Probiotic tablet production traceability method and system

Abstract

The invention relates to the field of image processing, in particular to a probiotic tablet production traceability method and system, and the method comprises the steps: building a traceability matrix through collecting the data of temperature and humidity, pH, equipment number, pressure and the like in each link of strain selection, culture amplification, fermentation, drying, formulation, forming, coating, quality inspection, packaging and the like; the method comprises the following steps of: acquiring a bacterial colony image through high-definition camera shooting, extracting geometric and textural features of the bacterial colony through preprocessing, watershed and dynamic threshold algorithms, forming supplementary parameters, finally splicing all data, generating a unique traceability code through SHA-256 Hash, spraying the unique traceability code on a packaging unit, and meanwhile, performing uplink evidence storage, thereby realizing full-process product traceability. And quality safety is guaranteed. According to the method, data of each link and a bacterial colony image are integrated, features are extracted through a dynamic threshold watershed, and a traceability code is generated through SHA-256 encryption.

Description

Technical Field

[0001] The present invention relates to the field of image processing, and particularly to a method and system for tracing the production of probiotic tablets. Background Art

[0002] With the continuous improvement of people's requirements for health and food safety, probiotic tablets, as a functional food with the functions of regulating intestinal flora and enhancing immunity, are receiving extensive attention from consumers and regulatory authorities. The production process of probiotic tablets involves multiple links. Therefore, it is particularly important to establish a full-process, real-time, and accurate traceability system. However, traditional traceability methods mainly focus on the collection of single production parameters such as temperature and humidity, pH value, equipment number, and pressure, and there is less research on the real-time dynamic monitoring of the growth state of colonies during the production process. Existing technologies usually only adopt fixed sampling frequencies and static data acquisition means, ignoring the rich information contained in colony images, such as the geometric morphology, edge texture, and growth rate changes of colonies. Due to the complex situations such as uneven colony distribution and noise interference during the amplification culture of probiotics, single parameter collection cannot comprehensively reflect the minute changes during the production process, thereby affecting the accurate evaluation of the quality of the final product. Thus, how to extract colony features by combining image processing technology on the basis of traditional traceability parameters to achieve comprehensive integration and real-time traceability of data has become a technical problem to be solved urgently.

[0003] Currently, some existing technologies use sensors to monitor production environment parameters in real time and upload the data to the traceability database, but these methods have great deficiencies in image information processing. For example, in the processing of colony images, fixed threshold segmentation algorithms are mostly used, which are prone to incorrect segmentation due to noise or gray level fluctuations, and cannot accurately identify the true boundaries of colonies, resulting in incomplete and inaccurate traceability data. Summary of the Invention

[0004] In view of the above problem of being unable to accurately identify the true boundary of colonies, in the first aspect, the present invention proposes a method for tracing the production of probiotic tablets, including: obtaining relevant parameters of each process during the production of probiotic tablets to obtain a traceability data matrix; collecting colony images at regular intervals during the probiotic pilot-scale amplification process, and using the watershed algorithm to segment the colony images to obtain corresponding colony segmentation images; extracting the geometric features and texture features of multiple colonies in the colony segmentation images, and splicing the geometric features and texture features as supplementary traceability parameters with the traceability data matrix to obtain a complete traceability set; processing the complete traceability set using an encryption algorithm to obtain a traceability code and spraying it on the minimum packaging unit of the probiotic tablets; the watershed algorithm further includes multiplying the basic threshold of the H transformation by an adjustment value to obtain a dynamic threshold, and the adjustment value is obtained through a series of predetermined mathematical operations including logarithmic operations and hyperbolic tangent operations by the growth rate gradient, colony spatial density, and a set adjustment coefficient; obtaining multiple connected regions in the colony image, and using the geometric centers of each connected region as the coordinates of the initial planting points of the colonies; calculating the coordinate variance and coordinate mean of all coordinates, obtaining the Euclidean distance between each coordinate and the coordinate mean, and obtaining the colony spatial density according to the ratio of the Euclidean distance to the coordinate variance; calculating the difference between the edge gradient amplitudes of each colony in the colony image and the colony image at a set time interval before to obtain the growth rate gradient.

[0005] The present invention organically integrates the data of each production process with the geometric and texture features extracted from the pilot-scale amplified colony images of probiotics, accurately segments the images using the dynamic threshold watershed algorithm, and then forms a unique traceability code through SHA-256 encryption, realizing full-process data tracking and encrypted traceability. This method effectively makes up for the deficiencies of traditional technologies that only rely on static production parameters, lack dynamic image monitoring, have incomplete data, and insufficient security, thus greatly improving the accuracy and anti-tampering ability of product traceability.

[0006] Further, the specific calculation method of the dynamic threshold is as follows: ; Where represents the reference threshold of the H transformation; represents the adjustment coefficient; and represent the growth rate gradient and the colony spatial density respectively; represents the hyperbolic tangent function; represents the sign function.

[0007] Further, the specific calculation method of the colony spatial density is as follows: ; Where represents the colony spatial density; represents the estimated number of colonies, specifically the number of independent connected regions in the colony image; ( ) represents the natural exponential function; represents the coordinates of the initial planting point of the th colony; represents the mean value of the coordinates of all initial planting points; represents the Euclidean norm.

[0008] By using the natural exponential function and the calculation of mean and variance, the present invention quantitatively analyzes the initial planting points of colonies, realizes the accurate calculation of the spatial density of colonies, objectively reflects the colony distribution characteristics more than the traditional rough estimation method, and provides accurate data support for subsequent dynamic feature extraction.

[0009] Further, the calculation method of the growth rate gradient is specifically as follows: ; where represents the growth rate gradient of the colony; represents the set time interval; represents the th moment in the colony image at the coordinate the Sobel gradient amplitude at the represents the region in the colony image where the gradient amplitude is greater than the set threshold; represents the Euclidean norm; represents a very small positive number.

[0010] The present invention calculates the growth rate gradient of colonies by using the edge gradient change in images at adjacent time points, effectively captures the dynamic growth of colonies, overcomes the deficiency in the prior art that the growth state of colonies cannot be reflected in real time, and thus realizes the fine monitoring of the dynamic process.

[0011] Further, the relevant parameters of each process include: The strain number in the strain selection process; the temperature, humidity and pH value in the culture and amplification process; the fermentation tank number in the fermentation process; the drying equipment number and the coding of the drying method in the drying and powder making process; the mass ratio of the bacterial powder to each auxiliary material in the formula design and auxiliary material mixing process; the granulation equipment number, tablet pressing equipment number and the pressure of the tablet pressing equipment in the granulation and tablet forming process; the coating equipment number and the coating thickness in the coating process; the viable bacteria count after sampling inspection in the quality inspection process; the packaging equipment number and the packaging date in the packaging and storage process.

[0012] Further, it also includes performing preprocessing operations on the colony image, specifically: converting the colony image to the HSV color space to obtain the HSV colony image; using the Otsu algorithm for threshold segmentation based on the S channel in the HSV colony image to obtain a segmented image; and performing morphological opening on the segmented image.

[0013] Image preprocessing is performed using HSV color space conversion, Otsu algorithm segmentation, and morphological opening, effectively removing noise and highlighting colony features. Compared with traditional unprocessed images, the segmentation results are clearer and more accurate, laying a solid foundation for subsequent feature extraction.

[0014] Further, obtaining multiple connected components in the colony image also includes: using 8-neighborhood connected component analysis on the preprocessed colony image to obtain multiple connected components.

[0015] Further, extracting the geometric and texture features of multiple colonies in the colony segmented image also includes: using the cv2.findContours() function in openCV to extract the contours of each colony and using cv2.contourArea() to calculate the contour area; using cv2.arcLength() to calculate the contour perimeter; using the Sobel operator to extract the edge gradient and calculate the mean value; and using the LBP algorithm to extract the texture features of each colony to obtain the LBP value.

[0016] By comprehensively extracting the colony contour, area, perimeter, edge gradient, and texture features using cv2.findContours, cv2.contourArea, cv2.arcLength, the Sobel operator, and the LBP algorithm, the present invention provides a comprehensive and quantitative description of colony morphology. Compared with traditional techniques that only collect a single parameter, the uniqueness and discrimination accuracy of traceability data are significantly enhanced.

[0017] Further, the encryption algorithm is the SHA-256 hash encryption algorithm.

[0018] In a second aspect, the present invention provides a probiotic tablet production traceability system, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the probiotic tablet production traceability method of the present invention is implemented.

[0019] The technical effects of the present invention are: The present invention integrates the traditional parameter data of each key production link with the dynamic features extracted from the colony images in the pilot-scale amplification stage, accurately segments the images through the dynamic threshold watershed algorithm, and calculates the spatial density and growth rate gradient of the colonies using a quantitative method, achieving a comprehensive capture of colony features. Combining SHA-256 encryption and blockchain evidence storage technology, a whole-process, dynamic, and tamper-proof probiotic tablet traceability system is constructed. Compared with the existing traceability solutions that only rely on static data, the present invention greatly improves the integrity, security, and traceability accuracy of the data, providing a new technical solution for the quality supervision of functional foods. Brief Description of the Drawings

[0020] By referring to the following detailed description with reference to the accompanying drawings, the above and other objects, features, and advantages of the exemplary embodiments of the present invention will become readily understandable. In the drawings, several embodiments of the present invention are shown in an exemplary rather than restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein: Figure 1 is a schematic flowchart showing the probiotic tablet production traceability method in an embodiment of the present invention; Figure 2 is a schematic block diagram showing the structure of the probiotic tablet production traceability system in an embodiment of the present invention. Detailed Embodiments

[0021] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts fall within the scope of the present invention.

[0022] Next, the specific embodiments of the present invention will be described in detail in conjunction with the accompanying drawings.

[0023] An embodiment of a probiotic tablet production traceability method: As Figure 1 shown, the probiotic tablet production traceability method of the present invention includes: S1. Obtain the traceability parameters of each process based on the production process of the probiotic tablet to obtain a traceability data matrix.

[0024] As a functional food, the quality and safety of probiotic tablets are directly related to the health of consumers. In the production process of probiotic tablets, multiple steps are usually involved, such as strain selection, amplification culture, and fermentation process, etc., all of which may affect the viable count, stability, and safety of the final product. Therefore, in this embodiment, a complete traceability system can be established to track every production step of probiotic tablets from raw materials to finished products, ensuring the controllability and consistency of product quality, and at the same time providing consumers with transparent product information to enhance trust.

[0025] For common probiotic tablets, their production process usually includes the following procedures, namely: strain selection, cultivation and amplification, fermentation, drying and powdering, formulation design and auxiliary material mixing, granulation and tablet forming, coating, quality inspection, packaging and storage, and shipment and transportation.

[0026] In one embodiment, the selected strain number can be recorded during the strain selection process and uploaded to the traceability database; during the cultivation and amplification process, the temperature of the incubator is monitored regularly through a temperature and humidity sensor and the humidity , and the pH value is collected regularly through a pH meter . After obtaining the parameter time series composed of the above multiple parameters, the Kalman filter algorithm is used to smooth the temperature time series and the humidity time series to eliminate sensor noise to obtain and , and the Z-score algorithm is used to standardize the pH value time series to obtain . Thus, the traceability parameter set of the cultivation and amplification process is obtained , and there is , and the set is uploaded to the traceability database; during the fermentation process, the fermentation tank number is recorded and uploaded to the traceability database; during the drying and powdering process, the number of the drying equipment is recorded , and then the drying method is recorded using One-Hot encoding . Exemplarily, it is explained that the freeze-drying method can be encoded as [1,0], and spray drying can be encoded as [0,1]. Then, the traceability parameter set of the drying and powdering process is denoted as , and there is and uploaded to the traceability database; during the formulation design and auxiliary material mixing process, the mass ratio of the bacterial powder to each auxiliary material is recorded , where represents the total number of auxiliary material categories. Here, the traceability parameter set of the formulation design and auxiliary material mixing process is denoted as , and there is and uploaded to the traceability database; during the granulation and tablet forming process, the granulation equipment number is recorded and the tablet pressing equipment number , then regularly obtain the pressure of the tablet pressing equipment through a pressure sensor , after obtaining the corresponding pressure time series, remove the noise generated by equipment vibration through wavelet transform to obtain , then the granulation and tablet forming process has a set of traceability parameters , and there is and upload it to the traceability database; in the coating process, record the coating equipment number and the coating thickness , then the coating process has a set of traceability parameters , and there is and upload it to the traceability database; in the quality inspection process, after sampling, use the plate counting method to record the viable bacteria count and upload it to the traceability database; in the packaging and storage process, record the packaging equipment number and the packaging date , then the packaging and storage process has a set of traceability parameters , and there is and upload it to the traceability database; finally, since the ex-factory and transportation processes of probiotic tablets are after the traceability code is spray-printed, no relevant data is obtained in this process. The preprocessing algorithms involved after collecting the traceability parameters of each process belong to well-known technologies and will not be elaborated here.

[0027] It should be noted that: during the production process of the above probiotic tablets, if traceability parameters need to be regularly recorded, such as the temperature, humidity, pH value in the cultivation and amplification process, and the pressure in the granulation and tablet forming process, etc., in this embodiment, the collection frequency can be set to once every two hours, and the implementer can also set the collection frequency to once every half hour, once every hour or once every four hours according to the cultivation duration of different probiotics.

[0028] So far, all the traceability parameter sets in the production process of probiotic tablets have been obtained and uploaded to the traceability database. The traceability parameter sets of each process are obtained from the traceability database and spliced to obtain a traceability data matrix , and there is: ; where represents the traceability parameters of strain selection; represents the traceability parameters of cultivation and amplification, represents the traceability parameters of fermentation; represents the traceability parameters of drying and powder making; represents the traceability parameters of formula design and excipient mixing; represents the traceability parameters of granulation and tablet forming; represents the traceability parameters of coating; represents the traceability parameters of quality inspection; Indicates the traceability parameters for packaging storage; Indicates the data splicing symbol.

[0029] S2. Take the colony images during the pilot-scale amplification process; obtain the colony spatial density by estimating the coordinates of the initial colony planting points and the number of colonies; obtain the growth rate gradient of the colonies based on the gradient magnitude of the colony images; obtain the regulation factor according to the colony spatial density and the growth rate gradient; adjust the threshold H through the regulation factor and complete the segmentation of the colony images based on the watershed algorithm with the dynamic threshold H to obtain the colony segmentation images.

[0030] During the production process of probiotic tablets, pilot-scale amplification is a key process. The colony images in this process reflect the growth status of probiotics during the amplification culture, and its quality determines the success rate of subsequent fermenter culture and also the quality of the subsequent produced probiotic tablets. Therefore, in this embodiment, the pilot-scale amplification process of probiotics can be used as a key traceability step.

[0031] S2.1. Take the colony images during the pilot-scale amplification process; obtain the colony spatial density by estimating the coordinates of the initial colony planting points and the number of colonies.

[0032] First, the characteristics of different colonies can be obtained by acquiring and segmenting the colony images during the pilot-scale amplification process, and then the characteristics of different colonies can be used as supplementary traceability parameters. Finally, the traceability data matrix and the supplementary traceability parameters are jointly used as a complete traceability set. It is clear that the growth processes of probiotics in different batches cannot be the same. Therefore, the characteristics of probiotic colonies obtained through pilot-scale amplification are unique, that is, the uniqueness of the final traceability code can be guaranteed. In addition, after segmenting the colony images to obtain different colonies, the status of the colonies in this batch can also be evaluated and analyzed to determine whether they can be used for fermentation culture.

[0033] In this embodiment, the watershed algorithm can be used to segment the colony images of pilot-scale amplification. The watershed algorithm regards the image as a topographic map, regards the areas with lower gray values as "basins", the higher areas as "mountains", and simulates the water flow spreading from low to high, and finally forms "watersheds" at the basin junctions. In practical applications, to avoid unnecessary segmentation due to noise or small gray fluctuations, the H-minima transform is usually combined to suppress local minima. However, when the distribution of the initial colony planting points is uneven, the growth rate of the mycelium is different, or the edge is interfered by gray light, the fixed threshold H cannot accurately distinguish the real colonies from the false candidate areas, which may lead to unsatisfactory segmentation effects.

[0034] The spatial distribution of the initial colony planting points is a key factor affecting the segmentation difficulty. Therefore, in this embodiment, the colony spatial density can be evaluated by estimating the coordinates of the initial colony planting points and the number of colonies. , and the specific calculation method is as follows: First, use a high-definition macro lens with a resolution of not less than 1080p to capture the RGB colony image of pilot-scale amplification; then convert the RGB colony image to the HSV color space format and extract the saturation S channel to enhance the contrast between the colonies and the background; further, based on the S channel, use the Otsu algorithm to generate a binary mask image, that is, set the colony area to 0 and the background area to 1; then use morphological opening operation to remove small noise points and retain the complete colony area; subsequently, use 8-neighborhood connectivity analysis on the image after the opening operation to label each independent region, and record the total number of independent connected regions as ; finally, calculate the geometric center of each independent connected region as the initial clustering center set of the K-means clustering algorithm. , and then use the K-means clustering algorithm to obtain multiple clustering clusters, and use the clustering center set of the clustering clusters as the coordinate set of the final planting points.

[0035] The input of the K-means clustering algorithm includes the image after the opening operation, the number of clusters, the initial centroid, the distance metric, the maximum number of iterations, and the convergence threshold. In this embodiment, the number of clusters can be set to the total number of independent connected regions , the initial centroid is set to , the distance metric is set to the Euclidean distance, the maximum number of iterations is set to the empirical value 100, and the convergence threshold is set to the empirical value 1e-4; the output of the algorithm is the clustering clusters and the centers of the clustering clusters. The coordinate set of the clustering cluster centers can be used as the coordinates of each planting point. The above K-means clustering algorithm, Otsu algorithm, and morphological opening operation are well-known technologies and will not be elaborated here. Then the colony spatial density is: ; where represents the colony spatial density; represents the estimated number of colonies; ( ) represents the natural exponential function; represents the initial planting point coordinates of the th colony; represents the coordinate mean of all planting points, and there is , which reflects the overall position center; represents the variance of the planting point coordinates, reflecting the degree of distribution dispersion, and there is ; represents the Euclidean norm.

[0036] When the distribution of planting points is more concentrated, is smaller, is closer to 0. At this time, the distance between each point and the mean value decreases, and the exponential function ( ) approaches 1, then is closer to 1, indicating that the colonies are relatively dense and prone to adhesion; conversely, when the distribution of planting points is more dispersed, is larger, and the exponential function ( ) approaches 0, then is closer to 0, indicating that the colonies are relatively independently distributed.

[0037] S2.2. Obtain the growth rate gradient of the colonies based on the gradient magnitude of the colony images; obtain the regulation factor according to the colony spatial density and the growth rate gradient.

[0038] The growth rate of the colonies and the edge expansion situation can be reflected by the image gradient change. Therefore, in this embodiment, the edge gradient change rate can be calculated using the images at adjacent time points, and then the growth dynamic information of the colony edge can be obtained, and the growth rate gradient of the colonies can be calculated , specifically: ; where represents the growth rate gradient of the colonies; represents the set time interval, which can be set to the empirical value of 120 min in this embodiment; represents the th moment of the colony image at the coordinate the Sobel gradient magnitude at the place, used to reflect the edge strength; represents the selected high-effective region, which can be set to the region with the top 10% gradient values in the colony image in this embodiment to ensure focusing on the main edges; represents the Euclidean norm; represents a very small positive number to prevent the denominator from being 0, which can be set to the empirical value of 1e-6 in this embodiment.

[0039] When the colony edge grows rapidly, the gradient magnitude increases. At this time, , then is also greater than 0, indicating that the edge density of the image increases; when the growth stagnates or slows down, the gradient change tends to 0. At this time, ; is close to 0 or negative. Further, based on the colony growth state, the segmentation reliability is quantified, and the regulation factor is constructed, specifically: ; where represents a regulatory factor; represents the sign function, which is used to indicate the growth direction for exemplary illustration. When is positive, it indicates accelerated growth; represents the natural logarithm function; represents the growth rate gradient of the colony; represents the spatial density of the colony.

[0040] When is positive and the spatial density of the colony is closer to 1, is larger, indicating that the colony is relatively dense and has a fast growth rate at this time. Then the regulatory factor is larger; when is negative or 0, and the spatial density of the colony is closer to 0, is also closer to 0. At this time, the colony is relatively dispersed and the growth has stagnated. Then the regulatory factor is smaller and close to 0.

[0041] S2.3. Adjust the threshold H through the regulatory factor and complete the segmentation of the colony image based on the watershed algorithm with the dynamic threshold H to obtain the colony segmentation image.

[0042] To make the suppression threshold of the H-minima transform better adapt to different colony images, in this embodiment, the regulatory factor is used to dynamically correct the reference threshold. The improved dynamic threshold calculation formula is as follows: ; where represents the dynamic threshold; represents the reference threshold; represents the adjustment coefficient, which is used to limit the amplitude of the threshold change. In this embodiment, it can be set to the empirical value 0.5; represents the hyperbolic tangent function, which is used to map the regulatory factor to the interval range of [-1, 1].

[0043] When is positive and larger, it indicates that the colony is relatively dense and has a fast growth rate at this time. At this time, the threshold H should be increased to avoid the adhesion of adjacent colonies caused by blurred edges. Then is larger; while when approaches 0, it indicates that the colony is relatively independent and has a slow growth rate. A relatively low threshold H should be maintained to capture the fine boundary. Then is relatively small; when is negative, it indicates that the colony may show the phenomenon of deterioration and contraction. It is necessary to further reduce the threshold H to identify potential separation points. Smaller.

[0044] Thus, the dynamic threshold of the H-minima transform is obtained. During the pilot-scale amplification of probiotics, colony images are collected every two hours and the watershed algorithm with the dynamic threshold is used to segment them to obtain the corresponding colony segmentation images.

[0045] S3. Extract colony features from the colony segmentation images obtained in step S2 to obtain supplementary traceability parameters, splice the supplementary traceability parameters with the traceability data matrix obtained in step S1 to obtain a complete traceability set, use a hash function to encrypt the complete traceability set to generate a traceability code and print it on the minimum packaging unit of the probiotic tablets, and at the same time upload it to the blockchain for evidence storage.

[0046] After obtaining the colony segmentation images in step S2, the area of each colony can be counted , perimeter , average edge gradient and texture features can be obtained. In this embodiment, the cv2.findContours() function in openCV can be used to extract the contours of each colony, and cv2.contourArea() can be used to calculate the contour area, cv2.arcLength() can be used to calculate the contour perimeter. Further, the Sobel operator can be used to extract the edge gradient and calculate the mean value, and the LBP algorithm can also be used to extract the texture features of each colony. The above algorithms are well-known technologies, and the specific calculation methods will not be elaborated here.

[0047] Thus, the feature set of the colony segmentation images is obtained , and there is , where represents the total number of colonies in the colony segmentation image and can also be recorded as supplementary traceability parameters. Since the colony images of the pilot-scale amplification are collected every two hours, multiple colony images will be obtained during the pilot-scale amplification of the same batch of probiotics, and correspondingly, there will also be multiple colony segmentation images. In this embodiment, the colony image at the 24th hour during the pilot-scale culture can be selected as the image for obtaining supplementary traceability parameters. In actual situations, the implementer can also adjust the time for obtaining images to 12 hours, 36 hours, 48 hours, etc. according to the selection of bacterial strains and the required duration of the pilot-scale culture process.

[0048] The traceability data matrix obtained in step S1 is spliced with the supplementary traceability parameters to obtain a complete traceability set, and there is: where Represents the complete traceability set; Represents the traceability data matrix; Represents supplementary traceability parameters. After obtaining the complete traceability set After that, the SHA-256 hashing algorithm is used for encryption to generate a unique traceability code. The generated traceability code is spray-coded on the smallest packaging unit of the probiotic tablets, such as a plastic packaging bottle, to facilitate consumers, manufacturers, or regulatory authorities to obtain the full-chain traceability information of the product through scanning; the complete traceability set and its corresponding traceability code can also be uploaded to the blockchain platform for deposit to achieve tamper-proof and publicly transparent traceability management.

[0049] Exemplary description: During the strain selection process, a batch of lactic acid bacteria strains were selected, and their strain numbers were BS2023-001, BS2023-002, BS2023-003 respectively. The strain numbers were uploaded to the traceability database and recorded as ; After strain selection, suitable and healthy strains were selected for cultivation and amplification. During this process, the monitored temperature time series data was: 36.8°C, 37.0°C,..., 37.1°C; the humidity time series data was: 60%, 62%,..., 60%; the humidity and temperature time series data were smoothed by Kalman filtering and uploaded to the traceability database; after obtaining the pH time series data, the standardized sequence 0.98, 1.02,..., 1.00 was obtained through Z-score standardization and uploaded to the traceability database and recorded as ; The same batch of probiotics after pilot-scale amplification was implanted into the fermentation tank for fermentation. During this process, the fermentation tank number F-001 was recorded and uploaded to the traceability database and recorded as ; Subsequently, after fermentation was completed, drying and powder-making were started. The drying equipment number D-02 and the drying method were recorded. Here, freeze-drying was used, and the code was [1,0]. It was uploaded to the traceability database and recorded as ; After the powder-making process, formula design and auxiliary material mixing were carried out. The ratios of the bacterial powder to the auxiliary materials were recorded, such as the ratio of the bacterial powder to the filler was 1:0.5, the ratio of the bacterial powder to the binder was 1:1, and the ratio of the bacterial powder to the disintegrant was 1:0.3. The ratios in this process were uploaded to the traceability database and recorded as ; The mixed auxiliary materials were granulated and tablet-shaped. The granulation equipment number G-01 and the tableting equipment number T-03 were recorded, and the tableting equipment pressure time series 100 kPa, 105 kPa,..., 100 kPa was collected. After noise reduction by wavelet transform, it was uploaded to the traceability database and recorded as ; After tableting, coating was carried out. The coating equipment number C-05 and the set coating thickness of 0.5 mm were recorded and uploaded to the traceability database and recorded as ; After coating, some samples were randomly inspected and the viable bacteria count was determined by the plate counting method to be 1×10 9CFU / tablet, 1.02×10 9 CFU / tablet, 1.035×10 9 CFU / tablet, …, 0.988×10 9 CFU / tablet, upload all viable counts to the traceability database and record as ; After sampling inspection, package all the probiotic tablets of this batch, record the packaging equipment number P-08 and the packaging date 2023-02-15, and upload these two parameters to the traceability database and record as .

[0050] Further, obtain the colony images at the 24th hour in the pilot-scale amplification process of the production of this batch of probiotic tablets, then obtain the colony segmentation images based on the watershed algorithm in step S2, and extract relevant parameters including the area, perimeter, average edge gradient, and texture features of each colony as: (250µm², 60µm, 5.6, 134), (360µm², 50µm, 4.9, 40), … (330µm², 55µm, 8.1, 98).

[0051] Then for this batch of probiotic tablets, its complete traceability set is: [BS2023-001, BS2023-002, BS2023-003, (36.8, 37.0, …, 37.1), (60, 62, …, 60), (0.98, 1.02, …, 1.00), F-001, D-02, (1, 0), (1:0.5, 1:1, 1:0.3), G-01, T-03, (100, 105, …, 100), C-05, 0.5, (1×10 9 , 1.02×10 9 , 1.035×10 9 , 0.988×10 9 ), P-08, 2023-02-15, {(250, 60, 5.6, 134), (360µm, 50µm, 4.9, 40), …, (330, 55µm, 8.1, 98)]; Then use the SHA-256 hash function to encrypt the above complete traceability set, and spray the generated traceability code on the outer surface of the probiotic plastic packaging bottle during the packaging process.

[0052] An embodiment of a probiotic tablet production traceability system: On the other hand, the present invention also provides a probiotic tablet production traceability system. As Figure 2 shown, the probiotic tablet production traceability system includes a processor and a memory, and the memory stores computer program instructions. When the computer program instructions are executed by the processor, a probiotic tablet production traceability method according to the first aspect of the present invention is implemented.

[0053] The production traceability system for probiotic tablets further includes other components well-known to those skilled in the art, such as communication interfaces, and their settings and functions are known in the art, so they will not be elaborated herein.

[0054] In the present invention, the aforementioned memory can be any tangible medium that contains or stores a program, and this program can be used by or in conjunction with an instruction execution system, device, or apparatus. For example, the computer-readable storage medium can be any suitable magnetic storage medium or magneto-optical storage medium, such as resistive random access memory (RRAM), dynamic random access memory (DRAM), static random access memory (SRAM), enhanced dynamic random access memory (EDRAM), high-bandwidth memory (HBM), hybrid memory cube (HMC), etc., or any other medium that can be used to store the required information and can be accessed by an application program, module, or both. Any such computer storage medium can be part of the device or accessible or connectable to the device. Any application or module described in the present invention can be implemented using computer-readable / executable instructions stored or otherwise held by such a computer-readable medium.

Claims

1. A method for tracing the production of probiotic tablets, characterized in that, The method includes: Obtaining relevant parameters of each process during the production of probiotic tablets to obtain a traceability data matrix; During the probiotic pilot-scale amplification process, colony images are collected at regular intervals, and the watershed algorithm is used to segment the colony images to obtain corresponding colony segmentation images; geometric features and texture features of multiple colonies in the colony segmentation images are extracted, and the geometric features and texture features are used as supplementary traceability parameters to be spliced with the traceability data matrix to obtain a complete traceability set; the complete traceability set is processed using an encryption algorithm to obtain a traceability code and is spray-coded on the minimum packaging unit of the probiotic tablets; The watershed algorithm further includes multiplying the basic threshold of the H transformation by an adjustment value to obtain a dynamic threshold, and the adjustment value is obtained by a series of predetermined mathematical operations including logarithmic operations and hyperbolic tangent operations through the growth rate gradient, colony spatial density, and a set adjustment coefficient; Obtaining multiple connected regions in the colony image, and taking the geometric center of each connected region as the coordinate of the initial planting point of the colony; calculating the coordinate variance and coordinate mean of all coordinates, obtaining the Euclidean distance between each coordinate and the coordinate mean, and obtaining the colony spatial density according to the ratio of the Euclidean distance to the coordinate variance; calculating the difference between the edge gradient amplitudes of each colony in the colony image and the colony image at a set time interval before to obtain the growth rate gradient.

2. The production traceability method of a probiotic tablet according to claim 1, characterized in that, The specific calculation method of the dynamic threshold is: ； wherein represents the reference threshold of the H transformation; represents the adjustment coefficient; and respectively represent the growth rate gradient and the colony spatial density; represents the hyperbolic tangent function; represents the sign function.

3. A method for tracing the production of probiotic tablets according to claim 2, characterized in that, The specific calculation method of the colony spatial density is: ； Among them represents the spatial density of colonies; represents the estimated number of colonies, specifically the number of independent connected regions in the colony image; ( ) represents the natural exponential function; represents the coordinates of the initial planting point of the th colony; represents the mean value of the coordinates of all initial planting points; represents the Euclidean norm.

4. A method for tracing the production of probiotic tablets according to claim 2, characterized in that, The specific calculation method of the growth rate gradient is: ； in represents the growth rate gradient of the colony; Indicates the set time interval; Indicates The coordinates of the colony image at the time The Sobel gradient amplitude at ; Indicates the area in the colony image where the gradient amplitude is greater than the set threshold; represents the Euclidean norm; Represents a very small positive number.

5. A method for tracing the production of probiotic tablets according to claim 1, characterized in that, The relevant parameters of each process include: The strain number of the strain selection process; the temperature, humidity, and pH value of the culture and amplification process; the fermentation tank number of the fermentation process; the drying equipment number and the coding of the drying method in the drying and powder-making process; the mass ratio of the bacterial powder to each auxiliary material in the formula design and auxiliary material mixing process; the granulation equipment number, tablet pressing equipment number, and the pressure of the tablet pressing equipment in the granulation and tablet forming process; the coating equipment number and the coating thickness in the coating process; the viable bacteria count after sampling inspection in the quality inspection process; the packaging equipment number and the packaging date in the packaging and storage process.

6. A method for tracing the production of probiotic tablets according to claim 1, characterized in that, It also includes performing preprocessing operations on the colony image, specifically: Performing HSV space conversion on the colony image to obtain an HSV colony image; Based on the S channel in the HSV colony image, using the Otsu algorithm for threshold segmentation to obtain a segmentation image; Performing morphological opening operation on the segmentation image.

7. A method for tracing the production of probiotic tablets according to claim 6, characterized in that, Obtaining multiple connected regions in the colony image, including: Using 8-neighborhood connectivity analysis on the preprocessed colony image to obtain multiple connected regions.

8. A method for tracing the production of probiotic tablets according to claim 1, characterized in that, Extracting geometric features and texture features of multiple colonies in the colony segmentation image, including: Using the cv2.findContours() function in openCV to extract the contours of each colony and using cv2.contourArea() to calculate the area of the contour; Using cv2.arcLength() to calculate the perimeter of the contour; Using the Sobel operator to extract the edge gradient and calculate the mean value; Using the LBP algorithm to extract the texture features of each colony to obtain LBP values.

9. A method for tracing the production of probiotic tablets according to claim 1, characterized in that, The encryption algorithm is the SHA-256 hash encryption algorithm.

10. A production traceability system for probiotic tablets, characterized in that, It includes a processor and a memory, and the memory stores computer program instructions. When the computer program instructions are executed by the processor, it implements a method for tracing the production of probiotic tablets according to any one of claims 1 to 9.

Images