analyzing the pixel displacement of the thin film edge or specific marker points. The ideal vibration value is typically the baseline vibration level of the system under stable operation without external interference, which can be obtained through system calibration, historical data analysis, or theoretical calculation. The upper limit of the system's permissible vibration is determined by equipment design specifications, thin film material properties, or manufacturing process requirements, representing the maximum tolerable vibration amplitude of the system. Through this normalization process, the thin film vibration index can objectively reflect the degree of deviation of the current vibration level from the system's tolerance.

[0030] This technique is used to quantify the tension of a film. Real-time film tension can be monitored by tension sensors installed along the film transport path, such as piezoelectric or strain gauge tension sensors. These sensors convert the pressure of the film on the guide rollers or the deformation of the guide rollers into an electrical signal, thereby measuring the tension. Ideally, the tension value is the optimal value set based on the film material, thickness, and printing process requirements. The permissible deviation range of the film tension value defines the acceptable range within which the film tension fluctuates around the ideal value. By calculating the absolute deviation between the real-time tension value and the ideal value and comparing it with the permissible deviation range, a standardized index reflecting the degree of tension deviation can be obtained, ensuring the comprehensiveness and accuracy of the tension assessment, regardless of whether the tension is too high or too low.

[0031] This technique aims to quantify the lateral positional stability of the film during transport. Real-time lateral drift can be measured using various sensors. For example, photoelectric edge detection sensors or ultrasonic sensors can be used to monitor the lateral position of the film edge relative to a preset baseline in real time; alternatively, a machine vision system can be employed to perform image recognition and positioning of the film edge, calculating its offset relative to a reference position. The maximum allowable offset threshold is determined based on factors such as the printing precision of the packaging bag, cutting requirements, and the mechanical precision of the equipment, representing the maximum acceptable range of lateral drift. By comparing the real-time drift with the maximum offset threshold, an index that intuitively reflects the lateral positional stability of the film can be obtained.

[0032] This application's solution addresses the inaccuracy of traditional methods by refining and standardizing key physical quantities in thin film operation, namely vibration amplitude, tension value, and lateral drift. These indices share the common characteristic of comparing actual measured values ​​with ideal values ​​or allowable ranges, and mapping the degree of deviation of physical quantities to an index range of 0 to 1 using a "1 minus quotient" method. Here, 1 represents the optimal or most stable state, while smaller values ​​indicate greater instability. This unified quantification method provides accurate and comparable input to the thin film operation monitoring module, making the calculation of the thin film operation stability index more reliable, thereby improving the accuracy and robustness of the entire automated color difference detection system.

[0033] In a preferred embodiment of the present invention, the thin film operating status monitoring model includes: ; in The coefficient representing the shortest-board effect. ; The coupling penalty coefficient is... ; This is an index of the degree of membrane vibration. The film tension index. This is the lateral drift index of the thin film. This is the stability index for thin film operation.

[0034] In this embodiment, the thin film operating status monitoring model is a mathematical expression, which serves to index the degree of thin film vibration. Thin film tension index and the lateral drift index of the thin film These three key indicators are integrated into a single, quantitative index for thin film operational stability. This allows for a more comprehensive and accurate reflection of the dynamic operating status of the thin film. The model can be deployed in a dedicated hardware computing unit, such as an embedded processor or field-programmable gate array, or run as a software module on an industrial control computer.

[0035] Shortest board effect coefficient This is used to quantify the impact of the "bottleneck effect" in thin film operation, that is, among multiple indicators, the worst-performing indicator has the greatest impact on overall stability. When When the value is close to 1, the effect of the weakest link effect is relatively weak; when When the value is small, the impact of the weakest link effect is more significant, meaning the worst-case indicator has a greater impact on the overall stability index. The lowering effect is stronger. This coefficient can be set by expert experience, for example, by preset according to the characteristics of different film materials or production lines; or by optimization and adjustment through historical data analysis and machine learning algorithms, so that the stability index output by the model is more consistent with the occurrence rate of problems in actual production.

[0036] Coupling penalty coefficient This is used to quantify the degree of penalty for "coupling degradation" between thin film performance indicators. Coupling degradation refers to the situation where, when multiple thin film performance indicators simultaneously deviate from ideal conditions, their negative impact on overall stability is not simply additive, but may produce a more severe synergistic deterioration. A value of 0 indicates that coupling penalty is not considered; when A value greater than 0 indicates the presence of a coupling penalty, and The higher the value, the stronger the penalty effect; that is, when multiple indicators deteriorate simultaneously, the thin film operation stability index... It will decrease more significantly. This coefficient can be initially set based on production experience; for example, a higher value can be set for production processes that are sensitive to multiple factors. Values; or by analyzing different failure modes and combining them with actual production data for calibration and optimization.

[0037] The film vibration index, film tension index, and film lateral drift index are quantitative indicators of the film's operating status, typically ranging from 0 to 1, where 1 represents the ideal or optimal state and 0 represents the worst or unacceptable state. These indices serve as inputs to the film operating status monitoring model, reflecting the film's real-time performance in terms of vibration, tension, and lateral position. They are obtained by acquiring data in real time using sensors (such as vibration sensors, tension sensors, and displacement sensors) and performing preprocessing and normalization calculations. The film operating stability index, an output of the film operating status monitoring model, is a comprehensive quantitative indicator representing the overall stability of the current film operation. Its value typically ranges from 0 to 1, where 1 indicates a very stable operating state and 0 indicates an extremely unstable operating state. This index provides crucial input for subsequent confidence assessment of color difference detection and adaptive transport speed control.

[0038] The thin-film operating condition monitoring model of this application comprehensively evaluates three key indicators—thin-film vibration index, thin-film tension index, and thin-film lateral drift index—through a structured mathematical expression to output a thin-film operating stability index. This model first... This study captures the "bottleneck effect" in thin film operation, where the overall stability of the thin film is primarily limited by the worst-performing index. Based on this, a bottleneck effect coefficient is introduced. ,pass The item makes the worst-case indicator affect the overall stability index. The degree of impact can be adjusted according to actual needs, when When the value is small, the negative impact of the worst-case indicator on overall stability is amplified, thus more sensitively reflecting potential stability problems. Meanwhile, the model calculates... The term introduces a "coupling penalty mechanism". This mechanism calculates the Euclidean distance of each exponent from the ideal value (1) and multiplies it by the coupling penalty coefficient. This imposes an additional penalty when multiple indicators deteriorate simultaneously. This means that when any two or three of the membrane's vibration, tension, or lateral drift parameters perform poorly at the same time, its impact on the overall stability index is reduced. The negative impact of a single indicator deterioration will be more significant than when a single indicator worsens, thus more accurately reflecting the synergistic degradation risk of the thin film's operating condition. Ultimately, these two parts are combined through multiplication to ensure the thin film operating stability index. This comprehensive evaluation method can fully and accurately reflect the dynamic characteristics of thin film operation, providing more reliable basic data for subsequent confidence assessment of color difference detection. It overcomes the problem of inaccurate assessment of thin film operating conditions in traditional methods, enabling the system to identify thin film instability that may affect the accuracy of color difference detection earlier and more accurately.

[0039] In a preferred embodiment of the present invention, the difference between the light source power volatility and the ideal volatility is divided by (the difference between the upper limit of volatility and the ideal volatility) to obtain a result, and 1 is subtracted from this result to obtain the light source power stability index. The light uniformity index is obtained by taking the absolute value of the difference between the real-time light uniformity and the ideal uniformity, and then dividing it by (the difference between the upper limit of the allowable light uniformity and the ideal uniformity). The difference between the real-time camera sensor temperature value and the optimal camera sensor temperature value is taken as the absolute value and then divided by (the maximum allowable camera sensor temperature value minus the optimal camera sensor temperature value) to obtain a result. Subtracting this result from 1 yields the camera sensor temperature index.

[0040] In this embodiment, the light source power fluctuation rate refers to the degree to which the output power of the light source changes over time during operation. It is typically obtained by measuring the deviation between the instantaneous value and the average value of the light source's output intensity. The ideal fluctuation rate is the minimum fluctuation level expected of the light source, set during system design or calibration, and is usually close to zero. The upper limit of the fluctuation rate is the maximum fluctuation range that the system can tolerate without affecting the detection accuracy. Various methods can be used to obtain the light source power fluctuation rate. For example, a highly sensitive photodetector (such as a photodiode) can be integrated near the light source to continuously monitor its output intensity. The analog signal is converted to a digital signal by an analog-to-digital converter, and the processing unit calculates its standard deviation or peak-to-peak fluctuation within a specific time window, then divides it by the average intensity to obtain the fluctuation rate.

[0041] Real-time illumination uniformity refers to the uniformity of the illumination intensity distribution provided by the light source within the thin film detection area. Ideal uniformity is the level of uniformity of illumination distribution that the system expects to achieve under optimal operating conditions, typically referring to minimal variation in illumination intensity within the detection area. The upper limit of permissible illumination uniformity is the maximum non-uniformity that the system can accept without affecting the accuracy of color difference detection. To obtain real-time illumination uniformity, the following methods can be used. For example, multiple illumination sensors can be arranged within the detection area, and the illumination uniformity can be quantified by comparing the differences in readings from these sensors at different locations.

[0042] Real-time camera sensor temperature refers to the actual operating temperature of the camera's internal image sensor (such as a CCD or CMOS chip). The optimal camera sensor temperature is the ideal operating temperature, determined during the camera sensor's design or through experimentation, that provides the best imaging performance (such as lowest noise and highest sensitivity). The maximum permissible camera sensor temperature is the highest operating temperature the camera sensor can withstand without significant performance degradation or damage. There are several ways to obtain real-time camera sensor temperature values. For example, many industrial cameras integrate temperature sensors, allowing real-time sensor temperature data to be read directly through the camera's software development kit or communication interface. Another approach is to install a high-precision thermistor or thermocouple near the camera sensor chip, measure the temperature using external circuitry, and transmit the measurement results to the processing unit.

[0043] This application's solution compares three key parameters—light source power fluctuation rate, real-time illumination uniformity, and real-time camera sensor temperature—with their respective ideal values, allowable ranges, or upper limits, and then normalizes them to obtain the light source power stability index, illumination uniformity index, and camera sensor temperature index. This quantification method unifies the originally dispersed and physically different parameters into standardized indices between 0 and 1, where 1 represents the optimal state and 0 represents the worst state. These indices, as inputs to the visual inspection system's health monitoring module, can accurately and objectively reflect the operating status of each core component of the visual inspection system. In this way, the visual inspection system's health monitoring module can more accurately assess the overall "health" of the visual system, rather than simply looking at whether a single parameter exceeds a simple threshold. This refined quantification and modeling provides more reliable and detailed basic data for subsequent color difference detection confidence assessment and adaptive transmission speed control, thereby improving the intelligence level and robustness of the entire automated color difference detection system.

[0044] In a preferred embodiment of the present invention, the health monitoring model of the visual inspection system includes: ; in For the synergistic degradation amplification factor, , The power stability index of the light source. The light uniformity index. For camera sensor temperature index, This is an index of visual system health.

[0045] In this embodiment, the vision inspection system health monitoring model aims to provide a quantitative and comprehensive assessment of the system's operational status and reliability. It generates a comprehensive health score by integrating multiple key performance indicators. This model can be deployed as a standalone software module on an industrial PC or embedded controller, acquiring various indices through a real-time data interface and performing periodic calculations. Furthermore, the model can also be integrated into a cloud-based data analytics platform, leveraging big data processing capabilities to centrally model and assess the health of vision system data from multiple production lines, enabling remote monitoring and predictive maintenance. (Light source power stability index) Illumination uniformity index and camera sensor temperature index These are quantitative representations of key performance parameters of the vision system, typically ranging from 0 to 1, where 1 represents the ideal or optimal state. As the basic inputs to the model, they reflect the independent health status of the vision system in terms of lighting conditions and its own thermal management, serving as fundamental dimensions for assessing the health of the vision system and providing a data foundation for subsequent co-deterioration analysis. Co-deterioration amplification factor. This is a non-negative parameter used to adjust the penalty strength for the overall health assessment when multiple vision system health indicators simultaneously deviate from the ideal state. Its value can be determined through regression analysis or machine learning training on historical operating data to capture the actual performance degradation patterns of different systems under multi-factor degradation. Alternatively, it can be set based on vision system design specifications, component interdependency analysis, and expert experience. For example, for systems highly sensitive to temperature and light fluctuations, a larger value can be set. value.

[0046] Exponential function term This is one of the core innovations of this scheme, specifically designed to quantify and amplify the synergistic degradation effect among the components of the visual system. By calculating the sum of squares of the deviations of each independent index from the ideal value (1), this term can sensitively reflect the simultaneous deterioration of multiple factors. When any one or more indices deviate from 1, the value of this term will decrease significantly, and this decrease will increase with the degree of deviation and The increase in the value accelerates the process, thus effectively incorporating the negative impact of collaborative degradation on system health into consideration. Visual System Health Index It is the final output of the visual inspection system health monitoring model. It is a comprehensive and normalized indicator that fully reflects the current operating status, performance reliability and potential failure risks of the visual system. It not only considers the independent health status of the light source, illumination, and sensor temperature, but more importantly, it accurately reflects the impact of these factors on the overall system performance when they interact through a synergistic degradation amplification mechanism, providing a more reliable basis for subsequent decision-making.

[0047] This application's solution addresses the deficiency in visual system health assessment that neglects the synergistic degradation among factors, thereby improving detection accuracy. The model multiplies the light source power stability index, illumination uniformity index, and camera sensor temperature index, ensuring that degradation of any single factor directly reduces the visual system health index, avoiding the one-sidedness of isolated assessments. Simultaneously, an exponential function term is introduced, based on the sum of squared deviations of each index from its ideal value, multiplied by a synergistic degradation amplification factor. When multiple indices deteriorate simultaneously, this term significantly reduces the visual system health index, capturing the coupling effect between factors and making the assessment more closely reflect actual system performance changes. The setting of the synergistic degradation amplification factor allows for adjustment of the penalty intensity according to system requirements, enhancing the model's adaptability and control flexibility. This modeling approach not only integrates various health factors but also dynamically quantifies the impact of synergistic degradation, providing a more comprehensive and reliable basis for visual system health assessment, thus supporting improved confidence in color difference detection.

[0048] In a preferred embodiment of the present invention, the difference between the real-time ambient temperature value and the optimal ambient temperature value is taken as the absolute value and then divided by (the maximum allowable ambient temperature value minus the optimal ambient temperature value) to obtain a result. Subtracting this result from 1 yields the ambient temperature index; in the color difference detection confidence assessment model: ; in This represents the environmental vulnerability coefficient. , This is the thin film operating stability index. As a visual system health index, The ambient temperature index. To detect the confidence index in real time.

[0049] In this embodiment, this application solves the problem of inaccurate confidence assessment caused by insufficient modeling of environmental factors by introducing the calculation of the ambient temperature index and integrating it into the color difference detection confidence assessment model. The calculation of the ambient temperature index involves comparing the real-time ambient temperature value with the optimal ambient temperature value and normalizing the deviation. This index can be obtained by real-time acquisition of ambient temperature data by temperature sensors installed around the detection system, and by executing preset calculation logic through a microcontroller or industrial computer. Alternatively, the system can preset multiple ambient temperature ranges, each corresponding to an ambient temperature index value. When the real-time ambient temperature falls into a certain range, the corresponding index value is directly assigned, thereby simplifying the calculation and providing a segmented temperature impact assessment.

[0050] The color difference detection confidence assessment model is a mathematical model that incorporates the thin film operational stability index. Visual system health index And the newly introduced ambient temperature index In summary, the real-time detection confidence index is calculated. The model can be deployed in the system's central processing unit or a dedicated evaluation module, receiving input data from the thin-film operating status monitoring module, the visual inspection system health monitoring module, and the ambient temperature sensor. The environmental vulnerability coefficient in the model... It is an adjustable parameter that can be configured according to the sensitivity of the actual production environment to temperature changes. For example, in a production line that is highly sensitive to temperature changes, a higher setting can be used. The value amplifies the impact of ambient temperature. Furthermore, the model can be optimized using machine learning algorithms, training model parameters with historical data to more accurately reflect the influence of various factors on the confidence level under different environmental conditions.

[0051] The solution presented in this application, through the aforementioned technical means, enables the color difference detection confidence assessment module to more comprehensively consider multiple factors affecting detection accuracy. This is based on the fundamental thin film operational stability index. and visual system health index Based on this, this application further introduces the ambient temperature index. This is cleverly integrated into the confidence assessment model. The model calculates... and The average value, combined with the ambient temperature index The degree of deviation is dynamically adjusted to adjust the real-time detection confidence index. When the ambient temperature deviates from the optimal value, the ambient temperature index... The confidence index will decrease, and the penalty term in the model will increase, thus affecting the real-time detection confidence index. The corresponding decrease is observed. This mechanism ensures that the system accurately reflects the reduction in detection reliability under adverse environmental conditions, avoiding inaccurate confidence assessments caused by unquantified environmental factors. In this way, the color difference detection confidence assessment module can provide a more robust and accurate confidence assessment result, providing a more reliable decision-making basis for the subsequent adaptive transmission speed control module.

[0052] In a preferred embodiment of the present invention, the difference between the current film conveying speed and the minimum allowable conveying speed is subtracted, and then divided by the allowable conveying speed range to obtain the current conveying speed index; in the adaptive conveying speed control model: ; in To adjust the strength coefficient, , For confidence bias sensitivity, , The stability speed regulation threshold, , The velocity-inertia damping coefficient, , For stability-velocity reference mapping coefficients, ; To detect the confidence index in real time. For the preset target confidence level, This is the thin film operating stability index. The stability speed regulation threshold, This represents the current transmission speed index. To adjust the cycle, The minimum permissible speed for the system. The maximum permissible speed of the system. Recommended transmission speed adjustment amount.

[0053] In this embodiment, the difference between the current film conveying speed and the minimum allowable conveying speed is divided by the allowable conveying speed range to obtain the current conveying speed index. This process aims to convert the actual physical speed value into a dimensionless normalized index. This index is typically between 0 and 1, allowing physical quantities with different dimensions to be compared and calculated within a unified mathematical model, thereby avoiding calculation errors caused by differences in units or magnitudes. This normalization process can be performed by the system's data processing unit, for example, by a microcontroller or industrial computer receiving the raw speed signal from a speed sensor and then performing preset mathematical operations to generate the current conveying speed index. Alternatively, it can be implemented by configuring smart sensors or programmable logic controllers, which can directly output the calculated and normalized speed index.

[0054] The adaptive conveyor speed control model is a core mathematical model that dynamically calculates the recommended adjustment amount for the thin film conveyor speed based on multiple real-time input parameters. This model optimizes system performance by comprehensively considering detection quality, thin film operating status, and current speed through complex nonlinear functional relationships. This model can be integrated as a software module into the system's central control unit, for example, running on an embedded system, industrial PC, or digital signal processor, and performing calculations periodically.

[0055] Adjusting intensity coefficient It is a positive value used to control the overall magnitude of speed regulation. This coefficient determines the aggressiveness of the system's response to the regulation demand calculated by the model. A larger value... Values ​​that result in more significant velocity changes, while smaller values... This value allows for smoother speed adjustments. This coefficient is typically stored as a system configuration parameter in non-volatile memory and can be adjusted by the operator via a human-machine interface to suit different production needs or material properties.

[0056] Confidence bias sensitivity It is also a positive value, which quantifies the system's confidence index in real-time detection. confidence level of the preset target The intensity of the reaction to the deviation between them. Higher This value means that even a small confidence level deviation can cause a large speed adjustment response, making the system more sensitive to fluctuations in detection quality. This parameter can be optimized during system debugging and calibration to achieve a balance between response speed and system stability.

[0057] Stability speed regulation threshold It is a value between 0 and 1, which sets a thin film operating stability index. The lowest acceptable level. When the thin film operating stability index Below this threshold, the system will limit or prevent further speed increases to ensure priority is given to detection quality when the film operation is unstable. This threshold can be preset based on empirical data or production process requirements, or it can be dynamically adjusted based on historical operating data using an adaptive algorithm.

[0058] Velocity inertia damping coefficient This is a non-negative value whose function is to introduce a damping effect during speed regulation to smooth speed changes and prevent overshoot or oscillation in the system. This coefficient helps ensure a smooth transition in film conveying speed, reduces mechanical shock, thereby extending equipment life and improving production line stability. This coefficient can be determined during the design phase based on the mechanical characteristics of the conveying system, or optimized through experiments.

[0059] Stability-velocity reference mapping coefficient It is a positive value used to establish the thin film operating stability index. The relationship between this coefficient and an ideal velocity reference. This coefficient allows the system to dynamically adjust its "expectation" for the current velocity based on the current thin-film stability, thus providing a dynamic reference point for the velocity-inertial damping term. This coefficient can be determined by analyzing and modeling the system's operating data at different stability conditions.

[0060] Real-time detection confidence index This is the output of the color difference detection confidence assessment module, which comprehensively reflects the impact of film operation stability, vision system health, and ambient temperature on the reliability of detection results. This index provides a quantitative assessment of the current detection quality for the speed control model.

[0061] Preset target confidence level It is the detection confidence level that the system expects to achieve, serving as a quality target for speed adjustment.

[0062] Thin film operating stability index This is the output of the thin film operation status monitoring module, which quantifies the comprehensive state of the thin film during transport, including vibration, tension fluctuations, and lateral drift. This index provides real-time feedback on the physical operating state of the thin film for the speed control model.

[0063] Current transmission speed index It is the normalized current film transport speed, which serves as the input to the speed control model and reflects the current operating speed state of the system.

[0064] Adjustment cycle The time interval between the speed regulation model performing calculations and applying regulation values ​​is defined. Shorter regulation intervals can make the system respond faster, but increase the computational load; longer intervals have the opposite effect.

[0065] System minimum allowable speed and the system's maximum allowed speed These are the lower and upper limits of the physical or technologically permissible transmission speed of the system, used to define the range of permissible transmission speed values.

[0066] Recommended transmission speed adjustment It is the final output of the adaptive transport speed control model, representing the specific amount of change that should be applied to the film transport speed within the current adjustment cycle.

[0067] The proposed solution normalizes the current film transport speed, enabling it to be calculated collaboratively with other key indices on a unified scale, thus avoiding adjustment deviations caused by differences in speed ranges. In the adaptive transport speed control model, this solution cleverly integrates several key factors. First, the `tanh` function is used to process the real-time detection confidence index. confidence level of the preset target The deviation between them, combined with the adjustment intensity coefficient and confidence bias sensitivity A regulation component based on confidence level deviation is generated. The introduction of the `tanh` function ensures the smoothness of the speed regulation response, effectively avoiding sudden speed changes caused by small fluctuations in confidence level, thereby improving the system's operational stability. Secondly, this confidence level regulation component is correlated with the thin film operational stability index. and stability speed regulation threshold The difference is multiplied, ensuring that the system only significantly increases speed when the film's operational stability is above a preset threshold. This prioritizes detection quality when the film's operating condition is poor, avoiding blind acceleration under unstable conditions that could compromise detection reliability. Furthermore, the model includes a damping term, which is exponentially based on the current transmission speed. With the thin film operation stability index Through stability-velocity reference mapping coefficients The deviation between the derived reference velocities is multiplied by the velocity inertia damping coefficient. This design fully considers the correlation between the current velocity state and the film stability, providing the necessary damping effect to effectively reduce oscillations during velocity regulation and ensure the smoothness and continuity of velocity changes. The entire model adjusts the period... and the allowed transmission speed range value The calculation of the relative adjustment is scaled up, converted into actual speed units, and periodically applied to adjust the film conveying speed.

[0068] As a preferred embodiment of the present invention, it also includes a thermal management optimization control module, which is used to construct a heat dissipation wind power control model based on the sensor target temperature, the sensor actual temperature, the real-time detection confidence index and the ambient temperature, and output a cooling fan power adjustment value, and optimize the heat dissipation efficiency of the detection system based on the cooling fan power adjustment value.

[0069] In this embodiment, the thermal management optimization control module is a unit responsible for monitoring, analyzing, and adjusting the system's heat dissipation performance. This module can be an independent hardware control unit with a built-in microprocessor and control algorithm, acquiring data through a sensor interface and controlling the cooling equipment through an actuator interface.

[0070] The target temperature of a sensor refers to a specific value within the ideal or optimal temperature range that a key sensor in a vision inspection system should maintain under normal operating conditions. This value can be determined and preset in the control system based on the specifications provided by the sensor manufacturer or through experimental testing. It can also be dynamically adjusted according to the production environment, film type, or detection accuracy requirements.

[0071] The actual temperature of the sensor refers to the temperature value actually measured by the temperature sensor during real-time operation of the key sensor in the vision inspection system. It can be directly measured and fed back to the control system by physical sensors such as thermistors, thermocouples or integrated temperature sensors, or the sensor surface temperature can be measured non-contactly by infrared thermometers.

[0072] The real-time detection confidence index is an indicator that quantifies the reliability of the current color difference detection results and reflects the performance level of the detection system under the current operating conditions. This index is derived by the color difference detection confidence assessment module based on a comprehensive evaluation of factors such as the film operation stability index, the vision system health index, and the ambient temperature.

[0073] Ambient temperature refers to the air temperature around the packaging bag production workshop or testing system. It can be obtained in real time through an ambient temperature sensor installed near the testing system, or through the environmental monitoring data interface provided by the workshop's central air conditioning system.

[0074] A cooling fan power control model is a mathematical model or algorithm used to calculate the required cooling fan power regulation value based on input parameters (sensor target temperature, sensor actual temperature, real-time detection confidence index, and ambient temperature). This model can be an intelligent control model built based on advanced control theories such as fuzzy logic, PID control, or neural networks, or a parameterized model based on a preset lookup table or empirical formula. The cooling fan power regulation value refers to the control signal or value used to adjust the cooling fan speed or output power to achieve precise control of heat dissipation. It can be a PWM (Pulse Width Modulation) signal, controlling the fan speed by adjusting the duty cycle, or an analog voltage signal or digital signal, directly controlling the power output of the variable frequency fan. Optimizing the heat dissipation efficiency of the detection system means precisely controlling the cooling fan power so that the detection system (especially key sensors) can maintain its optimal operating temperature range in the most energy-efficient way while ensuring detection accuracy. This involves dynamically adjusting the heat dissipation intensity to avoid energy waste caused by excessive cooling, while providing sufficient cooling when necessary to prevent overheating. Furthermore, by linking heat dissipation control to the confidence level of the detection task, "on-demand cooling" is achieved.

[0075] The proposed solution utilizes a thermal management optimization control module to construct a heat dissipation fan control model based on the sensor's target temperature, actual sensor temperature, real-time detection confidence index, and ambient temperature. This model outputs a cooling fan power adjustment value, which optimizes the heat dissipation efficiency of the detection system. Specifically, the thermal management optimization control module receives inputs from the sensor's target temperature, actual sensor temperature, real-time detection confidence index, and ambient temperature. The comparison between the sensor's target temperature and actual sensor temperature directly reflects whether the sensor is in an ideal operating state and what level of temperature adjustment is needed. The real-time detection confidence index provides quantitative information on the importance and reliability of the current system's detection task. When the confidence level is low, the system may require stricter temperature control to improve detection accuracy; when the confidence level is high, the heat dissipation intensity can be appropriately reduced to save energy while ensuring accuracy. Ambient temperature, as an external interference factor, directly affects the efficiency of the heat dissipation system and the sensor's thermal load. Incorporating it into the model makes the heat dissipation strategy more adaptable. These input parameters are fed into the heat dissipation fan control model. This model comprehensively considers sensor temperature deviation, detection task requirements, and external environmental influences, generating a cooling fan power adjustment value through complex calculation logic. This control value is a precise, dynamic output; it's no longer a simple on / off signal, but a continuous or multi-level signal that can finely adjust fan speed or power. Subsequently, based on this cooling fan power control value, the thermal management optimization control module directly controls the cooling fan's operation, thereby optimizing the heat dissipation efficiency of the detection system. For example, when the actual sensor temperature is higher than the sensor target temperature and the real-time detection confidence index is low, the model will output a higher cooling fan power control value to quickly reduce the sensor temperature and ensure detection accuracy. Conversely, when the actual sensor temperature is close to the sensor target temperature and the real-time detection confidence index is high, the model may output a lower cooling fan power control value to maintain temperature stability and save energy. This solution deeply couples thermal management with the confidence level of the core detection task and environmental factors, making heat dissipation control no longer isolated but an integral part of the intelligent operation of the entire automated color difference detection system. This, together with the thin film operation status monitoring module, the visual inspection system health monitoring module, and the color difference detection confidence assessment module, forms a comprehensive, adaptive intelligent control closed loop.

[0076] When the real-time detection confidence index output by the color difference detection confidence assessment module is low, in addition to potentially triggering the adaptive transmission speed control module to adjust the transmission speed, the thermal management optimization control module will also prompt the heat dissipation system to work more actively to eliminate the negative impact of temperature on detection accuracy. This multi-dimensional approach ensures and improves the stability and reliability of the entire color difference detection system. This interconnected mechanism allows the system to achieve a dynamic balance between energy consumption and performance output based on actual needs, while maintaining detection accuracy, significantly improving the system's intelligence and operational efficiency.

[0077] In a preferred embodiment of the present invention, the actual temperature of the sensor and the target temperature of the sensor are respectively substituted into the maximum-minimum normalization formula for processing, and the actual temperature index and the target temperature index of the sensor are obtained respectively; in the heat dissipation wind power control model: .

[0078] in, } indicates that the result is restricted to the range [0,1]. ; This is the energy-saving adjustment coefficient. , To enable real-time detection of the confidence index. The ambient temperature index. The preset target temperature index for the sensor. This refers to the actual temperature index of the sensor. This is the power control value for the cooling fan.

[0079] In this embodiment, This indicates that the positive value of the expression within the parentheses is taken, and if the expression is negative, it is taken as 0. This is used to distinguish the direction of temperature deviation. This is the energy-saving adjustment coefficient. Used to suppress unnecessary heat dissipation when the system is running well, thereby optimizing energy efficiency; A fixed value can be preset by engineers or process engineers based on factors such as system design specifications, equipment heat dissipation capacity, and production environment characteristics. This value can also be dynamically adjusted based on factors such as the system's current operating load, ambient temperature trends, and the importance of the testing task. Values, or by collecting system operation data (such as temperature, confidence level, energy consumption, etc.), can be used to train models for optimization. The value of ; Basic cooling requirements This reflects the direct impact of ambient temperature on heat dissipation requirements. Ambient Temperature Index The closer the value is to 1 (ideal state), the smaller this value is, and the lower the basic heat dissipation requirement; conversely, the greater the deviation of the ambient temperature from the ideal value, the stronger the heat dissipation requirement. This design enables the heat dissipation system to dynamically respond to environmental changes, enhancing environmental adaptability.

[0080] Overheating positive feedback regulation term This option is activated when the actual sensor temperature is higher than the target temperature. (Molecular part) Indicates temperature deviation, denominator Used for normalization to adapt the adjustment range to the target temperature setting. This item is also related to the real-time detection confidence index. Multiplication means that when the detection confidence is high, the system is more sensitive to temperature deviations and more proactive in heat dissipation regulation to ensure detection stability at high confidence levels.

[0081] Energy saving suppression regulation item This option is activated when the actual sensor temperature is lower than the target temperature. (Molecular) Indicates temperature margin, denominator Used for normalization. This item is also related to... Multiply, and combine with the energy-saving adjustment coefficient. This enables energy-saving adjustment based on detection confidence: when the detection confidence is high and the temperature is below the target value, the system can appropriately reduce the heat dissipation intensity, save energy, and avoid overcooling.

[0082] Finally passed The function restricts the sum of the three terms to a certain value. Within the specified range, ensure that the output fan power control value is always within a controllable range to avoid control failure due to calculation anomalies.

[0083] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An automated color difference detection system for packaging bag production, characterized in that, include: The thin film operation status monitoring module is used to construct a thin film operation status monitoring model based on the degree of thin film vibration, thin film tension and thin film lateral drift, and output the thin film operation stability index. The visual inspection system health monitoring module is used to construct a visual inspection system health monitoring model based on light source power stability, illumination uniformity, and camera sensor temperature, and outputs a visual system health index. The color difference detection confidence assessment module uses ambient temperature as an influencing factor to construct a color difference detection confidence assessment model based on the film operation stability index and the visual system health index, and outputs the real-time detection confidence index. The adaptive transport speed control module is used to construct an adaptive transport speed control model based on the real-time detection confidence index, the film operation stability index and the current film transport speed, and output a recommended transport speed adjustment amount, and periodically adjust the film transport speed according to the recommended transport speed adjustment amount.

2. The automated color difference detection system for packaging bag production according to claim 1, characterized in that, The difference between the quantized value of the real-time vibration amplitude and the vibration value under ideal conditions is divided by (the upper limit of the vibration allowed by the system minus the vibration value under ideal conditions) to obtain the quotient. The membrane vibration degree index is obtained by subtracting the quotient from 1. The difference between the real-time film tension value and the tension value under ideal conditions is taken as the absolute value and then divided by the allowable deviation range of the film tension value to obtain the quotient. The film tension index is obtained by subtracting the quotient from 1. The quotient is obtained by dividing the real-time lateral drift of the thin film by the maximum allowable offset threshold of the system. The lateral drift index of the thin film is obtained by subtracting the quotient from 1.

3. The automated color difference detection system for packaging bag production according to claim 2, characterized in that, In the thin film operating status monitoring model: The membrane operation stability index is calculated based on the membrane vibration degree index, membrane tension index, and membrane lateral drift index. The calculation method is as follows: first, take the minimum value among the three as the basic short board value, and then perform exponential weighting on it. Then, based on the coupling penalty coefficient, the square root of the sum of squares of the deviations of the membrane vibration degree index, membrane tension index, and membrane lateral drift index from the ideal value "1" is calculated. The square root is multiplied by the preset coupling penalty coefficient to obtain the penalty term. The penalty term is then subtracted from 1, and the result is multiplied by the aforementioned weighted short board value to obtain the final membrane operation stability index.

4. The automated color difference detection system for packaging bag production according to claim 1, characterized in that, The difference between the light source power volatility and the ideal volatility is divided by (the difference between the upper limit of volatility and the ideal volatility) to obtain a result. Subtracting this result from 1 yields the light source power stability index. The light uniformity index is obtained by taking the absolute value of the difference between the real-time light uniformity and the ideal uniformity, and then dividing it by (the difference between the upper limit of the allowable light uniformity and the ideal uniformity). The difference between the real-time camera sensor temperature value and the optimal camera sensor temperature value is taken as the absolute value and then divided by (the maximum allowable camera sensor temperature value minus the optimal camera sensor temperature value) to obtain a result. Subtracting this result from 1 yields the camera sensor temperature index.

5. The automated color difference detection system for packaging bag production according to claim 4, characterized in that, In the health monitoring model of the visual inspection system: The visual system health index is obtained by multiplying the light source power stability index, the illumination uniformity index, and the camera sensor temperature index by an exponential decay factor. The calculation method of the exponential attenuation factor is as follows: with the natural constant e as the base, the sum of the squares of the deviations of the light source power stability index, the illumination uniformity index and the camera sensor temperature index from the ideal value "1" is taken as a whole, multiplied by a preset synergistic degradation amplification factor, and the negative number is taken, and then exponential calculation is performed to obtain the final attenuation factor.

6. The automated color difference detection system for packaging bag production according to claim 1, characterized in that, The difference between the real-time ambient temperature value and the optimal ambient temperature value is taken as the absolute value, then divided by (the maximum allowable ambient temperature value minus the optimal ambient temperature value) to obtain a result. Subtracting this result from 1 yields the ambient temperature index. In the color difference detection confidence assessment model, the real-time detection confidence index is calculated as follows: First, the arithmetic mean of the thin film operation stability index and the vision system health index is calculated. Then, an adjustment factor is calculated based on the ambient temperature index and the preset environmental vulnerability coefficient. This adjustment factor is "1" minus a correction term consisting of the environmental vulnerability coefficient, the deviation of the ambient temperature index from the ideal value "1", and the deviation of the aforementioned arithmetic mean from the ideal value "1". Finally, the aforementioned arithmetic mean is multiplied by the adjustment factor to obtain the final real-time detection confidence index.

7. The automated color difference detection system for packaging bag production according to claim 1, characterized in that, The current conveying speed index is obtained by subtracting the minimum allowable conveying speed from the current film conveying speed, and then dividing by the allowable conveying speed range. In the adaptive conveying speed control model, the recommended conveying speed adjustment is calculated as follows: The adjustment period is multiplied by the system's allowable speed range to form the basic scaling factor; Based on the difference between the real-time detection confidence index and the target confidence level, the first adjustment component is calculated using the hyperbolic tangent function. This component is controlled by a preset adjustment intensity coefficient and a preset confidence deviation sensitivity. The second adjustment component is calculated based on the difference between the thin film operation stability index and the stability speed regulation threshold; Multiplying the first adjustment component by the second adjustment component yields the active adjustment term related to the confidence level; At the same time, subtract a reference value obtained by power mapping of the thin film operation stability index from the current transmission speed index, and then multiply by the preset speed inertia damping coefficient to obtain the speed inertia damping term that suppresses excessive speed change. Finally, the active adjustment term is subtracted from the inertial damping term, and multiplied by the aforementioned basic scaling factor to obtain the final recommended transmission speed adjustment.

8. The automated color difference detection system for packaging bag production according to claim 1, characterized in that, It also includes a thermal management optimization control module, which is used to construct a heat dissipation fan control model based on the sensor target temperature, sensor actual temperature, real-time detection confidence index and ambient temperature, and output cooling fan power adjustment value, and optimize the heat dissipation efficiency of the detection system based on the cooling fan power adjustment value.

9. The automated color difference detection system for packaging bag production according to claim 8, characterized in that, The actual sensor temperature and the target sensor temperature are respectively substituted into the maximum-minimum normalization formula for processing, and the actual sensor temperature index and the target sensor temperature index are obtained respectively; in the heat dissipation fan control model, the calculation method of the cooling fan power regulation value is as follows: The deviation between the ambient temperature index and the ideal value "1" is used as the basic cooling requirement; the portion of the actual temperature index of the sensor that exceeds the target temperature index is divided by the deviation between the target temperature index and the ideal value "1", and this result is multiplied by the real-time detection confidence index as the overheating positive feedback adjustment term. The portion of the sensor's target temperature index that exceeds the actual temperature index is divided by the target temperature index, and this result is multiplied by the real-time detection confidence index and the energy-saving adjustment coefficient to form an energy-saving suppression adjustment term. Add the basic cooling demand item and the overheating positive feedback adjustment item, and then subtract the energy saving suppression adjustment item to obtain the preliminary control value; finally, limit the preliminary control value within the range of [0, 1] to obtain the final cooling fan power control value.

Images