Method for manufacturing a stuffed plush toy

Abstract

The application relates to a production method of a cotton-filled plush toy in the field of cotton-filled plush, and the production method comprises the following steps: obtaining cotton raw materials to be treated, extracting the morphological characteristics of the cotton through an image recognition technology to obtain the loose degree and fiber length parameters of the cotton, judging whether the cotton raw materials meet the requirements of a bacteriostatic immersion process according to preset parameter thresholds, if yes, entering the bacteriostatic treatment, and if not, starting a raw material sorting device to screen qualified cotton; filling the toy into a sewing device, collecting sewing position parameters under the guidance of machine vision, controlling the motion track of a sewing needle according to the sewing position parameters, completing the sewing and sealing of the toy, obtaining a plush toy finished product, realizing the intelligent screening of the cotton raw materials, the accurate regulation and control of the bacteriostatic immersion process and the uniformity control of the toy filling, and significantly improving the quality stability and production efficiency of the plush toy.

Description

Technical Field

[0001] The present invention relates to the field of cotton-filled plush, in particular to the field of cotton-filled plush toys, and specifically to a method for manufacturing a cotton-filled plush toy. Background Art

[0002] In the plush toy production process, antibacterial treatment of cotton, softener impregnation, and filling are three critical and interdependent processes. Traditionally, these three steps are performed separately, posing numerous challenges that require urgent resolution. First, cotton exposed to air after antibacterial treatment is susceptible to recontamination, making it difficult to ensure hygienic safety. Second, the impregnated cotton has a high moisture content and requires drying before filling, which is not only time-consuming and energy-intensive but also reduces the cotton's fluffiness, affecting the toy's feel. Third, the cotton tends to clump during the filling process, making it difficult to fill evenly, resulting in a less plump and rounded appearance and shape.

[0003] The contradictions between these three steps lie in how to ensure thorough antibacterial treatment while maintaining the cotton's maximum softness and fluffiness and achieving uniform filling. In traditional processes, the cotton, after antibacterial treatment and impregnation, is exposed to air and dried at high temperatures, all of which affect its quality. Furthermore, the filling process makes it difficult to control the uniformity of the cotton, further affecting the appearance and feel of the product. In short, existing processes struggle to achieve a balanced balance of antibacterial, softness, and filling, while performing each separately is inefficient and energy-intensive.

[0004] Therefore, there is an urgent need to explore a new production method that integrates antibacterial treatment, impregnation, and filling to address the challenge of cotton quality control in plush toy production at the source. This new process must be able to seal the cotton after antibacterial treatment to prevent secondary contamination, and quickly fill it after softener impregnation to avoid drying that damages the cotton's fluffiness. It must also ensure uniform filling to preserve the toy's appearance and feel. For the plush toy industry, process innovation and optimization are key to improving product quality, increasing production efficiency, and reducing energy consumption. Summary of the Invention

[0005] The purpose of the present invention is to solve the above defects and provide a method for making a cotton-filled plush toy to solve the technical problems in the above background technology.

[0006] The object of the present invention is achieved in the following ways:

[0007] A method for manufacturing a cotton-filled plush toy, the method comprising:

[0008] Step S101: obtaining cotton raw material to be processed, extracting morphological features of the cotton using image recognition technology, obtaining parameters of looseness and fiber length of the cotton, and judging whether the cotton raw material meets the requirements of the antibacterial impregnation process according to preset parameter thresholds. If so, the cotton raw material enters the antibacterial treatment process; if not, the raw material sorting device is activated to screen out qualified cotton.

[0009] Step S102: The qualified cotton is conveyed to the antibacterial device, and an atomizing nozzle sprays an antibacterial liquid on the surface of the cotton. An infrared sensor collects the surface temperature of the cotton, and a concentration sensor collects the concentration of the antibacterial liquid. The spraying volume of the antibacterial liquid is adjusted within a range of 50 to 80 milliliters and the spraying time is adjusted within a range of 60 to 90 seconds based on the temperature and concentration values. After the antibacterial treatment, the antibacterial cotton is obtained;

[0010] Step S103: The antibacterial cotton is transported to an impregnation device. The vacuum degree is controlled within a range of 0.05 to 0.08 MPa and the impregnation time is controlled within a range of 180 to 240 seconds. A pressure sensor collects the cavity pressure value, and an online moisture meter collects the cotton moisture content value. The vacuum degree is adjusted according to the pressure value and the moisture content value to obtain soft and fluffy impregnated cotton.

[0011] Step S104: After the impregnated cotton is dehydrated by the dehydrating device, it enters the drying device. The temperature and humidity sensor collects the drying environment temperature and humidity values, and the moisture meter collects the cotton moisture content value. Based on the temperature, humidity and moisture content values, the hot air temperature is adjusted to a range of 60 to 80 degrees Celsius and the air volume is adjusted to a range of 2 to 3 cubic meters per minute. Image recognition is used to collect the cotton fluffiness value to obtain the dried cotton.

[0012] In step S105, the dry cotton enters the filling device through the air conveying pipe. Machine vision identifies the shape and size parameters of the toy cavity, generates filling path parameters and filling amount parameters for each filling point, and a multi-axis robotic arm controls the movement trajectory and filling speed of the filling head. A pressure sensor collects cavity filling density distribution values ​​every 5 centimeters. Based on the density distribution values, the filling amount is adjusted to a range of 15 to 20 grams per cubic decimeter to obtain a stuffed toy.

[0013] In step S106, the stuffed toy enters the sewing device, and the machine vision guides the collection of sewing position parameters. The automatic sewing controls the movement trajectory of the sewing needle according to the sewing position parameters to complete the sewing and sealing of the toy, thereby obtaining a finished plush toy.

[0014] The beneficial effects of the present invention are as follows: the morphological characteristics of cotton are extracted and screened through image recognition technology, the antibacterial process is monitored in real time in combination with infrared sensors and concentration sensors, the impregnation effect is accurately controlled by vacuum impregnation and pressure sensors, the drying process is regulated by temperature and humidity sensors and moisture meters, and finally precise filling is achieved by machine vision and multi-axis robotic arms. This invention realizes the intelligent screening of cotton raw materials, precise regulation of the antibacterial impregnation process, and uniformity control of toy filling, which significantly improves the quality stability and production efficiency of plush toys. The present invention is a method for manufacturing plush toys that mainly solves the technical problems of unstable cotton raw material quality, uneven antibacterial impregnation effect, and inconsistent filling density in the traditional plush toy production process. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a schematic flow chart of a method for making a cotton-filled plush toy according to this embodiment; DETAILED DESCRIPTION

[0016] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0017] In this embodiment, refer to Figure 1 The specific manufacturing method may include:

[0018] Step S101, obtain the cotton raw material to be processed, extract the morphological characteristics of the cotton through image recognition technology, obtain parameters such as the looseness and fiber length of the cotton, and judge whether the cotton raw material meets the requirements of the antibacterial impregnation process according to the preset parameter threshold. If it meets the requirements, it enters the antibacterial treatment. If not, the raw material sorting device is started to screen out qualified cotton.

[0019] Obtain the cotton raw material to be processed, collect cotton image data, and transmit the image data to the image processing module. The image processing module pre-processes the cotton image, including image denoising and enhancement, to improve image quality. Extract cotton morphological characteristic parameters, including looseness, fiber length, etc., and use a convolutional neural network algorithm for feature extraction. Compare the extracted cotton morphological characteristic parameters with the preset threshold to determine whether the cotton raw material meets the requirements of the antibacterial impregnation process. If the cotton raw material meets the process requirements, the qualified cotton is transferred to the antibacterial treatment unit for antibacterial impregnation. If the cotton raw material does not meet the process requirements, the raw material sorting device is activated, and the support vector machine algorithm is used to grade the cotton quality. The qualified cotton screened out is transferred to the antibacterial treatment unit, and the cotton that does not meet the standards is discarded to ensure the quality of the raw materials for subsequent processes.

[0020] Specifically, a 500-gram cotton sample was randomly sampled from a cotton raw material warehouse. Image data was captured using a high-definition digital camera with a resolution of 12 megapixels, a color depth of 24 bits, and a JPG image format. The captured image data was then transferred to an image processing module via the USB0 interface. This module used a Gaussian filter algorithm to denoise the cotton image, removing high-frequency noise. A histogram equalization algorithm was then used to enhance the image contrast and clarify the cotton fiber outlines. A convolutional neural network model was then used to extract features from the cotton image, extracting morphological parameters such as cotton bulkiness and fiber length. The bulkiness was calculated using a gray-level co-occurrence matrix algorithm, ranging from 0 to 1, with larger values ​​indicating greater bulkiness. The fiber length was calculated using a skeleton extraction algorithm, measured in millimeters, with larger values ​​indicating longer fibers. The extracted cotton morphological characteristic parameters are compared with preset thresholds, with the bulkiness threshold set at 6 and the fiber length threshold set at 25 mm. If the bulkiness of the cotton sample is greater than or equal to 6 and the fiber length is greater than or equal to 25 mm, the batch of cotton raw material is determined to meet the requirements of the antibacterial impregnation process and is transferred to the antibacterial treatment unit for antibacterial impregnation. If the bulkiness of the cotton sample is less than 6 or the fiber length is less than 25 mm, the raw material sorting device is activated to grade the cotton quality. A support vector machine algorithm is used to comprehensively evaluate the bulkiness and fiber length of the cotton sample. Based on the evaluation results, the cotton is divided into three grades: A, B, and C, with Grade A cotton having the highest quality and Grade C cotton having the lowest quality. Grade A cotton is transferred to the antibacterial treatment unit, while Grades B and C cotton are discarded to ensure the quality of raw materials for subsequent processes. This intelligent cotton raw material quality detection and sorting process can effectively improve the raw material quality of the antibacterial impregnation process and ensure the stability of product performance.

[0021] In step S102, the qualified cotton is transported to the antibacterial device, and the atomizing nozzle sprays the antibacterial liquid on the surface of the cotton. The infrared sensor collects the surface temperature value of the cotton, and the concentration sensor collects the concentration value of the antibacterial liquid. The spraying volume of the antibacterial liquid is adjusted within the range of 50 to 80 ml and the spraying time is adjusted within the range of 60 to 90 seconds according to the feedback of the temperature value and the concentration value. After the antibacterial treatment, the antibacterial cotton is obtained.

[0022] The batch information of qualified cotton is acquired, and the proportioning scheme of the bacteriostatic liquid is determined according to the batch information, and the corresponding bacteriostatic liquid proportioning data is called from the bacteriostatic liquid proportioning scheme database. The bacteriostatic liquid proportioning data is sent to the bacteriostatic device, and the bacteriostatic liquid proportioning unit is controlled to mix various bacteriostatic raw materials according to the proportioning data to obtain the proportioned bacteriostatic liquid. The proportioned bacteriostatic liquid is sprayed to the surface of the qualified cotton through the atomizing nozzle, and the temperature value of the surface of the cotton is collected in real time through the infrared sensor. The temperature value of the surface of the cotton is compared with the preset temperature threshold value, if the temperature value is lower than the temperature threshold value, the atomizing nozzle is controlled to increase the spraying amount of the bacteriostatic liquid, if the temperature value is higher than the temperature threshold value, the atomizing nozzle is controlled to reduce the spraying amount of the bacteriostatic liquid.

[0023] The concentration value of the bacteriostatic liquid is collected in real time through the concentration sensor, and the concentration value is compared with the preset concentration threshold value, if the concentration value is lower than the concentration threshold value, the spraying time of the bacteriostatic liquid is controlled to be extended, if the concentration value is higher than the concentration threshold value, the spraying time of the bacteriostatic liquid is controlled to be shortened. In the process of spraying the bacteriostatic liquid, the distribution of the bacteriostatic liquid on the surface of the cotton is monitored in real time through the machine vision system, and the spraying angle and spraying track of the atomizing nozzle are dynamically adjusted according to the monitoring result to ensure the uniform distribution of the bacteriostatic liquid on the surface of the cotton. The cotton after completing the bacteriostatic treatment is detected for moisture, and whether the bacteriostatic cotton reaches the predetermined bacteriostatic standard is judged according to the moisture detection result, if yes, the bacteriostatic cotton is transported to the next process, if no, the cotton is re-transported to the bacteriostatic device for secondary bacteriostatic treatment.

[0024] Specifically, first, the batch information of qualified cotton is acquired from the cotton information management system, such as the cotton variety is Xinjiang long-staple cotton, the place of origin is Xinjiang Uygur Autonomous Region Akesu, the batch number is AKS20221201, and the weight of the cotton is 5 tons. According to the batch information such as the cotton variety, the place of origin, and the weight, the bacteriostatic liquid proportioning data is called from the bacteriostatic liquid proportioning scheme database by using a fuzzy matching algorithm, wherein the addition proportion of bacteriostatic agent A is 5%, the addition proportion of bacteriostatic agent B is 2%, and the addition proportion of auxiliary agent C is 5%. The bacteriostatic liquid proportioning data is sent to the PLC controller of the bacteriostatic device through industrial Ethernet, and the bacteriostatic liquid proportioning unit is controlled to mix bacteriostatic agent A, bacteriostatic agent B, and auxiliary agent C according to the proportion of 5:2:5 to obtain the proportioned bacteriostatic liquid. The bacteriostatic device uniformly sprays the proportioned bacteriostatic liquid to the surface of the cotton through six atomizing nozzles, the atomizing particle size of the nozzle is 50-100 μm, and the atomizing pressure is 2-3 MPa.

[0025] At the same time, four infrared sensors installed above the conveyor belt collect real-time temperature readings on the cotton surface at a sampling frequency of 10 times per second. The system compares the collected cotton surface temperature with a preset temperature threshold of 45°C in real time. If the temperature readings are below 45°C for three consecutive times, the system controls the atomizing nozzle to increase the amount of antibacterial liquid sprayed by 20%. If the temperature readings are above 45°C for three consecutive times, the system controls the atomizing nozzle to reduce the amount of antibacterial liquid sprayed by 20%.

[0026] In addition, two concentration sensors installed below the conveyor belt collect real-time concentration data of the antibacterial liquid at a sampling frequency of 5 times per second. The system compares these concentration values ​​with a preset concentration threshold of 30%. If the concentration value is below 30% for two consecutive times, the system controls the spraying time to be extended by 5 seconds; if the concentration value is above 30% for two consecutive times, the system controls the spraying time to be shortened by 5 seconds. During the antibacterial liquid spraying process, two high-definition cameras installed above the conveyor belt monitor the distribution of the antibacterial liquid on the cotton surface in real time at a frequency of 5 frames per second. The system processes the captured images using an image segmentation algorithm to extract the area covered by the antibacterial liquid on the cotton surface and calculate the antibacterial liquid coverage rate. If the coverage rate falls below 90%, the system dynamically adjusts the spray angle and spray trajectory of the six atomizing nozzles based on the coverage distribution to ensure uniform distribution of the antibacterial liquid on the cotton surface. After the cotton is treated by the antibacterial device, the moisture content sensor is used to measure the moisture content of the cotton at a sampling frequency of 1 time per second. The system uses a moisture balance model to calculate the target moisture content of cotton, which is in the range of 15% to 15%, based on the variety and weight of the cotton.

[0027] The moisture detection results are compared with the target moisture content range. If the moisture content of 10 consecutive samples is within the target range, the antibacterial cotton is judged to have reached the predetermined antibacterial standard and is transported to the next drying process. If the moisture content of 10 consecutive samples exceeds the target range, the system will automatically transport the batch of cotton back to the antibacterial device for secondary antibacterial treatment until the antibacterial standard is met.

[0028] According to the cotton surface temperature value collected by the infrared sensor and the antibacterial liquid concentration value collected by the concentration sensor, the appropriate range of the antibacterial liquid spraying amount and spraying time is obtained through feedback adjustment, and the optimal antibacterial treatment parameters are determined.

[0029] An infrared sensor collects cotton surface temperature data in real time and transmits it to a data processing module. A concentration sensor acquires the antibacterial solution concentration in real time and transmits it to the data processing module. Within the data processing module, the received cotton surface temperature and antibacterial solution concentration data undergo data preprocessing, including data cleaning and normalization, to generate standardized temperature and concentration data. Based on preset temperature and concentration thresholds, the system determines whether the currently collected cotton surface temperature and antibacterial solution concentration are within a reasonable range. If they exceed these thresholds, an alarm mechanism is triggered. The standardized temperature and concentration data are then input into a pre-built BP neural network model. Through model training and parameter optimization, optimal predictions for the antibacterial solution spraying volume and time are obtained. Based on the predictions output by the BP neural network model, a fuzzy control algorithm is used to dynamically adjust the spraying volume and time of the antibacterial solution spraying device, achieving real-time optimized control of the antibacterial solution spraying process. The optimized antibacterial liquid spraying volume and spraying time parameters are fed back to the data processing module, and the changes in key parameters of the antibacterial treatment process are displayed in real time through data visualization technology, providing data support for subsequent process optimization.

[0030] Specifically, an infrared sensor collects cotton surface temperature data every 1 second and transmits this data to the data processing module via the RS485 bus. A concentration sensor acquires the antibacterial solution concentration every 2 seconds and transmits this value to the data processing module via a 4-20 mA current signal. After receiving the temperature and concentration data, the data processing module uses the least squares method to clean the data to remove outliers and noise. It then uses the maximum-minimum normalization method to map the temperature and concentration data to the [0, 1] range to generate standardized data. The system presets a threshold range of [40°C-50°C] for cotton surface temperature and [20%, 40%] for antibacterial solution concentration. If the currently collected temperature or concentration exceeds the threshold range, an alarm signal is sent to the host computer. The standardized temperature and concentration data are input into a three-layer BP neural network model with 10 hidden layer neurons, a learning rate of 0.1, and a maximum number of iterations of 1000. The model is trained using a backpropagation algorithm to obtain predicted values ​​for the antibacterial solution spraying amount and spraying time.

[0031] According to the predicted value, a fuzzy control rule base is established by using a Mamdani fuzzy reasoning method, input variables are temperature deviation E and concentration deviation EC, output variables are spraying amount deviation U and spraying time deviation UT, the domain range is [-3, 3], and fuzzy processing is performed by using a triangular membership function. Accurate spraying amount and spraying time control amount are obtained by a center average solution defuzzification method, and dynamic optimization control of the bacteriostatic liquid spraying process is realized. The optimized control parameters are fed back to the data processing module through a data bus, and real-time line graphs of temperature, concentration and control parameters are generated by using an Echarts tool, and dynamic change conditions of the bacteriostatic treatment process are intuitively displayed, thereby providing data reference for process optimization.

[0032] In step S103, the bacteriostatic cotton is conveyed to the impregnation device, the vacuum impregnation control is performed at a vacuum degree value in the range of 0.05 to 0.08 MPa and an impregnation time value in the range of 180 to 240 seconds, a pressure sensor collects a cavity pressure value, an online moisture meter collects a cotton moisture content value, the vacuum degree value is adjusted according to the pressure value and the moisture content value, and soft and fluffy impregnated cotton is obtained.

[0033] The initial moisture content value of the bacteriostatic cotton is obtained, and the bacteriostatic cotton is conveyed to the vacuum impregnation device for treatment. According to the preset vacuum degree range and impregnation time range, the process parameters of the vacuum impregnation device are initially set. The pressure value in the cavity of the vacuum impregnation device is collected in real time by a pressure sensor, and the moisture content value of the cotton is collected in real time by an online moisture meter. The collected pressure value and moisture content value are input into a neural network model, and the soft and fluffy degree of the cotton under the current process parameters is predicted by the trained model. If the predicted soft and fluffy degree does not meet the expectation, the vacuum degree parameter is optimized by a genetic algorithm, and the impregnation time parameter is optimized by a particle swarm algorithm. The optimized vacuum degree parameter and impregnation time parameter are output, and the vacuum degree and impregnation time of the vacuum impregnation device are dynamically adjusted according to the parameter values. Until the soft and fluffy degree of the cotton meets the expected requirement, the vacuum impregnation soft and fluffy treatment of the bacteriostatic cotton is completed.

[0034] Specifically, an online moisture meter first collects the initial moisture content of the antibacterial cotton, for example, 8%, and then transfers it to a vacuum impregnation device for processing. Based on the cotton variety and quality grade, the corresponding vacuum degree range of -0.8 MPa to -1 MPa and impregnation time range of 60 to 80 seconds are retrieved from the process parameter database and used as the initial process parameters for the vacuum impregnation device. During the vacuum impregnation process, a pressure sensor collects the pressure inside the vacuum impregnation device chamber every 5 seconds, and the online moisture meter collects the moisture content of the cotton every 10 seconds. The collected pressure and moisture content values ​​are input into a BP neural network model trained through 1000 iterations. Based on the input pressure and moisture content parameters, the model predicts the softness and fluffiness of the cotton under the current process parameters. If the predicted value is less than 85, the parameter optimization process is initiated. A genetic algorithm is used to optimize the vacuum degree parameters, with an initial population size of 50, a crossover probability of 6, and a mutation probability of 0.1. The optimal vacuum degree parameters are obtained after 500 iterations. The particle swarm algorithm was used to optimize the impregnation time parameters, setting the particle size to 100, the learning factor to 2, and the maximum number of iterations to 200. The optimal impregnation time parameters were obtained. The optimized vacuum and time parameters were output, and the vacuum level and time of the vacuum impregnation device were dynamically adjusted based on the parameter values, for example, adjusting the vacuum level to -0.95 MPa and the impregnation time to 75 seconds. This process was repeated until the cotton achieved a softness and fluffiness level of 95 or higher, completing the vacuum impregnation softening and fluffing treatment of the antibacterial cotton. The moisture content of the treated cotton was controlled at around 12%, resulting in a soft, fluffy texture and a delicate feel, laying a good foundation for subsequent processing.

[0035] The pressure value in the vacuum impregnation device is collected by a pressure sensor, and the moisture content value of the cotton is collected by an online moisture meter. The data is input into a neural network model to predict the softness and fluffiness of the cotton. If the expected level is not met, the genetic algorithm and particle swarm algorithm are used to optimize the vacuum degree and impregnation time parameters respectively. The equipment process parameters are dynamically adjusted according to the optimized parameter values ​​until the softness and fluffiness of the cotton meet the expected requirements, thus completing the antibacterial cotton vacuum impregnation softening and fluffiness treatment process.

[0036] A pressure sensor acquires real-time pressure within the vacuum impregnation device, while an online moisture meter obtains real-time moisture content of the cotton. These pressure and moisture content values ​​are input into a pre-built neural network model to predict the softness and fluffiness of the cotton under the current process parameters. A genetic algorithm is then used to determine whether the predicted softness and fluffiness of the cotton reaches a preset threshold, using the predicted softness and fluffiness as the fitness function to optimize the vacuum parameters of the vacuum impregnation device. The optimized vacuum parameter value is then obtained using a particle swarm algorithm, using the predicted softness and fluffiness as the objective function. The impregnation time parameter of the vacuum impregnation device is then optimized. Based on the optimized vacuum and time parameters, the process parameters of the vacuum impregnation device are dynamically adjusted, and the pressure and moisture content values ​​are re-acquired. When the predicted softness and fluffiness of the cotton reaches the preset threshold, the vacuum impregnation softening and fluffiness treatment process is completed, and qualified antibacterial cotton products are produced.

[0037] Specifically, during the vacuum impregnation process, a pressure sensor collects real-time pressure readings within the vacuum impregnation apparatus every 2 seconds, and an online moisture meter collects real-time moisture content readings of the cotton every 5 seconds. These collected pressure and moisture content values ​​are input into a BP neural network model that has been trained through 2000 iterations. Based on the input pressure and moisture content parameters, the model predicts the softness and fluffiness of the cotton under the current process parameters. If the predicted softness and fluffiness of the cotton is less than 90, a parameter optimization program is initiated. A genetic algorithm is used to optimize the vacuum parameters, with an initial population size of 80, a crossover probability of 7, and a mutation probability of 0.5. 800 iterations are used to obtain the optimal vacuum parameters. A particle swarm algorithm is used to optimize the impregnation time parameters, with a particle size of 150, a learning factor of 5, and a maximum number of iterations of 300. The optimized vacuum and time parameters are output and dynamically adjusted based on these values, for example, to adjust the vacuum to -0.85 MPa and the impregnation time to 2.10 seconds. The above process is repeated until the predicted softness and fluffiness of the cotton reaches above 95, the vacuum impregnation softening and fluffiness treatment of the antibacterial cotton is completed, and qualified antibacterial cotton products are output, with a moisture content controlled at about 12%, soft and fluffy, and delicate feel, realizing intelligent control of the vacuum impregnation softening and fluffiness treatment process of the antibacterial cotton.

[0038] By collecting pressure and moisture content values, inputting them into a neural network model to predict the softness and fluffiness, and dynamically adjusting the equipment process parameters according to the optimized vacuum value and impregnation time until the expected requirements are met, an antibacterial cotton vacuum impregnation softness and fluffiness treatment process is obtained.

[0039] The pressure and moisture content values ​​of the antibacterial cotton are obtained as input data for the neural network model. The trained neural network model is used to predict the softness and fluffiness of the antibacterial cotton. Based on the predicted softness and fluffiness, the initial values ​​of the process parameters of the vacuum impregnation equipment, including the vacuum level and impregnation time, are determined. The vacuum impregnation equipment is started to soften and fluff the antibacterial cotton. During the vacuum impregnation process, the pressure and moisture content values ​​are collected in real time and re-input into the neural network model. Based on the model's predictions, it is determined whether the softness and fluffiness meet the expected requirements. If not, the vacuum level and impregnation time are dynamically adjusted until the softness and fluffiness meet the expected requirements, completing the softening and fluffiness treatment of the antibacterial cotton.

[0040] Specifically, the pressure and moisture content of the antibacterial cotton were first acquired using a pressure sensor and a moisture sensor, which served as input data for a BP neural network model. The pressure range was 1-0 MPa, and the moisture range was 10%-30%. The 100 data sets were divided into training and test sets in an 8:2 ratio, and the data were normalized. The BP neural network model was trained using the Levenberg-Marquardt algorithm, with 10 hidden layer nodes, a learning rate of 0.1, and 1000 iterations, to determine the model weights and thresholds. The test set data was then fed into the trained model to predict the softness and fluffiness of the antibacterial cotton. The softness and fluffiness was represented by a value from 1 to 5, with a higher value indicating better softness and fluffiness. Based on the predicted softness and fluffiness, a fuzzy control algorithm was used to determine the initial values ​​of the process parameters for the vacuum impregnation equipment, including the vacuum level and impregnation time. The vacuum level ranged from -08 to -1 MPa, and the impregnation time ranged from 10 to 30 minutes. The vacuum impregnation equipment was activated to soften and fluff the antibacterial cotton. During the vacuum impregnation process, pressure and moisture content values ​​were collected every 5 minutes and fed into a neural network model to predict the degree of softness and fluffiness. If the predicted softness and fluffiness was less than 4, the vacuum level and impregnation time were dynamically adjusted according to fuzzy control rules until the desired softness and fluffiness was achieved. Ultimately, through real-time monitoring and dynamic adjustment, the antibacterial cotton was softened and fluffed, improving product quality and production efficiency.

[0041] In step S104, the impregnated cotton is dehydrated by the dehydrating device and then enters the drying device. The temperature and humidity sensors collect the temperature and humidity values ​​of the drying environment, and the moisture meter collects the moisture content value of the cotton. The hot air temperature value is adjusted within the range of 60 to 80 degrees Celsius and the air volume value is adjusted within the range of 2 to 3 cubic meters per minute based on the temperature, humidity and moisture content values. Image recognition is used to collect the cotton fluffiness value to obtain the dried cotton.

[0042] Temperature and humidity sensors collect the temperature and humidity of the drying environment and transmit these values ​​to the control system. A moisture meter collects the moisture content of the cotton after impregnation and dehydration and transmits this value to the control system. Based on the received temperature, humidity, and moisture content values, the control system uses a preset fuzzy control algorithm to determine the adjustment range for the hot air temperature and air volume. If the temperature falls below 60°C or rises above 80°C, the humidity exceeds the set threshold, or the moisture content exceeds the set threshold, the control system issues an alert and automatically adjusts the hot air temperature and air volume based on the deviation to maintain them within a reasonable range. After adjusting the hot air temperature and air volume, the cotton is dried. During the drying process, the moisture content of the cotton is monitored in real time, and the hot air temperature and air volume parameters are dynamically optimized based on the moisture content trend. Image recognition technology is used to measure the bulkiness of the dried cotton. The bulkiness value is calculated by analyzing the cotton's surface morphology. The cotton fluffiness value is compared with the preset threshold. If it is lower than the threshold, the hot air temperature and air volume parameters are adjusted to continue drying the cotton until the fluffiness of the cotton meets the requirements. The cotton drying process is completed and qualified dried cotton is output.

[0043] Specifically, first, the temperature and humidity values ​​of the drying environment are collected through the temperature and humidity sensor. For example, if the ambient temperature is 25°C and the relative humidity is 60%, the collected values ​​are transmitted to the control system. At the same time, the moisture content value of the cotton after impregnation and dehydration is collected using a moisture meter. For example, if the moisture content is 35%, it is also transmitted to the control system. After receiving the temperature, humidity and moisture content values, the control system determines that the hot air temperature adjustment range is 70°C to 75°C and the air volume adjustment range is 1000m 3 / h~1200m 3 During the drying process, if the ambient temperature drops to 18°C, the humidity rises to 80%, and the moisture content of the cotton is higher than 40%, the control system will issue a warning message and automatically increase the hot air temperature to 78°C and the air volume to 1300m3 according to the deviation value. 3 / h, so that the drying environment is maintained within a reasonable range. After the hot air temperature and air volume are adjusted, the cotton is dried. During the drying process, the moisture content of the cotton is monitored in real time. According to the rate of decrease of the moisture content, the fuzzy control algorithm is used to dynamically optimize and adjust the hot air temperature and air volume parameters. For example, in the later stage of drying, the hot air temperature and air volume can be appropriately reduced to prevent the cotton from being over-dried. Image recognition technology is used to analyze the surface morphological characteristics of the dried cotton, and the cotton fluffiness evaluation model is constructed to calculate the cotton fluffiness value. The cotton fluffiness value is compared with the preset threshold of 85%. If the fluffiness is detected to be lower than the threshold, the control system automatically adjusts the hot air temperature to 72°C and the air volume to 1100m 3 / h, continue to dry the cotton until the cotton fluffiness reaches more than 95%, complete the cotton drying process, and output fluffy and soft cotton with a moisture content of about 8%.

[0044] By analyzing the surface morphology of cotton, the cotton bulkiness value is calculated and compared with the preset threshold. If it is lower than the threshold, the hot air temperature and air volume parameters are adjusted and the cotton is continued to be dried until the cotton bulkiness meets the requirements. The cotton drying process is completed and qualified dried cotton is output.

[0045] Cotton surface morphological feature data is acquired and analyzed and extracted using image processing technology. A cotton bulk calculation model is constructed based on the extracted cotton surface morphological features, and the cotton bulk value is calculated using this model. A preset cotton bulk threshold is obtained and the calculated cotton bulk value is compared with the threshold. If the cotton bulk value is lower than the preset threshold, hot air temperature and air volume adjustment parameters are determined using a fuzzy control algorithm based on the difference between the cotton bulk value and the threshold. The determined hot air temperature and air volume adjustment parameters are input into a hot air drying device, and the cotton continues to be dried. After a certain period of hot air drying, the cotton surface morphological feature data is reacquired, the cotton bulk value is calculated, and a determination is made as to whether it meets the preset threshold. If the cotton bulk value meets the preset threshold, the hot air drying process is terminated, and qualified dried cotton is output; otherwise, the hot air temperature and air volume parameters are adjusted and drying continues until the cotton bulk meets the specified value.

[0046] Specifically, a high-definition camera was used to capture cotton surface images, and an image segmentation algorithm was used to extract cotton fiber contour features. A cotton bulk calculation model was constructed based on parameters such as fiber angle, length, and curl. The model was trained using a support vector machine algorithm based on a large amount of cotton sample data, generating a mapping between cotton bulk and fiber morphological parameters. A preset threshold for cotton bulk was set at 8, and the cotton image was fed into the trained model. The calculated bulk value was 75. Comparison revealed that this value was below the preset threshold, and a fuzzy control algorithm determined that the hot air temperature needed to be increased by 5°C and the air volume by 5 cubic meters per minute. The adjusted parameter settings were transmitted to the control unit of the hot air drying equipment, and the cotton was subjected to a 5-minute continuous drying process. After the process was completed, the cotton surface image was captured again, and the fiber morphological features were extracted. The calculated bulk value was 85, which met the preset threshold. The drying process was terminated, and the cotton was exported as qualified cotton that met the bulk requirements.

[0047] In step S105, the dry cotton enters the filling device through the air flow conveying pipe. Machine vision identifies the shape parameters and size parameters of the toy cavity, generates the filling path parameters and the filling amount parameters of each filling point, and the multi-axis robotic arm controls the movement trajectory and filling speed of the filling head. The pressure sensor collects the cavity filling density distribution value every 5 cm. The filling amount value is adjusted within the range of 15 to 20 grams per cubic decimeter according to the density distribution value to obtain a stuffed toy.

[0048] Machine vision captures the shape and size parameters of the toy cavity and transmits these parameters to the filling path planning module. Based on the cavity shape and size parameters, the filling path planning module uses a path optimization algorithm to generate a filling head trajectory and fill quantity parameters for each filling point. These generated filling trajectory and fill quantity parameters are transmitted to the robotic arm control system, which controls the filling head to follow the planned trajectory and speed, filling each filling point with cotton according to the set fill quantity parameters. During the filling process, a pressure sensor collects fill density distribution data within the cavity every 5 cm and transmits these data to the fill quantity adjustment module. Based on the received density distribution data, the fill quantity adjustment module determines whether the fill density is uniform. If the density distribution is uneven, it dynamically adjusts the fill quantity parameters for subsequent filling points to maintain the fill quantity within the range of 15 to 20 grams per cubic decimeter. Dry cotton is continuously transported to the filling device via an air conveyor pipe to ensure an adequate supply of cotton during the filling process. The filling device fills the cotton into each filling point in the toy cavity until the filling process is complete. After filling is completed, the stuffed toy undergoes quality inspection to determine whether the filling is satisfactory. If the filling quality is unqualified, the toy will be returned for refilling; if the filling quality is qualified, the stuffed toy will be transferred to the next process for subsequent processing.

[0049] Specifically, a high-precision industrial camera captures multi-angle images of the toy cavity, and image processing algorithms are used to extract cavity contour feature points to construct a 3D model. Based on the 3D model data, the cavity's shape parameters (such as length, width, and height) and dimensional parameters (such as internal volume) are extracted and passed to the filling path planning module. This module uses a modified A* algorithm to optimize filling efficiency and uniformity, generating a trajectory for the filling head. Through trajectory interpolation and discretization, the continuous trajectory is converted into a sequence of discrete filling points. Based on the spatial position of the filling points and the local shape of the cavity, the filling quantity parameters for each filling point are calculated, generating a filling instruction sequence. The robotic arm control system receives the filling instruction sequence and controls the filling head along the planned trajectory with an accuracy of 1 mm, controlling the filling quantity at each filling point with a resolution of 5 grams. A pressure sensor array scans and samples the filling density within the cavity every 5 cm, generating 3D density distribution data. The filling quantity adjustment module analyzes the collected density data in real time and determines the filling uniformity using a density gradient algorithm. When the density difference exceeds 10%, the filling amount of the filling point is adaptively adjusted to ensure that the local filling density is within the range of 18±2 grams per cubic decimeter. Dry cotton is continuously transported to the filling device at a rate of 1 cubic meter per minute through a blower and an air conveying duct. The filling device uses a servo-controlled extrusion mechanism to fill the cotton into each filling point of the cavity until the entire filling process is completed, thereby realizing the automated filling operation of the toy. After filling is completed, a combination of machine vision and pressure testing is used to inspect the quality of the filled toys. By comparing with the standard model, a comprehensive judgment is made on whether the filling effect is qualified. If the test fails, the toy is automatically returned to the filling station by a robot for refilling until it meets the quality requirements, and then it is transferred to the next process for subsequent processing, thereby realizing online detection and closed-loop control of the filling quality.

[0050] In step S106, the stuffed toy enters the sewing device, and the machine vision guides the collection of sewing position parameters. The automatic sewing controls the movement trajectory of the sewing needle according to the sewing position parameters to complete the sewing and sealing of the toy, thereby obtaining a finished plush toy.

[0051] The machine vision system captures high-precision images of the stitching area, obtaining detailed parameters of the stitching position. Based on the captured image data, image processing algorithms are used to optimize the stitching path, ensuring the accuracy of the stitching path. The automatic sewing machine receives the optimized stitching path data and adjusts the motion trajectory of the sewing needle to adapt to different shapes and sizes of toys. The sewing machine adjusts the speed and direction of the needle in real time during the stitching process to cope with different thickness and elasticity of materials. After completing the stitching, the automatic detection of the stitching quality is performed through machine learning algorithms to analyze the stitching results and ensure that the stitching quality of each plush toy meets the standard. The finished plush toys undergo final quality inspection, packaging and distribution, ready for sale or distribution. Throughout the process, the system optimizes stitching parameters and machine settings through a data feedback loop to improve production efficiency and product quality.

[0052] Specifically, the machine vision system uses a high-resolution industrial camera to capture images of the stitching area, with a resolution of 20 million pixels, which can clearly capture details of 1 millimeter. The captured image data is analyzed through image processing algorithms, using the Canny edge detection algorithm to extract the stitching line, and combining the Hough transform algorithm to optimize the line to generate the best stitching path. The automatic sewing machine receives the optimized stitching path data and converts it into the motion trajectory of the needle through motion control algorithms, and adjusts the motion range of the needle according to the shape and size of the toy. The speed of the needle in the XY plane can reach 500 mm / s, and the speed in the Z-axis direction can reach 200 mm / s, which can adapt to the stitching needs of different materials. The built-in pressure sensor of the sewing machine can detect the thickness change of the material in real time, and dynamically adjust the speed and direction of the needle through the PID control algorithm to ensure the stability of the stitching quality. After the stitching is completed, the system automatically detects the quality of the stitching result, uses a convolutional neural network algorithm to classify the stitching image, and identifies the areas of poor stitching. After training a large number of samples, the accuracy of the algorithm can reach more than 99%.

[0053] Finally, the system optimizes the stitching parameters based on the quality detection results, continuously adjusts the speed, direction and other parameters of the needle through genetic algorithms to improve the efficiency and quality of the stitching. After optimization, the production efficiency of the sewing machine has increased by 20%, and the poor stitching rate has decreased by 50%.

[0054] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention is disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technician familiar with this profession can make some changes or modifications to equivalent embodiments of equivalent changes by using the technical content disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modifications, equivalent changes and modifications made to the above embodiments according to the technology of the present invention are all within the scope of the technical solution of the present invention without departing from the content of the technical solution of the present invention.

Claims

1. A method for making a cotton-filled plush toy, characterized in that: The production method comprises: Step S101: obtaining cotton raw material to be processed, extracting morphological features of the cotton using image recognition technology, obtaining parameters of looseness and fiber length of the cotton, and judging whether the cotton raw material meets the requirements of the antibacterial impregnation process according to preset parameter thresholds. If so, the cotton raw material enters the antibacterial treatment process; if not, the raw material sorting device is activated to screen out qualified cotton. Step S102: The qualified cotton is conveyed to the antibacterial device, and an atomizing nozzle sprays an antibacterial liquid on the surface of the cotton. An infrared sensor collects the surface temperature of the cotton, and a concentration sensor collects the concentration of the antibacterial liquid. The spraying volume of the antibacterial liquid is adjusted within a range of 50 to 80 milliliters and the spraying time is adjusted within a range of 60 to 90 seconds based on the temperature and concentration values. After the antibacterial treatment, the antibacterial cotton is obtained; Step S103: The antibacterial cotton is transported to an impregnation device. The vacuum degree is controlled within a range of 0.05 to 0.08 MPa and the impregnation time is controlled within a range of 180 to 240 seconds. A pressure sensor collects the cavity pressure value, and an online moisture meter collects the cotton moisture content value. The vacuum degree is adjusted according to the pressure value and the moisture content value to obtain soft and fluffy impregnated cotton. Step S104: After the impregnated cotton is dehydrated by the dehydrating device, it enters the drying device. The temperature and humidity sensor collects the drying environment temperature and humidity values, and the moisture meter collects the cotton moisture content value. Based on the temperature, humidity and moisture content values, the hot air temperature is adjusted to a range of 60 to 80 degrees Celsius and the air volume is adjusted to a range of 2 to 3 cubic meters per minute. Image recognition is used to collect the cotton fluffiness value to obtain the dried cotton. In step S105, the dry cotton enters the filling device through the air conveying pipe. Machine vision identifies the shape and size parameters of the toy cavity, generates filling path parameters and filling amount parameters for each filling point, and a multi-axis robotic arm controls the movement trajectory and filling speed of the filling head. A pressure sensor collects cavity filling density distribution values ​​every 5 centimeters. Based on the density distribution values, the filling amount is adjusted to a range of 15 to 20 grams per cubic decimeter to obtain a stuffed toy. In step S106, the stuffed toy enters the sewing device, and the machine vision guides the collection of sewing position parameters. The automatic sewing controls the movement trajectory of the sewing needle according to the sewing position parameters to complete the sewing and sealing of the toy, thereby obtaining a finished plush toy.

2. The method for making a cotton-filled plush toy according to claim 1, wherein: In step S101, the cotton raw material to be processed is obtained, cotton image data is collected, and the image data is transmitted to the image processing module; The image processing module pre-processes the cotton image, including image denoising and enhancement, to improve image quality; Extract cotton morphological characteristic parameters, including looseness and fiber length, and use convolutional neural network algorithm for feature extraction; Compare the extracted cotton morphological characteristic parameters with the preset threshold value to determine whether the cotton raw material meets the requirements of the antibacterial impregnation process; If the cotton raw materials meet the process requirements, the qualified cotton will be sent to the antibacterial treatment unit for antibacterial dipping; If the cotton raw material does not meet the process requirements, the raw material sorting device is activated and the cotton is graded using the support vector machine algorithm; The qualified cotton that has been screened out will be sent to the antibacterial treatment unit, and the cotton that does not meet the standards will be removed to ensure the quality of raw materials for subsequent processes.

3. The method for making a cotton-filled plush toy according to claim 1, wherein: In step S102, the batch information of qualified cotton is obtained, the ratio scheme of the antibacterial liquid is determined according to the batch information, and the corresponding antibacterial liquid ratio data is retrieved from the ratio scheme database; The antibacterial liquid ratio data is sent to the antibacterial device, and the antibacterial liquid batching unit is controlled to mix various antibacterial raw materials according to the ratio data to obtain the antibacterial liquid after the ratio is obtained; The antibacterial liquid is sprayed onto the surface of qualified cotton through an atomizing nozzle, and the temperature value of the cotton surface is collected in real time through an infrared sensor; The cotton surface temperature value is compared with the preset temperature threshold. If the temperature value is lower than the temperature threshold, the atomizing nozzle is controlled to increase the spraying amount of the antibacterial liquid. If the temperature value is higher than the temperature threshold, the atomizing nozzle is controlled to reduce the spraying amount of the antibacterial liquid; The concentration sensor collects the concentration value of the antibacterial liquid in real time, compares the concentration value with the preset concentration threshold, and controls the spraying time of the antibacterial liquid to be extended if the concentration value is lower than the concentration threshold; If the concentration value is higher than the concentration threshold, the spraying time of the antibacterial liquid is shortened; During the antibacterial liquid spraying process, the distribution of the antibacterial liquid on the cotton surface is monitored in real time through a machine vision system. The spraying angle and spraying trajectory of the atomizing nozzle are dynamically adjusted according to the monitoring results to ensure the uniform distribution of the antibacterial liquid on the cotton surface. The moisture content of the cotton after antibacterial treatment is tested, and the moisture content is determined based on the test results to determine whether the antibacterial cotton meets the predetermined antibacterial standard. If the cotton meets the standard, it is transported to the next process; if the cotton does not meet the standard, it is transported back to the antibacterial device for secondary antibacterial treatment. Specifically, it also includes: according to the cotton surface temperature value collected by the infrared sensor and the antibacterial liquid concentration value collected by the concentration sensor, the appropriate range of the antibacterial liquid spraying amount and spraying time is obtained through feedback adjustment, and the optimal antibacterial treatment parameters are determined.

4. The method for making a cotton-filled plush toy according to claim 3, wherein: The infrared sensor is used to collect the cotton surface temperature data in real time, and the collected temperature data is transmitted to the data processing module; The concentration sensor is used to obtain the concentration value of the antibacterial liquid in real time, and the concentration value data is transmitted to the data processing module; In the data processing module, the received cotton surface temperature data and antibacterial liquid concentration data are preprocessed, including data cleaning and data normalization operations, to obtain standardized temperature and concentration data; According to the preset temperature threshold range and concentration threshold range, it is judged whether the currently collected cotton surface temperature and antibacterial liquid concentration are within a reasonable range. If they exceed the threshold range, the alarm mechanism is triggered; The standardized temperature and concentration data were input into the pre-built BP neural network model. Through model training and parameter optimization, the optimal prediction values ​​of the antibacterial liquid spraying amount and spraying time were obtained. According to the prediction results output by the BP neural network model, the fuzzy control algorithm is used to dynamically adjust the spraying amount and spraying time of the antibacterial liquid spraying device to achieve real-time optimization control of the antibacterial liquid spraying process; The optimized antibacterial liquid spraying volume and spraying time parameters are fed back to the data processing module, and the changes in key parameters of the antibacterial treatment process are displayed in real time through data visualization technology, providing data support for subsequent process optimization.

5. The method for making a cotton-filled plush toy according to claim 1, wherein: In step S103, the initial moisture content value of the antibacterial cotton is obtained and the cotton is transported to a vacuum impregnation device for treatment; Initially set the process parameters of the vacuum impregnation device according to the preset vacuum degree range and impregnation time range; The pressure value in the vacuum impregnation device cavity is collected in real time by a pressure sensor, and the moisture content value of the cotton is collected in real time by an online moisture meter; The collected pressure and moisture content values ​​are input into the neural network model, and the trained model is used to predict the softness and fluffiness of the cotton under the current process parameters. If the predicted softness and fluffiness does not meet expectations, the vacuum degree parameters are optimized using a genetic algorithm and the immersion time parameters are optimized using a particle swarm algorithm; Output the optimized vacuum degree parameters and impregnation time parameters, and dynamically adjust the vacuum degree and impregnation time of the vacuum impregnation device according to the parameter values; Until the softness and fluffiness of the cotton reaches the expected requirements, the vacuum impregnation softening and fluffiness treatment of the antibacterial cotton is completed; Specifically, it also includes: collecting the pressure value in the vacuum impregnation device through a pressure sensor and the cotton moisture content value through an online moisture meter, inputting the data into a neural network model to predict the softness and fluffiness of the cotton. If the expected level is not met, a genetic algorithm and a particle swarm algorithm are used to optimize the vacuum degree and impregnation time parameters respectively, and dynamically adjust the equipment process parameters according to the optimized parameter values ​​until the softness and fluffiness of the cotton meet the expected requirements, thus completing the antibacterial cotton vacuum impregnation soft and fluffiness treatment process.

6. The method for making a cotton-filled plush toy according to claim 5, characterized in that: The real-time pressure value in the vacuum impregnation device is obtained by the pressure sensor, and the real-time moisture content value of the cotton is obtained by the online moisture meter; The obtained pressure and moisture content values ​​are input into a pre-built neural network model to predict the softness and fluffiness of the cotton under the current process parameters; Determine whether the predicted softness and fluffiness of the cotton reaches a preset threshold; A genetic algorithm was used to optimize the vacuum parameters of the vacuum impregnation device with the predicted softness and fluffiness of the cotton as the fitness function, and the optimized vacuum parameter values ​​were obtained. The particle swarm algorithm was used to optimize the impregnation time parameter of the vacuum impregnation device with the predicted softness and fluffiness of the cotton as the objective function, and the optimized impregnation time parameter value was obtained. According to the optimized vacuum degree parameter value and impregnation time parameter value, the process parameter settings of the vacuum impregnation device are dynamically adjusted, and the pressure and moisture content values ​​are re-obtained; When the predicted softness and fluffiness of the cotton reaches a preset threshold, the vacuum impregnation softening and fluffiness treatment process of the antibacterial cotton is completed, and a qualified antibacterial cotton product is output; By collecting pressure and moisture content values ​​and inputting them into a neural network model to predict the softness and fluffiness, the equipment process parameters are dynamically adjusted according to the optimized vacuum value and impregnation time until the expected requirements are met, thus obtaining an antibacterial cotton vacuum impregnation softness and fluffiness treatment process; Obtain the pressure value and moisture content value of the antibacterial cotton as input data of the neural network model; The softness and fluffiness of antibacterial cotton are predicted through the trained neural network model; According to the predicted softness and fluffiness, determine the initial values ​​of the process parameters of the vacuum impregnation equipment, including the vacuum value and impregnation time; Start the vacuum impregnation equipment to make the antibacterial cotton soft and fluffy; During the vacuum impregnation process, the pressure and moisture content values ​​are collected in real time and input into the neural network model again; According to the model prediction results, determine whether the softness and fluffiness meet the expected requirements. If not, dynamically adjust the vacuum value and immersion time; Until the softness and fluffiness meet the expected requirements, the soft and fluffy processing of the antibacterial cotton is completed.

7. The method for making a cotton-filled plush toy according to claim 1, wherein: In step S104, the temperature and humidity values ​​of the drying environment are collected by the temperature and humidity sensor, and the collected temperature and humidity values ​​are transmitted to the control system; The moisture content of the cotton after impregnation and dehydration is collected by using a moisture meter, and the moisture content value is transmitted to the control system; The control system determines the adjustment range of hot air temperature and air volume based on the received temperature, humidity and moisture content values ​​through the preset fuzzy control algorithm; If the temperature is lower than 60℃ or higher than 80℃, the humidity exceeds the set threshold, and the moisture content is higher than the set threshold, the control system will issue a warning message and automatically adjust the hot air temperature and air volume according to the deviation value to keep them within a reasonable range; After the hot air temperature and air volume are adjusted, the cotton is dried; During the drying process, the moisture content of cotton is monitored in real time, and the hot air temperature and air volume parameters are dynamically optimized and adjusted according to the moisture content change trend; Image recognition technology is used to detect the bulkiness of dried cotton; By analyzing the surface morphology of cotton, the cotton bulkiness value is calculated; Compare the cotton bulkiness value with the preset threshold. If it is lower than the threshold, adjust the hot air temperature and air volume parameters and continue to dry the cotton until the cotton bulkiness reaches the requirement. Complete the cotton drying process and output qualified dried cotton; Specifically, it also includes: calculating the cotton fluffiness value by analyzing the surface morphological characteristics of the cotton, comparing the cotton fluffiness value with the preset threshold, and if it is lower than the threshold, adjusting the hot air temperature and air volume parameters, and continuing to dry the cotton until the cotton fluffiness meets the requirements, completing the cotton drying process, and outputting qualified dried cotton.

8. The method for making a cotton-filled plush toy according to claim 7, wherein: The cotton surface morphological features are obtained by analyzing and extracting the cotton surface morphological features using image processing technology; Based on the analysis and extraction of cotton surface morphological characteristics, a cotton bulkiness calculation model was constructed, and the cotton bulkiness value was calculated using this model. Obtaining a preset cotton bulkiness threshold, and comparing the calculated cotton bulkiness value with the threshold; If the cotton bulkiness value is lower than the preset threshold, the hot air temperature and air volume adjustment parameters are determined by the fuzzy control algorithm according to the difference between the cotton bulkiness value and the threshold; The determined hot air temperature and air volume adjustment parameters are input into the hot air drying equipment to continue drying the cotton; After a certain period of hot air drying, the cotton surface morphological characteristic data is re-obtained, the cotton bulkiness value is calculated, and it is determined whether it meets the preset threshold requirements; If the cotton fluffiness value reaches the preset threshold, the hot air drying process is stopped and the qualified dried cotton is output; Otherwise, continue to adjust the hot air temperature and air volume parameters and carry out drying treatment until the cotton fluffiness meets the standard.

9. The method for making a cotton-filled plush toy according to claim 1, wherein: In step S105, the shape parameters and size parameters of the toy cavity are obtained through machine vision, and the parameters are passed to the filling path planning module; The filling path planning module uses a path optimization algorithm to generate the motion trajectory of the filling head and the filling amount parameters of each filling point based on the cavity shape and size parameters; The generated filling trajectory and filling amount parameters are passed to the robotic arm control system. The robotic arm controls the filling head to move according to the planned trajectory and speed, and fills cotton at each filling point according to the set filling amount parameters. During the filling process, the pressure sensor collects the filling density distribution value in the cavity every 5 cm and transmits the collected density value to the filling amount adjustment module; The filling amount adjustment module determines whether the filling density is uniform based on the received density distribution value. If the density distribution is uneven, it dynamically adjusts the filling amount parameters of subsequent filling points to keep the filling amount within the range of 15 to 20 grams per cubic decimeter; The dry cotton is continuously transported to the filling device through the air conveying pipe to ensure sufficient cotton supply during the filling process; The filling device fills the cotton into various filling points of the toy cavity until the entire filling process is completed; After the filling is completed, the stuffed toys are inspected for quality and the filling effect is judged to be qualified based on the test results; If the filling quality is unsatisfactory, the toy will be returned for refilling; If the filling quality is qualified, the stuffed toy will be sent to the next step for further processing.

10. The method for making a cotton-filled plush toy according to claim 1, wherein: In step S106, a high-precision image of the suture area is captured by a machine vision system to obtain detailed parameters of the suture position; Based on the captured image data, the image processing algorithm is used to optimize the suture line to ensure the accuracy of the suture path; The automatic sewing machine receives the optimized sewing path data and adjusts the movement trajectory of the sewing needle to accommodate toys of different shapes and sizes; Sewing machines adjust the speed and direction of the needle in real time during the stitching process to cope with the different thickness and elasticity of the material; After stitching is completed, the stitching quality is automatically detected and the stitching results are analyzed through machine learning algorithms to ensure that the stitching quality of each plush toy meets the standards; After a final quality inspection, the finished plush toys are packaged and portioned, ready for sale or distribution.

Images