A soft capsule preparation system and preparation method
By measuring and predicting the rubber thickness before pressing the pill in the soft capsule preparation process and controlling the spray state, the problem of waste of unqualified products in the traditional process is solved, and the consistency of rubber thickness and production efficiency are improved.
Patent Information
- Application Number
- CN202411293576.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-14
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2044-09-14
AI Technical Summary
In the existing soft capsule preparation process, quality testing is carried out after pressing the pill, resulting in unqualified capsules being unable to be recycled, resulting in waste of raw materials and increased production costs.
Before pressing the pill, the rubber thickness is measured and predicted by the rubber thickness detection module to generate spray control data, regulate the opening or closing state of the spray inside the spray to achieve the pre-control of rubber quality.
It significantly reduces the generation of unqualified products, avoids waste of resources, ensures consistency of rubber thickness, and saves raw materials and energy.
Smart Images

Figure CN118806599B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of soft capsule preparation, and in particular to a soft capsule preparation system and a preparation method. Background Art
[0002] In the existing soft capsule preparation process, the formed capsules are usually tested for quality after the pill pressing process is completed to screen out products that meet the specification requirements. However, this post-testing method has obvious limitations: once unqualified capsules are found, since the soft capsules have been encapsulated, not only can the injected drugs or nutrients not be recovered, but additional resources are required to dispose of the unqualified products, resulting in material waste and increased production costs. Summary of the invention
[0003] In order to solve at least one of the technical problems mentioned above, the present invention provides a soft capsule preparation system to solve the problem that in traditional processes, quality inspection is usually carried out only after the soft capsule is pressed and formed. If it is found that the thickness of the rubber is not up to standard, the unqualified products can only be discarded, resulting in a waste of raw materials.
[0004] In one aspect, a soft capsule preparation system is provided, the system comprising:
[0005] The rubber thickness detection module is used to measure the thickness of the first rubber and the second rubber before pill pressing, and obtain the first rubber thickness data and the second rubber thickness data respectively;
[0006] A data processing module, used for receiving the first rubber thickness data and the second rubber thickness data, and obtaining spray control data according to the first rubber thickness data and the second rubber thickness data;
[0007] The spray body is used to receive the spray body control data and adjust the opening or closing state of different spray ports inside the spray body according to the spray body control data.
[0008] Preferably, the rubber thickness detection module includes an image acquisition unit and a rubber thickness prediction unit;
[0009] Among them, the image acquisition unit shoots perpendicularly to the plane where the rubber is located to obtain a first rubber image and a second rubber image; the rubber thickness prediction unit is used to input the first rubber image and the second rubber image into a pre-built rubber thickness prediction model to obtain the first rubber thickness data and the second rubber thickness data.
[0010] Preferably, after obtaining the first rubber thickness data and the second rubber thickness data, the method further includes:
[0011] Acquire roller mold specification data, wherein the roller mold specification includes mold cavity size data and mold cavity position data;
[0012] Based on the mold cavity size data and the mold cavity position data, the first rubber thickness data and the second rubber thickness data are subdivided into a plurality of corresponding sub-rubber thickness data.
[0013] Preferably, obtaining the spray control data according to the first rubber thickness data and the second rubber thickness data includes:
[0014] Performing thickness comparison processing on the sub-rubber thickness data corresponding to the first rubber thickness data and the second rubber thickness data one by one to obtain a thickness compliance sequence and a thickness compliance sequence;
[0015] Obtaining the rolling speed of the rolling die;
[0016] Based on the thickness reaching standard or failing standard sequence and the rolling speed of the rolling die, the opening and closing sequence of multiple sub-nozzles of the spray body is calculated to obtain the spray body control data.
[0017] Preferably, the step of performing thickness comparison processing on the sub-rubber thickness data corresponding to the first rubber thickness data and the second rubber thickness data one by one to obtain a thickness compliance sequence and a thickness non-compliance sequence comprises:
[0018] Based on the first rubber thickness data and the second rubber thickness data, a plurality of pill pressing thickness data pairs are obtained;
[0019] Determine whether all thickness data of the two data of the pill pressing thickness data pair are within a preset thickness range, if not, generate thickness non-standard data, if yes, calculate the difference in average thickness of the two data of the pill pressing thickness data pair, when the difference is greater than a preset thickness error threshold, generate thickness non-standard data, when the difference is less than or equal to the preset thickness error threshold, generate thickness standard data;
[0020] The thickness non-standard data and the thickness standard data are combined to generate the thickness standard-compliant and non-standard-compliant sequence.
[0021] Preferably, the step of inputting the first rubber image and the second rubber image into a pre-built rubber thickness prediction model to obtain the first rubber thickness data and the second rubber thickness data comprises:
[0022] grayscale the first rubber image and the second rubber image to obtain first rubber grayscale data and second rubber grayscale data;
[0023] The first rubber grayscale data and the second rubber grayscale data are input into the rubber thickness prediction model to obtain the first rubber thickness data and the second rubber thickness data.
[0024] Preferably, the system further comprises a model building module for building a rubber thickness prediction model, and the specific building process comprises:
[0025] Obtain the actual measured data of rubber thickness under the current preparation parameters and its corresponding grayscale image data from the historical database;
[0026] Performing cleaning, standardization, outlier and image enhancement preprocessing on the rubber thickness measured data and the grayscale image data;
[0027] Fit the preprocessed results to construct the relationship curve between rubber thickness and grayscale;
[0028] Using an independent validation set to evaluate the relationship curve between rubber thickness and grayscale and adjusting parameters until the preset evaluation criteria are met;
[0029] The rubber thickness and grayscale relationship curve is determined as the rubber thickness prediction model.
[0030] Preferably, regulating the opening or closing state of different nozzles inside the spray body according to the spray body control data includes:
[0031] Based on the thickness standard or non-standard sequence, monitor whether the rubber thickness corresponding to different nozzles of the lower nozzle body in the shot pressing area meets the standard. If the thickness does not meet the standard data, control the corresponding nozzle to open; if the thickness meets the standard data, control the corresponding nozzle to close; and control the opening time and closing time of the nozzle according to the rolling speed of the rolling die.
[0032] On the other hand, a method for preparing a soft capsule is provided, the method comprising:
[0033] Measuring a first rubber image and a second rubber image of the rubber before pill pressing;
[0034] The first rubber image and the second rubber image are grayed to obtain first rubber gray data and second rubber gray data;
[0035] Inputting the first rubber grayscale data and the second rubber grayscale data into a pre-built rubber thickness prediction model to obtain first rubber thickness data and second rubber thickness data;
[0036] Obtaining spray control data according to the first rubber thickness data and the second rubber thickness data;
[0037] The opening or closing state of different nozzles inside the spray body is regulated according to the spray body control data.
[0038] Preferably, before measuring the first rubber image and the second rubber image of the rubber before pill pressing, the method further comprises:
[0039] Obtain the actual measured data of rubber thickness under the current preparation parameters and its corresponding grayscale image data from the historical database;
[0040] Performing cleaning, standardization, outlier and image enhancement preprocessing on the rubber thickness measured data and the grayscale image data;
[0041] Fit the preprocessed results to construct the relationship curve between rubber thickness and grayscale;
[0042] Using an independent validation set to evaluate the relationship curve between rubber thickness and grayscale and adjusting parameters until the preset evaluation criteria are met;
[0043] The rubber thickness and grayscale relationship curve is determined as the rubber thickness prediction model.
[0044] The beneficial effects of the present invention are:
[0045] (1) In the conventional process, quality inspection is usually performed after the soft capsule is pressed and formed. If the thickness of the rubber skin is found to be substandard, the unqualified product can only be discarded, resulting in a waste of raw materials. However, the present invention measures and predicts the thickness of the rubber skin before pressing the pills, thereby achieving pre-control of the rubber skin quality, avoiding the production of a large number of unqualified products, and significantly reducing the waste of resources caused by the later removal of unqualified products;
[0046] (2) The traditional process lacks real-time monitoring and pre-control of the rubber thickness, resulting in the inability to adjust the process parameters in time according to the actual thickness during the production process. This method realizes precise regulation of the working state of the sprayer by dynamically generating spray control data through real-time monitoring of the thickness data of the first and second rubber layers. This active adjustment mechanism ensures the consistency of the thickness of the double-layer rubber, avoids secondary processing or product scrapping due to excessive thickness deviation, and effectively saves raw materials and energy.
[0047] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure. BRIEF DESCRIPTION OF THE DRAWINGS
[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the background technology, the drawings required for use in the embodiments of the present invention or the background technology will be described below.
[0049] The drawings herein are incorporated into the specification and constitute a part of the specification. These drawings illustrate embodiments consistent with the present disclosure and, together with the specification, are used to illustrate the technical solutions disclosed in the present invention.
[0050] Figure 1 A schematic diagram of the structure of a soft capsule preparation system provided by an embodiment of the present invention;
[0051] Figure 2 A schematic diagram of the structure of a rubber thickness detection module provided in an embodiment of the present invention;
[0052] Figure 3 The present invention provides a schematic flow chart of a method for preparing a soft capsule according to an embodiment of the present invention. DETAILED DESCRIPTION
[0053] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0054] The terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or devices.
[0055] The term "and / or" herein is only a description of the association relationship of the associated objects, indicating that there may be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the term "at least one" herein represents any combination of at least two of any one or more of a plurality of. For example, including at least one of A, B, and C can represent including any one or more elements selected from the set consisting of A, B, and C.
[0056] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.
[0057] In addition, in order to better illustrate the present invention, numerous specific details are provided in the following specific embodiments. It should be understood by those skilled in the art that the present invention can be implemented without certain specific details. In some examples, methods, means, components and circuits well known to those skilled in the art are not described in detail in order to highlight the subject matter of the present invention.
[0058] At present, in the existing soft capsule preparation, there is a common problem that the capsule quality is only tested after pilling, and the pre-control of the rubber thickness is ignored, resulting in a waste of resources.
[0059] Example 1 provides a soft capsule preparation system, referring to Figure 1 , the system includes:
[0060] The rubber thickness detection module 100 is used to measure the thickness of the first rubber and the second rubber before pilling, and obtain the first rubber thickness data and the second rubber thickness data respectively;
[0061] The data processing module 200 is used to receive the first rubber thickness data and the second rubber thickness data, and obtain the spray control data according to the first rubber thickness data and the second rubber thickness data;
[0062] The spray body 300 is used to receive the spray body control data and adjust the opening or closing state of different spray ports inside the spray body according to the spray body control data.
[0063] Preferably, reference Figure 2 , the rubber thickness detection module 100 includes an image acquisition unit 110 and a rubber thickness prediction unit 120;
[0064] The image acquisition unit 110 shoots perpendicularly to the plane where the rubber is located to obtain a first rubber image and a second rubber image; the rubber thickness prediction unit 120 is used to input the first rubber image and the second rubber image into a pre-built rubber thickness prediction model to obtain first rubber thickness data and second rubber thickness data.
[0065] In this embodiment, the image acquisition unit 110 is set at an appropriate position in front of the pill press, and uses a high-definition industrial camera to shoot perpendicular to the rubber conveying direction. The camera is equipped with a precise servo control system to ensure that it always remains perpendicular to the plane where the rubber is located, so as to obtain clear and distortion-free first rubber image and second rubber image. The image acquisition frequency matches the rubber conveying speed to ensure that each frame of the image corresponds to an accurate rubber position.
[0066] The rubber thickness prediction unit 120 has a pre-trained rubber thickness prediction model built in. The model is based on deep learning technology and is trained using a large number of annotated rubber images and their corresponding thickness data. It has the ability to infer the actual thickness from visual features such as rubber surface texture, edge contour or brightness. After the image acquisition unit 110 acquires the first rubber image and the second rubber image, the rubber thickness prediction unit 120 uses the two images as input, calculates the corresponding first rubber thickness data and second rubber thickness data through the prediction model, and sends the prediction model calculation results to the data processing module 200 in real time.
[0067] The data processing module 200 receives the first rubber thickness data and the second rubber thickness data transmitted by the rubber thickness detection module 100, and performs comparative analysis. If the thickness difference between the two exceeds the preset tolerance range, the data processing module 200 will generate the spray control data according to the difference value and the preset compensation strategy. The data includes the opening or closing instructions for the different nozzles inside the spray body 300, as well as the parameters such as the spray pressure and speed that may need to be adjusted, in order to accurately control the rubber spraying process and ensure that the final liquid medicine and the rubber pressed pill are tightly combined.
[0068] The spray body 300 is provided with a plurality of independently controllable nozzles, each of which is equipped with a solenoid valve or a servo drive device, and can quickly respond to the spray body control data sent by the data processing module 200. When receiving an opening command, the corresponding nozzle is immediately opened, and the glue liquid is sprayed according to the set spraying parameters to form a rubber skin; conversely, when receiving a closing command, the nozzle is quickly closed and the glue liquid spraying is stopped. By accurately controlling the state and spraying parameters of each nozzle, the spray body can dynamically adapt to the changing requirements of the rubber skin thickness, ensuring that the soft capsules produced have an ideal double-layer rubber skin thickness.
[0069] In this embodiment, the image recognition technology is combined with the deep learning prediction model. The system can directly predict the thickness information from the lightness and darkness of the rubber image, avoiding the errors and physical interference to the rubber caused by traditional contact measurement, and improving the accuracy and reliability of rubber thickness measurement. At the same time, the use of artificial intelligence technology to replace manual visual inspection and manual adjustment greatly reduces the dependence on manpower, improves production efficiency, and also reduces product quality fluctuations caused by human factors. In addition, controlling the opening and closing states of different nozzles through accurate rubber thickness data helps to reduce the waste of glue, reduce material consumption, save production costs, and is also beneficial to environmental protection.
[0070] Preferably, after obtaining the first rubber thickness data and the second rubber thickness data, the method further includes:
[0071] Obtaining roller mold specification data, the roller mold specification includes mold cavity size data and mold cavity position data;
[0072] Based on the mold cavity size data and the mold cavity position data, the first rubber thickness data and the second rubber thickness data are subdivided into a plurality of corresponding sub-rubber thickness data.
[0073] In this embodiment, the specification data of the currently used rolling mold is obtained by connecting to the rolling mold database or directly reading the electronic tag on the rolling mold. The rolling mold specification data includes mold cavity size data (such as the diameter and depth of a single mold cavity) and mold cavity position data (such as the distribution position of the mold cavity on the circumference of the rolling mold). The data processing module 200 performs subdivision processing on the first rubber thickness data and the second rubber thickness data that have been obtained according to the obtained rolling mold specification data, as follows:
[0074] According to the cavity position data, areas corresponding to each cavity are divided on the first rubber image and the second rubber image. These areas are accurately divided according to the circumferential arrangement order of the roller mold to ensure one-to-one correspondence with the actual cavity position. For each cavity area, the corresponding first rubber image and second rubber image sub-area are extracted and sent to the rubber thickness prediction unit 120 for separate calculation. The prediction unit re-predicts the first sub-rubber thickness data and the second sub-rubber thickness data corresponding to the cavity according to the sub-area image. After the above processing, the original overall rubber thickness data is subdivided into sub-rubber thickness data corresponding to each cavity.
[0075] In this embodiment, by subdividing the thickness data of the first rubber and the second rubber at the mold cavity level, the system achieves independent and precise control of the thickness of the soft capsule rubber in each mold cavity, eliminating the problem of uniform overall thickness of the rolling mold but possible thickness deviation between local mold cavities, thereby further improving the consistency of soft capsule product quality.
[0076] Preferably, the spray control data is obtained according to the first rubber thickness data and the second rubber thickness data, including:
[0077] Perform thickness comparison processing on the sub-rubber thickness data corresponding to the first rubber thickness data and the second rubber thickness data one by one to obtain a thickness sequence that meets the standard and fails to meet the standard;
[0078] Get the rolling speed of the rolling die;
[0079] Based on the thickness compliance and non-compliance sequence and the rolling speed of the rolling die, the opening and closing sequence of multiple sub-nozzles of the nozzle body is calculated to obtain the nozzle control data.
[0080] In this embodiment, the data processing module 200 compares the first rubber thickness data and the second rubber thickness data obtained after the subdivision processing one by one. During the comparison process, a preset thickness threshold range is set as the standard. For each sub-rubber thickness data, if its value falls within the threshold range, it is marked as "thickness meets the standard", otherwise it is marked as "thickness does not meet the standard". In this way, a thickness meeting standard and non-meeting standard sequence is formed, which clearly reflects whether the thickness of each mold cavity sub-rubber meets the set standard. At the same time, through the communication interface or sensor monitoring with the roller mold drive motor, the current rolling speed of the roller mold is obtained in real time. The rolling speed of the roller mold is an important factor affecting the timing and duration of glue injection, and is closely related to the control of the spray body. The data processing module 200 calculates the opening and closing sequence of multiple sub-nozzles of the spray body according to the thickness meeting standard and non-meeting standard sequence and the rolling speed of the roller mold. The specific steps are as follows:
[0081] (1) Establish an association between the thickness data of each sub-rubber and the sub-nozzle group in the corresponding mold cavity area to ensure that the control command can accurately act on the corresponding nozzle; (2) For sub-rubbers whose thickness does not meet the standard, the system decides not to open the corresponding nozzle to avoid spraying glue when the roller mold passes through the mold cavity, thereby preventing the production of products with substandard thickness; (3) For sub-rubbers whose thickness does not meet the standard, the system decides not to open the corresponding nozzle to avoid spraying glue when the roller mold passes through the mold cavity, thereby preventing the production of products with substandard thickness. The specific steps are as follows:
[0082] A) When the thickness difference is small (close to the standard thickness), the minimum duration of opening the corresponding sub-nozzle is calculated according to the rolling speed of the roller mold and the preset spraying parameters (such as the amount of glue required per unit time) to maintain the target thickness without over-spraying;
[0083] B) When the thickness difference is moderate, adjust the injection rate to adapt to the target thickness. For slightly thicker sub-rubbers, reduce the injection rate to reduce the glue output per unit time; for slightly thinner sub-rubbers, increase the injection rate appropriately to increase the glue supply;
[0084] C) When the thickness difference is large, the glue spraying state can be finely controlled by changing the spraying pressure. For thicker sub-rubbers, reduce the spraying pressure to reduce the amount of glue attached; for thinner sub-rubbers, appropriately increase the spraying pressure to enhance the coating effect of the glue;
[0085] D) Integrate the information of the opening and closing status, opening time, injection rate or injection pressure of all sub-nozzles into an ordered sequence, namely the nozzle control data. This data set contains the precise control instructions of each sub-nozzle during the rolling process of the roller mold, which is used to guide the nozzle to adjust the injection behavior in real time.
[0086] In this embodiment, by comparing the thickness data of the sub-rubbers one by one and generating the spray control data accordingly, the system can accurately control the thickness of the soft capsule rubber in each mold cavity, ensure that the product meets the preset thickness standard, and improve the consistency of product quality. Furthermore, for the rubber pairs whose thickness does not meet the standard, the system chooses not to open the corresponding nozzle, avoiding the waste of glue liquid; for the rubber pairs whose thickness meets the standard, the injection parameters are finely adjusted according to the thickness difference value and the rolling speed of the rolling die, optimizing the use of resources, reducing energy consumption, and helping to improve production efficiency and reduce costs.
[0087] Preferably, thickness comparison processing is performed one by one on the sub-rubber thickness data corresponding to the first rubber thickness data and the second rubber thickness data to obtain a thickness standard-reaching and non-standard-reaching sequence, including:
[0088] Based on the first rubber thickness data and the second rubber thickness data, a plurality of pill pressing thickness data pairs are obtained;
[0089] Determine whether all thickness data of the two data of the pill pressing thickness data pair are within the preset thickness range, if not, generate thickness non-standard data, if so, calculate the difference in average thickness of the two data of the pill pressing thickness data pair, when the difference is greater than the preset thickness error threshold, generate thickness non-standard data, when the difference is less than or equal to the preset thickness error threshold, generate thickness standard data;
[0090] Combine the thickness non-standard data and thickness standard data to generate a thickness standard and non-standard sequence.
[0091] In this embodiment, the first rubber thickness data and the second rubber thickness data are paired according to the corresponding cavity positions to form a plurality of pill pressing thickness data pairs. Each data pair contains the sub-rubber thickness data of the corresponding areas of the two rubbers in the same cavity, which is used to evaluate the thickness uniformity of the rubber pair in the cavity. For each pill pressing thickness data pair, check whether its two data (i.e., the thickness values of the sub-rubbers corresponding to the first rubber and the second rubber) are both within the preset thickness range. The preset thickness range is usually the standard value of the rubber thickness required by the soft capsule product design and its allowed deviation value. If any data is not within the range, it is determined that the pill pressing thickness data pair is not up to standard thickness, and the corresponding thickness not up to standard data record is generated. For the case where the thickness data are all within the preset range, the average thickness difference of the two data in the pill pressing thickness data pair is calculated. If the average thickness difference is greater than the preset thickness error threshold, it means that although the single thickness value is within the allowable range, the thickness difference of the rubber pair is too large, which may lead to uneven quality of the soft capsule. At this time, the thickness not up to standard data is generated. On the contrary, if the average thickness difference is less than or equal to the preset thickness error threshold, it means that the thickness difference of the rubber pair is within the acceptable range, and the thickness compliance data is generated. All thickness non-compliance data and thickness compliance data are merged according to the corresponding cavity position sequence to form a complete thickness compliance and non-compliance sequence. This sequence clearly reflects whether the thickness of the rubber pair in each cavity meets the established quality standards, providing a key basis for subsequent spray control.
[0092] It should be noted that, in this embodiment, by pairing and analyzing the thickness data of the first rubber and the second rubber one by one, the system not only pays attention to whether the thickness of a single rubber meets the standard, but also evaluates the thickness difference between the rubber pairs, thereby realizing a refined evaluation of the uniformity of the thickness of the soft capsule rubber, which helps to improve the consistency of product quality. Furthermore, by setting a double judgment standard (a single thickness value is within a preset range and the thickness difference is within an error threshold), it is ensured that only rubber pairs with uniform thickness and meeting strict standards can be identified as meeting the standard, strictly controlling product quality and reducing soft capsule performance defects caused by uneven thickness.
[0093] Preferably, the first rubber image and the second rubber image are input into a pre-built rubber thickness prediction model to obtain the first rubber thickness data and the second rubber thickness data, including:
[0094] grayscale the first rubber image and the second rubber image to obtain first rubber grayscale data and second rubber grayscale data;
[0095] The first rubber grayscale data and the second rubber grayscale data are input into a rubber thickness prediction model to obtain the first rubber thickness data and the second rubber thickness data.
[0096] In this embodiment, the collected first rubber image and second rubber image are grayed. By converting the color image into a single grayscale channel, the interference of color information is eliminated, so that the image only retains the brightness information, which is convenient for the subsequent model to accurately extract the thickness feature. After processing, the first rubber grayscale data and the second rubber grayscale data are obtained. The obtained first rubber grayscale data and the second rubber grayscale data are respectively used as inputs and sent to the pre-trained rubber thickness prediction model. The model is based on deep learning technology, such as convolutional neural network (CNN). By learning the mapping relationship between a large number of known rubber images and their corresponding thickness measurement values, it has the ability to accurately infer the rubber thickness from the grayscale image. The rubber thickness prediction model performs feature extraction, nonlinear transformation and multi-level abstraction on the input grayscale data, and finally outputs the corresponding first rubber thickness data and second rubber thickness data. These data accurately reflect the actual thickness of the two pieces of rubber in the area covered by the image in numerical form, and can be used for subsequent quality control, process optimization and other links.
[0097] This embodiment combines image processing with deep learning technology to obtain the thickness data of the rubber skin in a non-contact and non-destructive manner without destroying the soft capsule. Compared with the traditional mechanical measurement method, it significantly reduces the physical interference and potential damage risk to the product. Compared with manual measurement or traditional equipment detection, the method of combining image recognition with prediction model can quickly process a large amount of image data, realize online and real-time thickness monitoring, greatly improve the detection efficiency in the production process, and adapt to the real-time monitoring needs of high-speed production lines.
[0098] Preferably, the system further includes a model building module for building a rubber thickness prediction model, and the specific building process includes:
[0099] Obtain the actual measured data of rubber thickness under the current preparation parameters and its corresponding grayscale image data from the historical database;
[0100] Cleaning, standardization, outlier and image enhancement preprocessing are performed on the measured data of rubber thickness and grayscale image data;
[0101] Fit the preprocessed results to construct the relationship curve between rubber thickness and grayscale;
[0102] Use an independent validation set to evaluate the relationship curve between rubber thickness and grayscale and adjust the parameters until the preset evaluation criteria are met;
[0103] The relationship curve between rubber thickness and grayscale is determined as the rubber thickness prediction model.
[0104] In this embodiment, the model building module collects the measured data of rubber thickness under the current preparation parameters and its corresponding grayscale image data from the historical database, performs data cleaning, standardization, outlier processing and image enhancement preprocessing, and then constructs the relationship curve between rubber thickness and grayscale based on the preprocessing results. The model is evaluated and adjusted through an independent validation set until the preset evaluation criteria are met. The final determined relationship curve is used as a rubber thickness prediction model for real-time thickness prediction of the soft capsule preparation system, realizing data-driven, customized, high-precision and adaptive optimization-capable rubber thickness prediction, and promoting the intelligent upgrade of the production line.
[0105] Preferably, the opening or closing state of different nozzles inside the spray body is regulated according to the spray body control data, including:
[0106] Based on the thickness standard and non-standard sequence, monitor whether the rubber thickness corresponding to different nozzles of the lower nozzle body in the shot pressing area meets the standard. If the thickness does not meet the standard, control the corresponding nozzle to open; if the thickness meets the standard, control the corresponding nozzle to close, and control the opening and closing time of the nozzle according to the rolling speed of the roller mold.
[0107] By precisely controlling the opening and closing states of the nozzle and the injection time, the system of this embodiment can effectively correct the rubber thickness deviation, ensure that the double-layer rubber thickness of each soft capsule meets the preset standard, and significantly improve the consistency and stability of product quality.
[0108] Example 2 provides a method for preparing a soft capsule, referring to Figure 3 , methods include:
[0109] S100, measuring a first rubber image and a second rubber image of the rubber before pill pressing;
[0110] S200, graying the first rubber image and the second rubber image to obtain first rubber grayscale data and second rubber grayscale data;
[0111] S300, inputting the first rubber grayscale data and the second rubber grayscale data into a pre-built rubber thickness prediction model to obtain the first rubber thickness data and the second rubber thickness data;
[0112] S400, obtaining spray control data according to the first rubber thickness data and the second rubber thickness data;
[0113] S500, regulating the opening or closing state of different nozzles inside the spray body according to the spray body control data.
[0114] In a possible embodiment, before pressing the pellets, a high-definition industrial camera is used to shoot the rubber to obtain a first rubber image and a second rubber image, respectively. The shooting angle, distance and lighting conditions are carefully set to ensure that the image is clear and distortion-free, and can accurately reflect the surface characteristics of the rubber.
[0115] The first rubber image and the second rubber image are grayed to convert the colorful color image into a single gray channel image, namely the first rubber gray data and the second rubber gray data. The gray image only retains the brightness information, which is conducive to the subsequent model to efficiently extract features related to the rubber thickness.
[0116] The grayscale data of the first rubber and the grayscale data of the second rubber are input into a pre-built rubber thickness prediction model. The model is based on deep learning technology, such as convolutional neural network (CNN), and after a large number of sample training, it can accurately infer the rubber thickness from the grayscale image. After the model is processed, the first rubber thickness data and the second rubber thickness data are output, indicating the actual thickness of the two pieces of rubber in the area covered by the image.
[0117] Based on the obtained first rubber thickness data and second rubber thickness data, the system calculates the spray control data. Specifically, the thickness data of the two rubbers are compared to determine whether there is a thickness deviation. If the deviation exceeds the preset threshold, the spray control data is generated based on the deviation value and the preset compensation strategy. These control data not only include the opening or closing instructions of each nozzle, but may also include specific values for adjusting parameters such as injection pressure and speed, aiming to accurately control the subsequent rubber injection process and ensure the consistency of the thickness of the double rubber.
[0118] The spray body receives the spray body control data and adjusts the opening or closing state of different nozzles inside the spray body according to the instructions. When receiving the opening instruction, the corresponding nozzle is immediately started to spray the glue according to the set spray parameters to form a rubber skin; conversely, when receiving the closing instruction, the nozzle is quickly closed and the glue spraying is stopped. By accurately controlling the state and spraying parameters of each nozzle, the spray body can dynamically adapt to the changing requirements of the rubber skin thickness, ensuring that the soft capsules produced have an ideal double-layer rubber skin thickness.
[0119] In general, the existing methods usually perform quality inspection after soft capsules are pressed and formed. If the thickness of the rubber is found to be substandard, the unqualified products can only be discarded, resulting in a waste of raw materials. However, this embodiment achieves pre-control of the rubber quality by measuring and predicting the thickness of the rubber before pressing, avoiding the production of a large number of unqualified products and significantly reducing the waste of resources caused by the later removal of unqualified products.
[0120] Preferably, before measuring the first rubber image and the second rubber image of the rubber before pill pressing, the method further includes:
[0121] Obtain the actual measured data of rubber thickness under the current preparation parameters and its corresponding grayscale image data from the historical database;
[0122] Cleaning, standardization, outlier and image enhancement preprocessing are performed on the measured data of rubber thickness and grayscale image data;
[0123] Fit the preprocessed results to construct the relationship curve between rubber thickness and grayscale;
[0124] Use an independent validation set to evaluate the relationship curve between rubber thickness and grayscale and adjust the parameters until the preset evaluation criteria are met;
[0125] The relationship curve between rubber thickness and grayscale is determined as the rubber thickness prediction model.
[0126] The foregoing is merely a specific embodiment of the present invention, which enables those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features claimed herein.
Claims
1. A soft capsule preparation system, characterized in that: The system comprises: The rubber thickness detection module is used to measure the thickness of the first rubber and the second rubber before pill pressing, and obtain the first rubber thickness data and the second rubber thickness data respectively; A data processing module, used for receiving the first rubber thickness data and the second rubber thickness data, and obtaining spray control data according to the first rubber thickness data and the second rubber thickness data; A spray body, used to receive the spray body control data and adjust the opening or closing state of different spray ports inside the spray body according to the spray body control data; The rubber thickness detection module includes an image acquisition unit and a rubber thickness prediction unit; the image acquisition unit takes pictures perpendicular to the plane where the rubber is located to obtain a first rubber image and a second rubber image; the rubber thickness prediction unit is used to input the first rubber image and the second rubber image into a pre-built rubber thickness prediction model to obtain the first rubber thickness data and the second rubber thickness data; Furthermore, after obtaining the first rubber thickness data and the second rubber thickness data, the method further includes: Acquire rolling mold specification data, wherein the rolling mold specification includes mold cavity size data and mold cavity position data; Based on the mold cavity size data and the mold cavity position data, the first rubber thickness data and the second rubber thickness data are subdivided into a plurality of corresponding sub-rubber thickness data; Wherein, obtaining the spray control data according to the first rubber thickness data and the second rubber thickness data includes: Performing thickness comparison processing on the sub-rubber thickness data corresponding to the first rubber thickness data and the second rubber thickness data one by one to obtain a thickness sequence that meets the standard and fails to meet the standard; Get the rolling speed of the rolling die; Based on the thickness reaching standard or failing standard sequence and the rolling speed of the rolling die, the opening and closing sequence of the plurality of sub-nozzles of the spray body is calculated to obtain the spray body control data; Further, the step of performing thickness comparison processing on the sub-rubber thickness data corresponding to the first rubber thickness data and the second rubber thickness data one by one to obtain a thickness compliance sequence and a thickness non-compliance sequence includes: Based on the first rubber thickness data and the second rubber thickness data, a plurality of pill pressing thickness data pairs are obtained; Determine whether all thickness data of the two data of the pill pressing thickness data pair are within a preset thickness range, if not, generate thickness non-standard data, if yes, calculate the difference in average thickness of the two data of the pill pressing thickness data pair, when the difference is greater than a preset thickness error threshold, generate thickness non-standard data, when the difference is less than or equal to the preset thickness error threshold, generate thickness standard data; Combining the thickness non-standard data and the thickness standard data to generate the thickness standard-compliant and non-standard-compliant sequence; Further, the step of inputting the first rubber image and the second rubber image into a pre-built rubber thickness prediction model to obtain the first rubber thickness data and the second rubber thickness data includes: grayscale the first rubber image and the second rubber image to obtain first rubber grayscale data and second rubber grayscale data; The first rubber grayscale data and the second rubber grayscale data are input into the rubber thickness prediction model to obtain the first rubber thickness data and the second rubber thickness data.
2. A soft capsule preparation system according to claim 1, characterized in that: The system also includes a model building module for building a rubber thickness prediction model. The specific building process includes: Obtain the actual measured data of rubber thickness under the current preparation parameters and its corresponding grayscale image data from the historical database; Performing cleaning, standardization, outlier and image enhancement preprocessing on the rubber thickness measured data and the grayscale image data; Fit the preprocessed results to construct the relationship curve between rubber thickness and grayscale; Using an independent validation set to evaluate the relationship curve between rubber thickness and grayscale and adjusting parameters until the preset evaluation criteria are met; The rubber thickness and grayscale relationship curve is determined as the rubber thickness prediction model.
3. A soft capsule preparation system according to claim 1, characterized in that: The method of regulating the opening or closing state of different nozzles in the spray body according to the spray body control data includes: Based on the thickness standard or non-standard sequence, monitor whether the rubber thickness corresponding to different nozzles of the lower nozzle body in the shot pressing area meets the standard. If the thickness does not meet the standard data, control the corresponding nozzle to open; if the thickness meets the standard data, control the corresponding nozzle to close; and control the opening time and closing time of the nozzle according to the rolling speed of the rolling die.
4. A method for preparing soft capsules, applied to the soft capsule preparation system according to any one of claims 1 to 3, characterized in that: The method comprises: Measuring a first rubber image and a second rubber image of the rubber before pill pressing; The first rubber image and the second rubber image are grayed to obtain first rubber gray data and second rubber gray data; Inputting the first rubber grayscale data and the second rubber grayscale data into a pre-built rubber thickness prediction model to obtain first rubber thickness data and second rubber thickness data; Obtaining spray control data according to the first rubber thickness data and the second rubber thickness data; The opening or closing state of different nozzles inside the spray body is regulated according to the spray body control data.
5. A method for preparing soft capsules according to claim 4, characterized in that: Before measuring the first rubber image and the second rubber image of the rubber before pill pressing, the method further comprises: Obtain the actual measured data of rubber thickness under the current preparation parameters and its corresponding grayscale image data from the historical database; Performing cleaning, standardization, outlier and image enhancement preprocessing on the rubber thickness measured data and the grayscale image data; Fit the preprocessed results to construct the relationship curve between rubber thickness and grayscale; Using an independent validation set to evaluate the relationship curve between rubber thickness and grayscale and adjusting parameters until the preset evaluation criteria are met; The rubber thickness and grayscale relationship curve is determined as the rubber thickness prediction model.
Citation Information
Patent Citations
Automatic soft capsule production equipment and speed regulation method
CN115366385A
Material layer thickness control system and method based on ore grade detection device
CN117191824A
Capsule shell thickness detection device
CN204470139U
Adjusting device is measured from dynamic testing to rubber thickness
CN208206078U
Automatic detection and thickness adjustment device for soft capsule rubber
CN210932798U
