A method for automated production control of radial bamboo curtains integrating AI quality inspection

Abstract

This invention belongs to the field of intelligent manufacturing technology and discloses an automated production control method for radial bamboo curtains that integrates AI quality inspection. The method includes: real-time monitoring of the electrostatic field distribution during the radial bamboo curtain production process, constructing a spatiotemporal distribution dataset of the electrostatic field, and calculating a dust-electrostatic adsorption potential energy map by combining the real-time collected dust concentration at each workstation; predicting the dust deposition rate at each workstation based on the dust-electrostatic adsorption potential energy map, and predicting the dust distribution density map by combining it with a predefined bamboo curtain movement trajectory, and dynamically identifying high-risk areas and surface defect types; quantitatively evaluating the surface quality coefficient of the radial bamboo curtain surface based on the dust distribution density map, high-risk areas, and surface defect types, and intelligently adjusting the working parameters of the dust removal equipment. This invention deeply integrates AI quality inspection with automated production control, effectively improving the surface quality of the radial bamboo curtain and increasing production efficiency.

Description

Technical Field

[0001] This invention relates to the field of intelligent manufacturing technology, and more specifically, to an automated production control method for radial bamboo curtains that integrates AI quality inspection. Background Technology

[0002] Radial bamboo blinds, as an environmentally friendly woven material, are widely used in architectural decoration, furniture manufacturing, and handicraft production. During the automated production of radial bamboo blinds, the cutting, weaving, and friction of the bamboo generate a large amount of fine dust. This dust not only affects the surface quality and appearance of the product but also harms the production environment and worker health. Traditional dust removal methods mainly employ passive cleaning methods such as mechanical vacuuming or compressed air blowing, which suffer from incomplete dust removal, high energy consumption, and inability to prevent dust adsorption. Especially during the high-speed weaving process of the bamboo blinds, the friction between the bamboo and the equipment generates a significant electrostatic effect, causing dust to be firmly adsorbed onto the surface of the bamboo blinds by electrostatic force, making it difficult for conventional dust removal methods to remove effectively. Furthermore, existing production lines lack the ability to predict dust distribution and cannot adjust dust removal strategies accordingly, resulting in wasted dust removal resources and unstable product quality.

[0003] With the development of intelligent manufacturing technology, integrating technologies such as artificial intelligence, IoT sensing, and big data analysis into traditional manufacturing has become an important way to improve product quality and production efficiency. However, in the bamboo processing field, especially in the production of radial bamboo curtains, the application of intelligent technology is still in its early stages. Existing technologies have not fully recognized the intrinsic relationship between electrostatic field distribution and dust adsorption, lack dust prediction models based on electrostatic field monitoring, and cannot achieve precise control from the perspective of dust adsorption mechanism. At the same time, traditional quality inspection relies on manual visual inspection, which is inefficient and lacks standardized procedures, making it difficult to meet the needs of large-scale automated production. Therefore, it is urgent to develop an automated production control method that can monitor electrostatic field distribution in real time, predict dust adsorption behavior, and intelligently regulate dust removal strategies to improve the surface quality of radial bamboo curtains, reduce production costs, and thus promote the intelligent upgrading of the bamboo processing industry.

[0004] In view of this, the present invention proposes an automated production control method for radial bamboo curtains that integrates AI quality inspection to solve the above problems. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the existing technology and achieve the above objectives, the present invention provides the following technical solution: a radial bamboo curtain automated production control method integrating AI quality inspection, comprising: Step S1: Monitor the electrostatic field distribution during the radial bamboo curtain production process in real time, record the electric field intensity at each workstation at different times, and construct a spatiotemporal distribution dataset of the electrostatic field. Step S2: Collect the dust concentration distribution of each workstation in real time, and combine it with the electrostatic field spatiotemporal distribution dataset to calculate the dust-electrostatic adsorption potential energy map of each workstation. Predict the dust deposition rate at each workstation based on the dust-electrostatic adsorption potential energy map. Step S3: Based on the predefined bamboo curtain movement trajectory and the dust deposition rate at each workstation, predict the dust distribution density map on the radial surface of the bamboo curtain, and dynamically identify high-risk areas and surface defect types. Step S4: Perform intelligent fusion analysis on the dust distribution density map, high-risk areas and surface defect types of the radial bamboo curtain surface to quantitatively evaluate the surface quality coefficient of the radial bamboo curtain surface; Step S5: Gather high-risk areas, surface defect types, and surface quality coefficients on the radial bamboo curtain surface, intelligently adjust the working parameters of the dust removal equipment, and generate and execute a personalized dust removal strategy.

[0006] Furthermore, methods for constructing spatiotemporal distribution datasets of electrostatic fields include: The unit volume is preset, and the three-dimensional space of each station is divided into grids based on the unit volume to obtain multiple grid units corresponding to each station; the electrostatic field distribution during the radial bamboo curtain production process is collected in real time at each station of the radial bamboo curtain production line; based on the electric field intensity in the electrostatic field distribution collected at different times, an electrostatic field spatiotemporal distribution dataset is constructed.

[0007] Furthermore, the method for calculating the dust-electrostatic adsorption potential energy diagram for each workstation includes: The dust concentration distribution and particle size distribution of each workstation are collected in real time. The particle size range corresponding to each particle size interval is obtained from the particle size distribution, and the median particle size of each particle size interval is calculated based on the particle size range. The average electric field intensity of each workstation at different times is calculated based on the spatiotemporal distribution data of the electrostatic field. Based on the average electric field intensity of each workstation at different times, the representative charge of each particle size range at different times is calculated; the number of particles in each particle size range is obtained from the particle size distribution, and the particle weight of each particle size range at different times is calculated; based on the particle weight, the representative charge of each particle size range at the same time is weighted and summed, and the calculation results are corrected according to the preset charge correction coefficient to obtain the average charge of each workstation at different times. Based on the spatiotemporal distribution dataset of the electrostatic field, the cell potential of each grid cell is calculated; based on the average charge and cell potential of each grid cell at the same time, the electrostatic potential energy of each grid cell at different times is calculated; the dust concentration of each grid cell is obtained from the dust concentration distribution, and based on the electrostatic potential energy and dust concentration of each grid cell at the same time, the effective adsorption potential energy of each grid cell at different times is calculated; the effective adsorption potential energy of each grid cell at different times corresponding to the same workstation is summarized to construct the dust-electrostatic adsorption potential energy map of each workstation.

[0008] Furthermore, methods for predicting dust deposition rates at various workstations include: Different digital tags are set for different workstations and marked as workstation tags; environmental influencing factors of each workstation are collected in real time, and the environmental influencing factors, effective adsorption potential energy and workstation tags of the same grid unit at the same time are combined to obtain the settlement prediction data of each grid unit at different times. The settlement prediction data of each grid cell at different times are input into the trained settlement prediction model to predict the dust settling rate of each grid cell at different times. The average dust settling rate of the same grid cell at different times is calculated to obtain the dust deposition rate of each grid cell. The dust deposition rates of each grid cell at the same workstation are summarized to obtain the dust deposition rate at each workstation. The settlement prediction model is a deep neural network model.

[0009] Furthermore, methods for predicting the dust distribution density map on the radial surface of the bamboo curtain include: The movement trajectory of the bamboo curtain is discretized into a time series, and each moment in the time series corresponds to the position coordinates of the radial bamboo curtain on the production line; the projected area of ​​the grid cell is obtained, and the surface of the radial bamboo curtain is divided into multiple bamboo curtain cells based on the projected area; the bamboo curtain coordinates corresponding to each bamboo curtain cell at each moment in the time series are calculated according to the position coordinates at each moment in the time series and the projected area of ​​the bamboo curtain cell. Obtain the spatial range of each grid cell corresponding to each workstation, compare the bamboo curtain coordinates of each bamboo curtain cell at each moment in the time series with the spatial range of each grid cell corresponding to each workstation, and use the grid cell corresponding to the spatial range into which the bamboo curtain coordinates fall as the matching cell of the corresponding bamboo curtain cell at the corresponding moment. Based on the matching cells corresponding to each bamboo curtain cell on the radial surface of the bamboo curtain at different moments, establish the mapping relationship between the bamboo curtain cell and the grid cell at different moments. Based on the mapping relationship and the dust deposition rate at each workstation, the instantaneous dust deposition amount of each bamboo curtain unit at each moment in the time series is calculated; the entire production process of the radial bamboo curtain passing through all workstations is integrated over time to calculate the cumulative dust amount of each bamboo curtain unit; the cumulative dust amount of all bamboo curtain units corresponding to the radial bamboo curtain surface is summarized to construct a dust distribution density map of the radial bamboo curtain surface.

[0010] Furthermore, methods for dynamically identifying high-risk areas on the radial surface of bamboo curtains include: The cumulative dust amount of each bamboo curtain unit in the dust distribution density map is compared with the preset dust threshold. Bamboo curtain units with a cumulative dust amount greater than the dust threshold are marked as risk units. Connectivity analysis is performed on the risk units on the radial surface of the bamboo curtain to merge spatially adjacent risk units into a high-risk area, resulting in multiple high-risk areas on the radial surface of the bamboo curtain.

[0011] Furthermore, methods for dynamically identifying surface defect types on the radial surface of bamboo curtains include: The cumulative dust amount of each bamboo curtain unit in the dust distribution density map is normalized to obtain the standard dust amount of each bamboo curtain unit and form a dust distribution standard map. The standard dust amount of each bamboo curtain unit in the dust distribution standard map is mapped to the corresponding pixel gray value in turn, and a dust gray value image is formed based on the pixel gray value of each bamboo curtain unit. The grayscale image of dust is input into the trained defect recognition model to predict the predicted label vector. The predicted label vector includes multiple binary labels, each of which corresponds to a type of surface defect. If the value of the binary label is 1, it indicates that the corresponding surface defect type exists. If the value of the binary label is 0, it indicates that the corresponding surface defect type does not exist. Based on the binary labels with a value of 1 in the predicted label vector, all surface defect types on the radial bamboo curtain surface are determined.

[0012] Furthermore, methods for quantitatively evaluating the surface quality coefficient of radial bamboo blind surfaces include: Global distribution features are extracted from the dust distribution density map, including total dust volume, distribution uniformity, and distribution concentration. Quantitative analysis is performed on each high-risk area to extract regional risk features, including the number of risk areas, the area of ​​the largest risk area, and the risk coverage rate. A defect mapping table is constructed, and the corresponding defect coefficients are obtained from the defect mapping table based on the surface defect type of the radial bamboo curtain. The dust distribution density map is divided into multiple local windows. The variance of the cumulative dust amount of all bamboo curtain units corresponding to each local window is calculated to obtain the local variance of each local window. The mean of the local variances of all local windows is calculated to obtain the average local variance. The variance of the cumulative dust amount of all bamboo curtain units in the dust distribution density map is calculated to obtain the global variance. The ratio of the average local variance to the global variance is calculated to obtain the dust distribution complexity. The global distribution characteristics, regional risk characteristics, defect coefficients, and dust distribution complexity are combined to form a quality assessment vector. Each dimension in the quality assessment vector is normalized sequentially, and a corresponding quality weight is set for each dimension. Based on the quality weights, the normalized quality assessment vector is weighted and summed to obtain the surface quality coefficient of the radial bamboo curtain surface.

[0013] Furthermore, methods for intelligently adjusting the operating parameters of dust removal equipment include: A dust removal equipment capability matrix is ​​constructed, which includes capability feature vectors corresponding to different dust removal equipment. The capability feature vectors include the applicable particle size range, the effective area, the energy consumption level, and the treatment effect score for different surface defect types. An equipment defect matching matrix is ​​constructed, and the optimal matching is solved using the Hungarian algorithm to obtain the dust removal equipment matched for each surface defect type. The dust removal intensity requirement is determined based on the surface quality coefficient, and the treatment effect score of the dust removal equipment matched for each surface defect type is compared with the dust removal intensity requirement. If the treatment effect score is less than the dust removal intensity requirement, the corresponding surface defect type is marked as the type to be enhanced, and dust removal equipment matching the type to be enhanced is obtained; the enhancement effect score is calculated based on the treatment effect scores of all dust removal equipment matching the type to be enhanced; the enhancement effect score is compared with the dust removal intensity requirement again; if the enhancement effect score is less than the dust removal intensity requirement, the dust removal equipment matching the type to be enhanced is obtained iteratively until the enhancement effect score is greater than or equal to the dust removal intensity requirement. Each dust removal device matching each surface defect type is marked as a matching device, and the parameter range corresponding to each matching device is obtained. Multiple sets of different working parameter vectors are constructed based on the parameter range, and each set of working parameter vectors includes a set of working parameters corresponding to each matching device. The best working parameters are selected from multiple sets of different working parameter vectors, and each matching device is intelligently allocated based on the working parameters of each matching device in the best working parameters.

[0014] Furthermore, methods for generating personalized dust removal strategies include: Define a device synergy enhancement matrix, which includes the effect gain coefficient of each dust removal device working after another dust removal device; arrange all enhancement devices corresponding to the enhancement type to obtain multiple device execution orders; for each device execution order, obtain the preceding device corresponding to each enhancement device in turn, where the preceding device is the dust removal device located before the enhancement device in the device execution order; based on the preceding device, obtain the effect gain coefficient of each enhancement device in each device execution order from the device synergy enhancement matrix in turn; Define a time-series cumulative effect function. For each device execution order, calculate the cumulative effect score of each enhanced device in sequence using the time-series cumulative effect function. Based on the cumulative effect scores of all dust removal devices matching the type to be enhanced, calculate the enhancement effect score again and mark it as the enhancement cumulative score. Compare the enhancement cumulative scores of each device execution order and select the device execution order with the highest enhancement cumulative score as the optimal execution order. Generate a personalized dust removal strategy based on the optimal execution order.

[0015] The technical effects and advantages of this invention, which integrates AI quality inspection into the automated production control method for radial bamboo curtains, are as follows: By real-time monitoring of the electrostatic field distribution during the radial bamboo curtain production process and constructing a dust-electrostatic adsorption potential energy map, accurate prediction of dust deposition behavior is achieved from the perspective of dust adsorption mechanism, overcoming the technical challenge of traditional methods being unable to identify electrostatically adsorbed dust. A deep neural network model combined with the bamboo curtain's movement trajectory is used to predict dust distribution density maps, and a convolutional neural network is employed to automatically identify surface defect types, enabling full-process tracking of dust distribution and intelligent identification of defect types, providing a dual basis for precise dust removal through spatial positioning and type identification. Finally, a surface quality coefficient constructed through multi-dimensional feature fusion provides a quantitative basis for dust removal intensity requirements, effectively addressing the problem of unreasonable dust removal resource allocation. The system innovatively establishes an equipment-defect matching mechanism and a multi-equipment synergistic efficiency model. By optimizing equipment allocation through the Hungarian algorithm, optimizing working parameters through heuristic algorithms, and optimizing the execution sequence through time-series cumulative effect functions, it achieves personalized customization of dust removal strategies, improving the targeting and effectiveness of dust removal. Furthermore, the system deeply integrates AI quality inspection with automated production control, effectively improving the surface quality of radial bamboo curtains, reducing energy consumption, and increasing production efficiency. It also replaces traditional experience-based operations with data-driven intelligent decision-making, providing a complete technical solution for the transformation of the bamboo processing industry from extensive to intensive and intelligent processes. This has significant engineering application value and industrial upgrading significance. Attached Figure Description

[0016] Figure 1 This is a flowchart of the automated production control method for radial bamboo curtains that integrates AI quality inspection, according to Embodiment 1 of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Example 1:

[0018] Please see Figure 1 As shown in this embodiment, the radial bamboo curtain automated production control method integrating AI quality inspection includes: Step S1: Monitor the electrostatic field distribution during the radial bamboo curtain production process in real time, record the electric field intensity at each workstation at different times, and construct a spatiotemporal distribution dataset of the electrostatic field.

[0019] Methods for constructing spatiotemporal distribution datasets of electrostatic fields include: A preset unit volume is established, which is pre-set by those skilled in the art based on actual conditions. The three-dimensional space of each workstation is divided into grids based on the unit volume, resulting in multiple grid units corresponding to each workstation. At each workstation of the radial bamboo curtain production line, arrays of non-contact electrostatic field sensors (such as vibration capacitive sensors, capacitive electrostatic sensors, etc.) are deployed to collect the electrostatic field distribution during the radial bamboo curtain production process in real time. The electrostatic field distribution refers to the variation of the magnitude and direction of the electric field intensity vector of each grid unit corresponding to each workstation with its position, used to describe the spatial distribution characteristics of the electrostatic force field. A workstation refers to an independent operating position on the radial bamboo curtain production line with a specific processing or handling function, such as a bamboo opening workstation, planing workstation, weaving workstation, gluing workstation, drying workstation, etc. Based on the electric field intensity in the electrostatic field distribution collected at different times, a spatiotemporal distribution dataset of the electrostatic field is constructed.

[0020] Step S2: Collect the dust concentration distribution of each workstation in real time, and combine it with the electrostatic field spatiotemporal distribution dataset to calculate the dust-electrostatic adsorption potential energy map of each workstation. Predict the dust deposition rate at each workstation based on the dust-electrostatic adsorption potential energy map.

[0021] Methods for calculating the dust-electrostatic adsorption potential energy diagram for each workstation include: An array of online light scattering particle size monitors deployed at each workstation collects real-time data on dust concentration and particle size distribution. Dust concentration distribution refers to the variation of the number of particles per unit volume of air in each grid cell corresponding to each workstation with location; the particle number is the number of dust particles. Particle size distribution refers to the number of particles in different particle size ranges within each grid cell corresponding to each workstation. The particle size range corresponding to each particle size range is obtained from the particle size distribution, and the median particle size of each range is calculated based on the particle size range. Specifically, the particle size range includes the maximum and minimum particle sizes; the median particle size is obtained by averaging the maximum and minimum particle sizes. The average electric field strength at different times is calculated for each workstation within the spatiotemporal distribution dataset of the electrostatic field, where the electric field strength is the same at all times and workstations. It should be noted that the electric field strength used in the calculation of the average electric field strength is in scalar form. Based on the average electric field intensity at each workstation at different times, the representative charge of each particle size range at different times is calculated; the expression for the representative charge is: In the formula, To represent the amount of charge, The vacuum permittivity, The relative permittivity of the particles, The median particle size, Let be the average electric field strength; where, ; The settings are pre-defined by those skilled in the art based on factors such as the type of product material and practical experience. For example, when the dust particles are made of bamboo fiber, the corresponding... ; The particle count for each particle size interval is obtained from the particle size distribution, and the particle counts for different particle size intervals at the same time are summed sequentially to obtain the total number of particles at different times. The ratio between the particle count for each particle size interval at different times and the total number of particles at the corresponding time is calculated to obtain the particle weight for each particle size interval at different times. Based on the particle weight, the representative charge of each particle size interval at the same time is weighted and summed to obtain the weighted charge of each station at different times. The product of the weighted charge of each station at different times and a preset charge correction coefficient is calculated to obtain the average charge of each station at different times. The charge correction coefficient is preset by those skilled in the art based on the spontaneous charge effect of dust particles caused by friction, collision, etc. in the actual production process. In this embodiment, the preferred value range of the charge correction coefficient is [insert range here]. ; From the spatiotemporal distribution dataset of the electrostatic field, the electric field intensity of each grid cell at different times is obtained, and the cell potential of each grid cell is calculated using the finite difference method based on the Poisson equation. It should be noted that the electric field intensity used in the calculation of the cell potential is in vector form, and the finite difference method based on the Poisson equation is an existing technology, the specific process of which will not be elaborated here. The product of the average charge and the cell potential of each grid cell at the same time is calculated to obtain the electrostatic potential energy of each grid cell at different times. The dust concentration of each grid cell is obtained from the dust concentration distribution of each workstation, and the product of the electrostatic potential energy and the dust concentration of each grid cell at the same time is calculated to obtain the effective adsorption potential energy of each grid cell at different times. The effective adsorption potential energy of each grid cell at the same workstation at different times is summarized to construct the dust-electrostatic adsorption potential energy map of each workstation, which is used to reflect the adsorption trend of dust particles in each workstation under electrostatic action, that is, the probability of dust being electrostatically adsorbed at different locations.

[0022] Methods for predicting dust deposition rates at various workstations include: The system collects environmental impact factors at each workstation in real time. These factors include ambient temperature, ambient humidity, and air velocity. The environmental impact factors are collected by environmental monitoring sensors deployed at each workstation, including temperature sensors (for collecting ambient temperature), humidity sensors (for collecting ambient humidity), and wind speed sensors (for collecting air velocity). Different digital labels are assigned to different workstations and marked as workstation labels to distinguish the production characteristics of different workstations. The environmental influencing factors and effective adsorption potential energy of the same grid cell at the same time are combined with the workstation labels to obtain the sedimentation prediction data of each grid cell at different times. The environmental influencing factors corresponding to all grid cells within the same workstation at the same time are the same, i.e., equal to the environmental influencing factors of the corresponding workstation. The sedimentation prediction data of each grid cell at different times are input into the trained sedimentation prediction model to predict the dust settling rate of each grid cell at different times. The dust settling rate refers to the number of particles falling into the grid cell per unit time. The average dust settling rate of the same grid cell at different times is calculated to obtain the dust deposition rate of each grid cell. The dust deposition rates of each grid cell corresponding to the same workstation are summarized to obtain the dust deposition rate at each workstation.

[0023] The settlement prediction model is specifically a deep neural network model, comprising an input layer, hidden layers, and an output layer. Each hidden layer contains multiple neurons, and each neuron is connected to neurons in the next layer. These connections contain weights that determine the importance and impact of data transmitted within the neural network. An activation function is applied to each neuron between the hidden and output layers. This activation function introduces non-linearity, allowing the network to learn more complex patterns and features. The training process of the settlement prediction model includes: Pre-collection Different sets of sedimentation prediction data were used, and corresponding dust settling rates were set for each set of sedimentation prediction data in sequence. The values ​​are integers greater than 1; the sedimentation prediction data and the corresponding dust sedimentation rate are converted into a set of corresponding feature vectors; the dust sedimentation rate corresponding to the sedimentation prediction data is collected by those skilled in the art during the historical prediction of dust deposition rates at various work sites. Different sets of sedimentation prediction data were analyzed sequentially based on historical dust deposition rate records for each workstation, environmental influencing factors, and production process conditions. The dust deposition rate for each set of prediction data was determined, and the results were then used to... Set the corresponding dust settling rate for each set of different settling prediction data; Each set of feature vectors is used as input to the settling prediction model. The settling prediction model outputs a set of predicted dust settling velocities corresponding to each set of settling prediction data, and uses the actual dust settling velocity corresponding to each set of settling prediction data as the prediction target. The actual dust settling velocity is the pre-set dust settling velocity corresponding to the settling prediction data. The training objective is to minimize the sum of prediction biases of all settling prediction data. The formula for calculating the prediction bias is: In the formula, For prediction error, This represents the group number of the feature vector corresponding to the settlement prediction data. For the first The predicted dust settling rate corresponding to the settling prediction data. For the first The actual dust settling rate corresponding to the settling prediction data is determined; the settling prediction model is trained until the sum of prediction errors converges, at which point training stops.

[0024] Step S3: Based on the predefined bamboo curtain movement trajectory and the dust deposition rate at each workstation, predict the dust distribution density map on the radial surface of the bamboo curtain, and dynamically identify high-risk areas and surface defect types.

[0025] Methods for predicting the dust distribution density map on the radial surface of bamboo curtains include: The predefined bamboo curtain movement trajectory is obtained. The bamboo curtain movement trajectory is provided by the production line control system, including the movement path, movement speed and dwell time of the bamboo curtain between each workstation. The bamboo curtain movement trajectory is discretized into a time series. Each moment in the time series corresponds to the radial position coordinates of the bamboo curtain on the production line. The position coordinates are represented by a three-dimensional coordinate system, with the origin set at the starting position of the production line, the X-axis along the production line direction, the Y-axis along the width direction of the bamboo curtain, and the Z-axis along the vertical direction. Obtain the projected area of ​​the grid cell on the XY plane, and divide the radial bamboo curtain surface into multiple bamboo curtain cells based on the projected area; wherein, the XY plane is a horizontal reference plane jointly formed by the X-axis (production line direction) and the Y-axis (bamboo curtain width direction); calculate the position coordinates of each bamboo curtain cell at each moment in the time series (i.e., the position coordinates of the center point of the bamboo curtain cell) based on the position coordinates of each moment in the time series and the projected area of ​​the bamboo curtain cell, and mark them as bamboo curtain coordinates; obtain the spatial range of each workstation corresponding to each grid cell, the spatial range including the horizontal coordinate range (including the minimum horizontal coordinate and the maximum horizontal coordinate), the vertical coordinate range (including the minimum vertical coordinate and the maximum vertical coordinate), and the height range (including the minimum height and the maximum height); The bamboo curtain coordinates of each bamboo curtain unit at each time point in the time series are compared with the spatial range of each grid unit at each workstation. The grid unit corresponding to the spatial range in which the bamboo curtain coordinates fall is used as the matching unit of the corresponding bamboo curtain unit at the corresponding time point. Based on the matching units corresponding to each bamboo curtain unit on the radial bamboo curtain surface at different times, the mapping relationship between the bamboo curtain unit and the grid unit at different times is established. Specifically, when the radial bamboo curtain moves to a certain workstation, the bamboo curtain unit on the radial bamboo curtain surface corresponds to the grid unit in the corresponding workstation. Obtain the time interval between two adjacent moments in the time series and mark it as the time step; calculate the dust deposition rate of the matching unit corresponding to each bamboo curtain unit at each moment in the time series, and multiply it by the time step to obtain the instantaneous dust deposition amount of each bamboo curtain unit at each moment in the time series; if there is no matching unit corresponding to the bamboo curtain unit, the instantaneous dust deposition amount of the corresponding bamboo curtain unit is 0, indicating that the bamboo curtain coordinates do not fall within the spatial range of any grid unit (such as when transferring between workstations); add the instantaneous dust deposition amounts of the same bamboo curtain unit at different moments in the time series to obtain the cumulative dust amount of each bamboo curtain unit corresponding to the radial bamboo curtain surface; summarize the cumulative dust amounts of all bamboo curtain units corresponding to the radial bamboo curtain surface to construct a dust distribution density map of the radial bamboo curtain surface.

[0026] It should be noted that dust deposition during the transfer of radial bamboo curtains between workstations is not considered in the entire production process because the radial bamboo curtains move rapidly during the transfer process and are usually located in relatively enclosed or dust-free work areas. The amount of dust deposition is very limited and can be ignored compared to the amount of dust deposition during the stay and processing stages of the radial bamboo curtains at each workstation. Therefore, it has a minimal impact on the overall dust distribution density map.

[0027] Methods for dynamically identifying high-risk areas on the radial surface of bamboo curtains include: A preset dust threshold is established, which is pre-set by those skilled in the art based on product quality standards and historical production experience. The cumulative dust amount of each bamboo curtain unit in the dust distribution density map is compared with the dust threshold. Bamboo curtain units with a cumulative dust amount greater than the dust threshold are marked as risk units, while bamboo curtain units with a cumulative dust amount less than or equal to the dust threshold are not marked. Connectivity analysis is performed on the risk units on the radial surface of the bamboo curtain to merge spatially adjacent risk units into a high-risk area, resulting in multiple high-risk areas on the radial surface of the bamboo curtain.

[0028] Methods for dynamically identifying surface defect types on the radial surface of bamboo curtains include: The cumulative dust amount of each bamboo curtain unit in the dust distribution density map is normalized to obtain the standard dust amount of each bamboo curtain unit, forming a standard dust distribution map. The standard dust amount of each bamboo curtain unit in the standard dust distribution map is then mapped sequentially to the corresponding pixel grayscale values, and a dust grayscale image is formed based on the pixel grayscale values ​​of each bamboo curtain unit. Specifically, the maximum dust (the largest cumulative dust amount) and the minimum dust (the smallest cumulative dust amount) in the dust distribution density map are obtained, and the difference between the maximum and minimum dust is calculated to obtain the dust difference. The difference between the cumulative dust amount and the minimum dust amount of each bamboo curtain unit is calculated and divided by the dust difference to obtain the standard dust amount of each bamboo curtain unit. The standard dust amount of each bamboo curtain unit is then mapped sequentially to the range of pixel grayscale values. The pixel grayscale value of each bamboo curtain unit is obtained; that is, the standard dust amount of each bamboo curtain unit is calculated by multiplying it by 255. The larger the standard dust amount, the higher the corresponding pixel grayscale value. The grayscale image of dust is input into the trained defect recognition model to predict the predicted label vector. The predicted label vector includes multiple binary labels, each corresponding to a surface defect type, indicating whether the corresponding surface defect type exists. If the value of the binary label is 1, it means that the corresponding surface defect type exists; if the value of the binary label is 0, it means that the corresponding surface defect type does not exist. Based on the binary labels with a value of 1 in the predicted label vector, all surface defect types on the radial bamboo curtain surface are determined. Surface defect types include, but are not limited to, local isolated defects (single point or small area of ​​dust deposition), strip continuous defects (dust deposition that extends continuously along a certain direction), regional cluster defects (large area, irregular shape of dust deposition), grid texture defects (dust deposition with a periodic grid or interwoven structure), and ring structure defects (dust deposition that is sparse in the center and distributed in a ring shape around it).

[0029] The defect recognition model is specifically a convolutional neural network (CNN) model, an extension of deep neural networks. It mainly includes an input layer, multiple convolutional layers, pooling layers (downsampling layers), fully connected layers, and an output layer. Each convolutional layer consists of multiple convolutional kernels (filters), each sliding with the input data to extract local features. Weight parameters are included in the convolution operation to learn the importance of features. Pooling layers are typically followed by convolutional layers to reduce the size of feature maps, reducing computational complexity while retaining key information. Fully connected layers or global pooling layers follow the convolutional and pooling layers, organizing the deep features extracted by the convolutional layers into fixed-dimensional feature vectors to represent the texture features of each seed. In the convolutional and fully connected layers, each neuron typically applies an activation function, introducing non-linearity to enhance the model's ability to express complex image patterns and its generalization capabilities. The specific training process of the defect recognition model includes: Multiple grayscale images of dust are pre-collected and labeled as training images. Those skilled in the art analyze each training image sequentially, identifying the surface defect type corresponding to each image, and constructing a corresponding prediction label vector based on the identified surface defect type. Each training image is then labeled sequentially according to the prediction label vector. The labeled training images are divided into a training set and a test set, with 70% of the training images used as the training set and 30% as the test set. The defect recognition model is trained using the training set and tested using the test set. A preset error threshold is established. When the mean prediction error of all grayscale images of dust in the test set is less than the error threshold, the defect recognition model is output, indicating that the defect recognition model training is complete. The prediction error is calculated using a binary cross-entropy loss function, and the error threshold is preset according to the required accuracy of the defect recognition model.

[0030] Step S4: Perform intelligent fusion analysis on the dust distribution density map, high-risk areas and surface defect types of the radial bamboo curtain surface to quantitatively evaluate the surface quality coefficient of the radial bamboo curtain surface.

[0031] Methods for quantitatively evaluating the surface quality coefficient of radial bamboo blinds include: Global distribution features are extracted from the dust distribution density map, including total dust volume, distribution uniformity, and distribution concentration. The total dust volume is calculated by summing the cumulative dust volume of all bamboo curtain units in the dust distribution density map. Distribution uniformity is calculated by calculating the mean and standard deviation of the cumulative dust volume of all bamboo curtain units in the dust distribution density map; the ratio of the mean to the standard deviation is then calculated to obtain the distribution uniformity. Distribution concentration is calculated by counting the number of bamboo curtain units in the dust distribution density map whose cumulative dust volume is greater than the mean, marking them as high-dust units; counting the total number of bamboo curtain units in the dust distribution density map, marking them as the total number of units; and calculating the ratio of the total number of units to the high-dust units to obtain the distribution concentration. A quantitative analysis was conducted on each high-risk area to extract regional risk characteristics, including the number of risk areas, the area of ​​the maximum risk area, and the risk coverage rate. The number of risk areas equals the number of high-risk areas. The maximum risk area was calculated by counting the number of bamboo curtain units corresponding to each high-risk area, obtaining the area of ​​each high-risk area, and then comparing the areas of all high-risk areas, taking the area with the largest value as the maximum risk area. The risk coverage rate was calculated by summing the areas of all high-risk areas sequentially and dividing by the total number of units. A defect mapping table is constructed, which includes a defect coefficient corresponding to each type of surface defect. This coefficient is pre-set by those skilled in the art based on the degree of influence of each type of surface defect on the quality of the radial bamboo curtain product. The corresponding defect coefficient is obtained from the defect mapping table according to the type of surface defect on the radial bamboo curtain surface. The dust distribution density map is divided into multiple local windows, each containing... Two adjacent bamboo curtain units, The variable is an integer greater than 1. The variance of the cumulative dust amount for all bamboo curtain units corresponding to each local window is calculated to obtain the local variance of each local window. The mean of the local variances of all local windows is calculated to obtain the average local variance. The variance of the cumulative dust amount for all bamboo curtain units in the dust distribution density map is calculated to obtain the global variance. The ratio of the average local variance to the global variance is calculated to obtain the dust distribution complexity. The global distribution characteristics, regional risk characteristics, defect coefficient, and dust distribution complexity are combined to form a quality assessment vector. Specifically, the quality assessment vector = {total dust, distribution uniformity, distribution concentration, number of risk areas, area of ​​the largest risk area, risk coverage, defect coefficient, dust distribution complexity}. Each dimension of the quality assessment vector is then normalized sequentially, mapping the value of each dimension to... Interval; set corresponding quality weights for each dimension in the quality assessment vector, and perform weighted summation calculation on each dimension in the normalized quality assessment vector based on the quality weights to obtain the surface quality coefficient of the radial bamboo curtain surface.

[0032] Among them, the quality weights are pre-set by those skilled in the art based on product quality requirements and customer focus, and are used to reflect the importance of each dimension to the final quality assessment; the quality weight corresponding to the distribution uniformity is a positive number, while the quality weights corresponding to the total dust volume, distribution concentration, number of risk areas, area of ​​the largest risk area, risk coverage rate, defect coefficient and dust distribution complexity are all negative numbers. The specific reasons are as follows: the greater the uniformity of distribution, the more uniform the dust distribution on the radial bamboo curtain surface, and therefore the greater the surface quality coefficient; the greater the total amount of dust, the more dust on the radial bamboo curtain surface, and therefore the smaller the surface quality coefficient; the greater the concentration of distribution, the more concentrated the dust on the radial bamboo curtain surface, and the more serious the local pollution, and therefore the smaller the surface quality coefficient; the greater the number of risk areas, the more risk units on the radial bamboo curtain surface, and therefore the smaller the surface quality coefficient; the greater the area of ​​the maximum risk area, the larger the most serious high-risk area on the radial bamboo curtain surface, and therefore the smaller the surface quality coefficient; the greater the risk coverage rate, the greater the proportion of high-risk areas on the radial bamboo curtain surface, and therefore the smaller the surface quality coefficient; the greater the defect coefficient, the more serious the defects on the radial bamboo curtain surface, and therefore the smaller the surface quality coefficient; the greater the complexity of dust distribution, the more irregular the dust distribution on the radial bamboo curtain surface, and therefore the smaller the surface quality coefficient.

[0033] Step S5: Gather high-risk areas, surface defect types, and surface quality coefficients on the radial bamboo curtain surface, intelligently adjust the working parameters of the dust removal equipment, and generate and execute a personalized dust removal strategy.

[0034] Methods for intelligently adjusting the operating parameters of dust removal equipment include: A dust removal equipment capability matrix is ​​constructed, which includes capability feature vectors corresponding to different dust removal equipment. This matrix is ​​used to quantitatively evaluate the capabilities of various dust removal equipment configured at the end of the production line. The dust removal equipment includes cyclone dust collectors (suitable for large particles), electrostatic precipitators (using a reverse electric field to adsorb charged dust), pulse jet cleaning devices (using high-pressure airflow to peel off attached dust), ultrasonic dust collectors (using high-frequency vibration to loosen dust), and composite dust removal units (integrating multiple dust removal principles). The capability feature vectors include the applicable particle size range, effective area, energy consumption level, and treatment effect score for different surface defect types. The applicable particle size range, effective area, and energy consumption level are obtained by those skilled in the art based on the technical specifications of each dust removal equipment. The treatment effect score is obtained by those skilled in the art based on historical dust removal data from radial bamboo curtain dust removal processes, and is used to quantify the cleaning effect of each dust removal equipment on different surface defect types. For example, the cyclone dust collector has a treatment effect score of 0.9 for isolated defects and a treatment effect score of 0.3 for grid-textured defects. Constructing a device defect matching matrix ,in The number of dust removal equipment, The matrix elements represent the number of surface defect types on the radial surface of the bamboo curtain. For the first The dust removal equipment is for the first The treatment effect of each surface defect type is evaluated. Based on the equipment defect matching matrix, the Hungarian algorithm is used to solve for the optimal match, obtaining the dust removal equipment matched for each surface defect type. That is, each surface defect type on the radial bamboo curtain surface is matched with a dust removal equipment that meets the matching constraints, such as the effective area covering the high-risk area corresponding to the corresponding surface defect type, and the particle size range covering all particle size intervals corresponding to the corresponding surface defect type. The Hungarian algorithm is an existing technology, and the specific process will not be elaborated here. The difference between the surface quality coefficient and the dust removal intensity requirement is calculated. The treatment effect score of the dust removal equipment matched for each surface defect type is compared with the dust removal intensity requirement. If all treatment effect scores are greater than or equal to the dust removal intensity requirement, the operating parameters of all matched dust removal equipment are intelligently adjusted. If any treatment effect score is less than the dust removal intensity requirement, the corresponding surface defect type is marked as the type to be enhanced, and the dust removal equipment matching the surface defect type is removed from the equipment defect matching matrix. Based on the remaining dust removal equipment in the equipment defect matching matrix, the Hungarian algorithm is used again to solve for the optimal match, obtaining the dust removal equipment matching the type to be enhanced. Based on the treatment effect scores of all dust removal equipment matching the type to be enhanced, the enhancement effect score is calculated. The enhancement effect score is compared with the dust removal intensity requirement again. If the enhancement effect score is less than the dust removal intensity requirement, the dust removal equipment matching the type to be enhanced is obtained iteratively until the enhancement effect score is greater than or equal to the dust removal intensity requirement, and the operating parameters of all matched dust removal equipment are intelligently adjusted. Each dust removal device matching each surface defect type is labeled as a matching device, and the parameter range corresponding to each matching device is obtained. The parameter range includes the value range of each working parameter corresponding to the matching device, determined by those skilled in the art based on the technical specifications and historical working parameters of each dust removal device. Working parameters include general parameters (applicable to all dust removal devices) and independent parameters (applicable only to specific types of dust removal devices). General parameters include, but are not limited to, moving speed and working time, while independent parameters include, but are not limited to, the suction strength of cyclone dust collectors and the scanning frequency of pulse jet cleaning devices and ultrasonic dust collectors. A value is randomly selected from each value range corresponding to each matching device, and a set of working parameters corresponding to each matching device is constructed. A separate set of working parameters is constructed for each matching device. Group working parameters, It is an integer greater than 1; for each matching device... The working parameters are combined to obtain multiple sets of different working parameter vectors. Each set of working parameter vectors includes a set of working parameters corresponding to each matching device. Based on the high-risk areas, surface defect types, and surface quality coefficients of the radial bamboo curtain surface, a heuristic optimization algorithm (such as ant colony optimization algorithm, water wave optimization algorithm, genetic algorithm, etc.) is used to select a set of working parameter vectors from the multiple sets of different working parameter vectors, which is then used as the optimal working parameters. Based on the working parameters of each matching device in the optimal working parameters, each matching device is intelligently allocated.

[0035] Specifically, the expression for the enhancement effect score is: ; In the formula, To enhance the performance rating, For the first The processing effect score of the enhanced equipment To increase the number of devices, The coefficient represents the synergy factor; where the enhancing equipment is the dust removal equipment of the type to be enhanced. The expression for the synergy coefficient is: In the formula, This is the cooperative gain ratio coefficient, which is preset by those skilled in the art based on the actual situation.

[0036] Methods for generating personalized dust removal strategies include: Define a synergistic effect matrix for equipment, which includes the effect gain coefficient of each dust removal device working after another dust removal device. The effect gain coefficient is obtained by those skilled in the art through statistical analysis based on the working principles of different dust removal devices and the synergistic dust removal situation in the historical radial bamboo curtain dust removal process. For example, after the ultrasonic dust collector loosens the dust through high-frequency vibration, the dust removal efficiency of the cyclone dust collector can be improved by 30% compared with direct use, so the corresponding effect gain coefficient is set to 1.3. Define a time-series cumulative effect function, the specific expression of which is: ; In the formula, For the first The cumulative performance score of each dust removal device For the first The dust removal equipment's treatment effect rating. For the first The dust removal equipment is for the first The effect gain coefficient of a dust removal device; All enhancement devices corresponding to the enhancement type are arranged to obtain multiple device execution sequences. For each device execution sequence, the preceding device for each enhancement device is obtained sequentially. The preceding device is the dust removal device that precedes the enhancement device in the device execution sequence. Based on the preceding device, the effect gain coefficient of each enhancement device in each device execution sequence is obtained sequentially from the device synergy enhancement matrix. For each device execution sequence, the cumulative effect score of each enhancement device is calculated sequentially using the time-series cumulative effect function. Based on the cumulative effect scores of all dust removal devices matching the enhancement type, the enhancement effect score is calculated again and marked as the enhancement cumulative score. Specifically, the processing effect score in the expression of the enhancement effect score is replaced with the corresponding cumulative effect score, thereby recalculating the enhancement effect score. The enhancement cumulative scores of each device execution sequence are compared, and the device execution sequence with the highest enhancement cumulative score is selected as the optimal execution sequence. Based on the optimal execution sequence, a personalized dust removal strategy is generated.

[0037] This embodiment achieves accurate prediction of dust deposition behavior from the dust adsorption mechanism level by real-time monitoring of the electrostatic field distribution during the radial bamboo curtain production process and constructing a dust-electrostatic adsorption potential energy map, overcoming the technical difficulty of traditional methods in identifying electrostatically adsorbed dust. It predicts dust distribution density maps by combining a deep neural network model with the bamboo curtain's movement trajectory and uses a convolutional neural network to automatically identify surface defect types, achieving full-process tracking of dust distribution and intelligent identification of defect types, providing a dual basis for precise dust removal through spatial positioning and type identification. Finally, the surface quality coefficient constructed through multi-dimensional feature fusion provides a quantitative basis for dust removal intensity requirements, effectively addressing the problem of unreasonable dust removal resource allocation. The project innovatively establishes an equipment-defect matching mechanism and a multi-equipment collaborative efficiency model. It optimizes equipment allocation through the Hungarian algorithm, optimizes working parameters through heuristic algorithms, and optimizes the execution sequence through a time-series cumulative effect function, thereby achieving personalized customization of dust removal strategies and improving the targeting and effectiveness of dust removal. Furthermore, it deeply integrates AI quality inspection with automated production control, effectively improving the surface quality of radial bamboo curtains, reducing energy consumption, and increasing production efficiency. It also replaces traditional experience-based operations with data-driven intelligent decision-making, providing a complete technical solution for the transformation of the bamboo processing industry from extensive to intensive and intelligent processes. This has significant engineering application value and industrial upgrading significance. Example 2:

[0038] This application also provides an electronic device. The electronic device may include one or more processors and one or more memories. The memories store computer-readable code that, when executed by the one or more processors, can perform the radial bamboo curtain automated production control method incorporating AI quality inspection as described above.

[0039] The method according to the embodiments of this application can also be implemented using the architecture of the electronic device shown in this application. The electronic device may include a bus, one or more CPUs, ROM, RAM, a communication port connected to a network, input / output, a hard disk, etc. The storage device in the electronic device, such as a ROM or hard disk, may store the radial bamboo curtain automated production control method integrating AI quality inspection provided in this application. Furthermore, the electronic device may also include a user interface. Of course, the architecture shown in this application is merely exemplary; when implementing different devices, one or more components in the electronic device shown in this application may be omitted according to actual needs. Example 3:

[0040] Please refer to the accompanying drawings. One embodiment of this application discloses a computer-readable storage medium. The computer-readable storage medium stores computer-readable instructions. When the computer-readable instructions are executed by a processor, the radial bamboo curtain automated production control method integrating AI quality inspection according to an embodiment of this application, as described above, can be performed. The storage medium includes, but is not limited to, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0041] Furthermore, according to embodiments of this application, the processes described in the above-referenced flowcharts can be implemented as computer software programs. For example, this application provides a non-transitory machine-readable storage medium storing machine-readable instructions that can be executed by a processor to perform instructions corresponding to the method steps provided in this application, such as a radial bamboo curtain automated production control method incorporating AI quality inspection. When this computer program is executed by a central processing unit (CPU), it performs the functions defined in the method of this application.

[0042] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0043] All formulas in this manual are dimensionless and calculated numerically. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0044] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A radial bamboo curtain automated production control method integrating AI quality inspection, characterized in that, include: Step S1: Monitor the electrostatic field distribution during the radial bamboo curtain production process in real time, record the electric field intensity at each workstation at different times, and construct a spatiotemporal distribution dataset of the electrostatic field. Step S2: Collect the dust concentration distribution of each workstation in real time, and combine it with the electrostatic field spatiotemporal distribution dataset to calculate the dust-electrostatic adsorption potential energy map of each workstation. Predict the dust deposition rate at each workstation based on the dust-electrostatic adsorption potential energy map. Step S3: Based on the predefined bamboo curtain movement trajectory and the dust deposition rate at each workstation, predict the dust distribution density map on the radial surface of the bamboo curtain, and dynamically identify high-risk areas and surface defect types. Step S4: Perform intelligent fusion analysis on the dust distribution density map, high-risk areas and surface defect types of the radial bamboo curtain surface to quantitatively evaluate the surface quality coefficient of the radial bamboo curtain surface; Step S5: Gather high-risk areas, surface defect types, and surface quality coefficients on the radial bamboo curtain surface, intelligently adjust the working parameters of the dust removal equipment, and generate and execute a personalized dust removal strategy.

2. The radial bamboo curtain automated production control method integrating AI quality inspection according to claim 1, characterized in that, Methods for constructing spatiotemporal distribution datasets of electrostatic fields include: The unit volume is preset, and the three-dimensional space of each station is divided into grids based on the unit volume to obtain multiple grid units corresponding to each station; the electrostatic field distribution during the radial bamboo curtain production process is collected in real time at each station of the radial bamboo curtain production line; based on the electric field intensity in the electrostatic field distribution collected at different times, an electrostatic field spatiotemporal distribution dataset is constructed.

3. The radial bamboo curtain automated production control method integrating AI quality inspection according to claim 2, characterized in that, Methods for calculating the dust-electrostatic adsorption potential energy diagram for each workstation include: The dust concentration distribution and particle size distribution of each workstation are collected in real time. The particle size range corresponding to each particle size interval is obtained from the particle size distribution, and the median particle size of each particle size interval is calculated based on the particle size range. The average electric field intensity of each workstation at different times is calculated based on the spatiotemporal distribution data of the electrostatic field. Based on the average electric field intensity of each workstation at different times, the representative charge of each particle size range at different times is calculated; the number of particles in each particle size range is obtained from the particle size distribution, and the particle weight of each particle size range at different times is calculated; based on the particle weight, the representative charge of each particle size range at the same time is weighted and summed, and the calculation results are corrected according to the preset charge correction coefficient to obtain the average charge of each workstation at different times. Based on the spatiotemporal distribution dataset of the electrostatic field, the cell potential of each grid cell is calculated; based on the average charge and cell potential of each grid cell at the same time, the electrostatic potential energy of each grid cell at different times is calculated; the dust concentration of each grid cell is obtained from the dust concentration distribution, and based on the electrostatic potential energy and dust concentration of each grid cell at the same time, the effective adsorption potential energy of each grid cell at different times is calculated; the effective adsorption potential energy of each grid cell at different times corresponding to the same workstation is summarized to construct the dust-electrostatic adsorption potential energy map of each workstation.

4. The radial bamboo curtain automated production control method integrating AI quality inspection according to claim 3, characterized in that, Methods for predicting dust deposition rates at various workstations include: Different digital tags are set for different workstations and marked as workstation tags; environmental influencing factors of each workstation are collected in real time, and the environmental influencing factors, effective adsorption potential energy and workstation tags of the same grid unit at the same time are combined to obtain the settlement prediction data of each grid unit at different times. The settlement prediction data of each grid cell at different times are input into the trained settlement prediction model to predict the dust settling rate of each grid cell at different times. The average dust settling rate of the same grid cell at different times is calculated to obtain the dust deposition rate of each grid cell. The dust deposition rates of each grid cell at the same workstation are summarized to obtain the dust deposition rate at each workstation. The settlement prediction model is a deep neural network model.

5. The radial bamboo curtain automated production control method integrating AI quality inspection according to claim 4, characterized in that, Methods for predicting the dust distribution density map on the radial surface of bamboo curtains include: The movement trajectory of the bamboo curtain is discretized into a time series, and each moment in the time series corresponds to the position coordinates of the radial bamboo curtain on the production line; the projected area of ​​the grid cell is obtained, and the surface of the radial bamboo curtain is divided into multiple bamboo curtain cells based on the projected area; the bamboo curtain coordinates corresponding to each bamboo curtain cell at each moment in the time series are calculated according to the position coordinates at each moment in the time series and the projected area of ​​the bamboo curtain cell. Obtain the spatial range of each grid cell corresponding to each workstation, compare the bamboo curtain coordinates of each bamboo curtain cell at each moment in the time series with the spatial range of each grid cell corresponding to each workstation, and use the grid cell corresponding to the spatial range into which the bamboo curtain coordinates fall as the matching cell of the corresponding bamboo curtain cell at the corresponding moment. Based on the matching cells corresponding to each bamboo curtain cell on the radial surface of the bamboo curtain at different moments, establish the mapping relationship between the bamboo curtain cell and the grid cell at different moments. Based on the mapping relationship and the dust deposition rate at each workstation, the instantaneous dust deposition amount of each bamboo curtain unit at each moment in the time series is calculated; the entire production process of the radial bamboo curtain passing through all workstations is integrated over time to calculate the cumulative dust amount of each bamboo curtain unit; the cumulative dust amount of all bamboo curtain units corresponding to the radial bamboo curtain surface is summarized to construct a dust distribution density map of the radial bamboo curtain surface.

6. The radial bamboo curtain automated production control method integrating AI quality inspection according to claim 5, characterized in that, Methods for dynamically identifying high-risk areas on the radial surface of bamboo curtains include: The cumulative dust amount of each bamboo curtain unit in the dust distribution density map is compared with the preset dust threshold. Bamboo curtain units with a cumulative dust amount greater than the dust threshold are marked as risk units. Connectivity analysis is performed on the risk units on the radial surface of the bamboo curtain to merge spatially adjacent risk units into a high-risk area, resulting in multiple high-risk areas on the radial surface of the bamboo curtain.

7. The radial bamboo curtain automated production control method integrating AI quality inspection according to claim 6, characterized in that, Methods for dynamically identifying surface defect types on the radial surface of bamboo curtains include: The cumulative dust amount of each bamboo curtain unit in the dust distribution density map is normalized to obtain the standard dust amount of each bamboo curtain unit and form a dust distribution standard map. The standard dust amount of each bamboo curtain unit in the dust distribution standard map is mapped to the corresponding pixel gray value in turn, and a dust gray value image is formed based on the pixel gray value of each bamboo curtain unit. The grayscale image of dust is input into the trained defect recognition model to predict the predicted label vector. The predicted label vector includes multiple binary labels, each of which corresponds to a type of surface defect. If the value of the binary label is 1, it indicates that the corresponding surface defect type exists. If the value of the binary label is 0, it indicates that the corresponding surface defect type does not exist. Based on the binary labels with a value of 1 in the predicted label vector, all surface defect types on the radial bamboo curtain surface are determined.

8. The radial bamboo curtain automated production control method integrating AI quality inspection according to claim 7, characterized in that, Methods for quantitatively evaluating the surface quality coefficient of radial bamboo blinds include: Global distribution features are extracted from the dust distribution density map, including total dust volume, distribution uniformity, and distribution concentration. Quantitative analysis is performed on each high-risk area to extract regional risk features, including the number of risk areas, the area of ​​the largest risk area, and the risk coverage rate. A defect mapping table is constructed, and the corresponding defect coefficients are obtained from the defect mapping table based on the surface defect type of the radial bamboo curtain. The dust distribution density map is divided into multiple local windows. The variance of the cumulative dust amount of all bamboo curtain units corresponding to each local window is calculated to obtain the local variance of each local window. The mean of the local variances of all local windows is calculated to obtain the average local variance. The variance of the cumulative dust amount of all bamboo curtain units in the dust distribution density map is calculated to obtain the global variance. The ratio of the average local variance to the global variance is calculated to obtain the dust distribution complexity. The global distribution characteristics, regional risk characteristics, defect coefficients, and dust distribution complexity are combined to form a quality assessment vector. Each dimension in the quality assessment vector is normalized sequentially, and a corresponding quality weight is set for each dimension. Based on the quality weights, the normalized quality assessment vector is weighted and summed to obtain the surface quality coefficient of the radial bamboo curtain surface.

9. The radial bamboo curtain automated production control method integrating AI quality inspection according to claim 8, characterized in that, Methods for intelligently adjusting the operating parameters of dust removal equipment include: A dust removal equipment capability matrix is ​​constructed, which includes capability feature vectors corresponding to different dust removal equipment. The capability feature vectors include the applicable particle size range, the effective area, the energy consumption level, and the treatment effect score for different surface defect types. An equipment defect matching matrix is ​​constructed, and the optimal matching is solved using the Hungarian algorithm to obtain the dust removal equipment matched for each surface defect type. The dust removal intensity requirement is determined based on the surface quality coefficient, and the treatment effect score of the dust removal equipment matched for each surface defect type is compared with the dust removal intensity requirement. If the treatment effect score is less than the dust removal intensity requirement, the corresponding surface defect type is marked as the type to be enhanced, and dust removal equipment matching the type to be enhanced is obtained; the enhancement effect score is calculated based on the treatment effect scores of all dust removal equipment matching the type to be enhanced; the enhancement effect score is compared with the dust removal intensity requirement again; if the enhancement effect score is less than the dust removal intensity requirement, the dust removal equipment matching the type to be enhanced is obtained iteratively until the enhancement effect score is greater than or equal to the dust removal intensity requirement. Each dust removal device matching each surface defect type is marked as a matching device, and the parameter range corresponding to each matching device is obtained. Multiple sets of different working parameter vectors are constructed based on the parameter range, and each set of working parameter vectors includes a set of working parameters corresponding to each matching device. The best working parameters are selected from multiple sets of different working parameter vectors, and each matching device is intelligently allocated based on the working parameters of each matching device in the best working parameters.

10. The radial bamboo curtain automated production control method integrating AI quality inspection according to claim 9, characterized in that, Methods for generating personalized dust removal strategies include: Define a device synergy enhancement matrix, which includes the effect gain coefficient of each dust removal device working after another dust removal device; arrange all enhancement devices corresponding to the enhancement type to obtain multiple device execution orders; for each device execution order, obtain the preceding device corresponding to each enhancement device in turn, where the preceding device is the dust removal device located before the enhancement device in the device execution order; based on the preceding device, obtain the effect gain coefficient of each enhancement device in each device execution order from the device synergy enhancement matrix in turn; Define a time-series cumulative effect function. For each device execution order, calculate the cumulative effect score of each enhanced device in sequence using the time-series cumulative effect function. Based on the cumulative effect scores of all dust removal devices matching the type to be enhanced, calculate the enhancement effect score again and mark it as the enhancement cumulative score. Compare the enhancement cumulative scores of each device execution order and select the device execution order with the highest enhancement cumulative score as the optimal execution order. Generate a personalized dust removal strategy based on the optimal execution order.

Images