Real-time monitoring method and system for crystallization process of cistanche

Abstract

The invention belongs to the technical field of image processing, and particularly relates to a cistanche crystallization process real-time monitoring method and system, and the method comprises the steps: carrying out the segmentation model processing and image analysis of an obtained cistanche image, and obtaining the geometric parameters of crystal particles, including the particle area, the contour perimeter, and the length and width of a minimum enclosing rectangle; obtaining a crystal morphology isomerism index; calculating a crystal neighborhood disturbance degree; calculating a crystal signal-to-noise ratio; calculating a crystal comprehensive early warning value; performing time sequence confirmation on the candidate particles with high matching degree; on the basis of global crystal particle distribution in the current process stage, carrying out global confirmation of abnormal states on the candidate particles confirmed through the time sequence; and comparing the globally confirmed candidate particles with a final early warning threshold value and carrying out early warning. The method solves the technical problem of detection desensitization of abnormal crystal forms under the background of massive normal crystal samples in the prior art.

Description

Technical Field

[0001] This invention relates to the field of image processing technology. More specifically, this invention relates to a method and system for real-time monitoring of the crystallization process of Cistanche deserticola. Background Technology

[0002] In the industrial production of active ingredients from Cistanche deserticola, such as echinacoside, purification through crystallization is a crucial step in obtaining high-purity products. The core objective of crystallization is not only to precipitate crystals but also to stably produce normal crystalline forms with specific solubility, high stability, and bioavailability. However, within the crystallization vessel, normal crystalline forms coexist with abnormal crystalline forms such as metastable crystalline forms and solvates, which have drastically different physical properties, competing for solute molecules in the solution. Any minute fluctuation in process parameters can induce preferential nucleation of abnormal crystalline forms, triggering catastrophic and irreversible crystalline transformations, resulting in the entire batch of products failing to meet quality standards.

[0003] To achieve precise control over the crystallization process, existing technologies typically employ online image analysis methods. By capturing microscopic images inside the crystallization vessel and using image processing algorithms to statistically analyze macroscopic parameters such as crystal size and quantity, the efficiency of control over the Cistanche deserticola crystallization process is greatly improved.

[0004] However, in the early or stable stages of the crystallization process, abnormal crystal nuclei are a rare occurrence, and their number and size are far smaller than those of the dominant normal crystal form. Traditional image analysis methods typically rely on standard classification models, which, when dealing with scenarios with extremely unbalanced class distributions, tend to heavily favor the majority class, i.e., the normal crystal form. This results in extremely low sensitivity in identifying the very few abnormal crystal nuclei, making it difficult to effectively identify and warn of abnormal crystal forms before they trigger large-scale crystal transformation. Therefore, it cannot meet the ultra-high sensitivity detection requirement of accurately identifying a single abnormal crystal nucleus among millions of normal crystals. Summary of the Invention

[0005] To address the technical problem of insensitivity in detecting abnormal crystal forms in the context of a large number of normal crystal samples, the present invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides a method for real-time monitoring of the crystallization process of Cistanche deserticola, comprising: The acquired Cistanche deserticola images were processed using a segmentation model and image analysis to obtain the geometric parameters of the crystal particles, including particle area, perimeter, and the length and width of the minimum bounding rectangle. The theoretical area of ​​the crystal was calculated based on the perimeter of the crystal particles, and the crystal morphology heterogeneity index was obtained by comparing the theoretical area with the actual area. A circular neighborhood was pre-defined with each individual crystal particle as the center. The crystal morphology heterogeneity index and Euclidean distance from the central crystal were calculated for all crystal particles within the circular neighborhood except for the central crystal. Based on the sparse distribution of the individual crystal particle and its neighboring crystal particles, and the crystal morphology heterogeneity index, the crystal neighborhood was calculated. The perturbation degree is calculated based on the difference between the crystal morphology heterogeneity index and the perturbation degree of the crystal neighborhood. The crystal signal-to-noise ratio is calculated by combining the numerical correlation and dynamic adjustment relationship between the crystal signal-to-noise ratio and the crystal morphology heterogeneity index. Candidate crystal particles that exceed the average of all crystal comprehensive warning values ​​for the first time are subjected to time-series growth fingerprint matching verification, and candidate particles with high matching degree are confirmed in time sequence. Based on the global crystal particle distribution in the current process stage, the candidate particles that have passed the time sequence confirmation are globally confirmed for abnormal states. The globally confirmed candidate particles are compared with the final warning threshold and a warning is issued.

[0007] Traditional methods struggle to identify a small number of abnormal crystals within a large pool of normal crystals, often resulting in insensitive detection. This invention first accurately acquires crystal data and filters out useless information, laying a reliable foundation for subsequent analysis. Then, it comprehensively assesses abnormal crystals based on multiple indicators, including crystal morphology, surrounding environment, time variations, and overall distribution, thus accurately identifying and eliminating false positives. Whether it's a minor anomaly in a stable environment or a significant anomaly in a deteriorating environment, it can accurately identify both potential risks and avoids blindly issuing warnings. Ultimately, it only triggers warnings for high-confidence abnormal crystals, effectively solving the problem of insensitivity in traditional methods, ensuring a stable crystallization process, and preventing product quality defects caused by abnormal crystals.

[0008] Preferably, the crystal grain geometric parameters, including grain area, perimeter, and the length and width of the minimum bounding rectangle, are obtained, including: Using an immersion high-resolution microscopic imaging probe, real-time image streams of the solid-liquid two-phase mixture inside the crystallization vessel are continuously acquired at a preset frame rate. Each acquired image frame is processed using an instance segmentation model to accurately identify and segment each individual crystal particle in the image. For each identified particle, image analysis algorithms are used to obtain the crystal particle's geometric parameters, including: particle area, outline perimeter, and the length and width of the minimum bounding rectangle. A screening algorithm is then applied to identify and exclude crystal agglomerates that are too large or have irregular shapes based on the crystal particle's geometric parameters.

[0009] Preferably, the crystal morphological isomerism index satisfies the following expression: ; In the formula, Indicates the first The crystal morphological isomerism index of each crystal grain, dimensionless; Indicates the first The perimeter of the outline of a crystal grain; Indicates the first The length of the smallest bounding rectangle of a crystal grain; Indicates the first The width of the minimum bounding rectangle of a crystal grain; Represents the natural exponential function; Pi is the mathematical constant of a circle.

[0010] This invention derives a crystal morphology isomerism index through a specific calculation method, which can clearly distinguish crystals of different morphologies. That is, normal crystals and abnormal crystal particles will correspond to different crystal morphology isomerism indices. This allows for the rapid identification of crystal particles with abnormal morphologies, providing a clear basis for further judgment on whether there are problems with the crystal particles, and enabling staff to discover potentially abnormal crystal particles earlier.

[0011] Preferably, the crystal neighborhood perturbation degree satisfies the following expression: ; In the formula, Indicates the first The crystal neighborhood perturbation degree of each crystal grain, dimensionless; Indicates the first The set of all crystal particles in the circular neighborhood of a crystal particle, excluding the central crystal. Indicates the first In the circular neighborhood of the nth crystal grain Crystal morphological isomerism index of individual crystal grains; Indicates the first The first crystal grain and the first In the circular neighborhood of the nth crystal grain Euclidean distance between individual crystal grains; It is a very small positive number, used to prevent the inner denominator from being zero; It is a very small positive number used to prevent the outer denominator from being zero.

[0012] When calculating the perturbation degree of a crystal neighborhood, this invention considers the situation of surrounding crystal particles and their distance from the central crystal. The impact of abnormal crystals that are close to the central crystal on the stability of the surrounding environment of the central crystal will be more accurately reflected, which can more realistically reflect the state of the microenvironment in which the central crystal is located. This helps staff to fully understand the situation around the crystal and determine whether there are any local environmental anomalies.

[0013] Preferably, calculating the crystal signal-to-noise ratio includes: The crystal morphological heterogeneity index of the current crystal particle is compared with the perturbation degree of the crystal neighborhood, and the calculation result is used as the crystal signal-to-noise ratio. The crystal signal-to-noise ratio is used to evaluate the degree of anomaly of the current crystal particle relative to its surrounding background.

[0014] Preferably, the calculated crystal comprehensive early warning value satisfies the following expression: ; In the formula, Indicates the first The overall early warning value of each crystal particle is dimensionless. Indicates the first Crystal morphological isomerism index of individual crystal grains; Indicates the first The signal-to-noise ratio of each crystal particle; This represents the mean crystal morphological isomerism index of all crystal particles; It is a very small dimensionless positive number used to prevent the denominator from being zero; It is a natural exponential function.

[0015] When calculating the comprehensive early warning value of a crystal, this invention combines the degree of anomaly of the crystal particle itself with its prominence in the surrounding environment. It can also dynamically adjust according to the overall crystal environment. When the overall environment is stable, even minor anomalies can be keenly detected. When the environment deteriorates, it can focus on the most prominent anomalies, making the early warning more in line with the actual situation, neither omitting potential risks nor issuing blind warnings.

[0016] Preferably, time-series verification is performed on candidate particles with high matching degree, including: For candidate crystal particles whose comprehensive early warning value exceeds the average of all comprehensive early warning values ​​for the first time, the system initiates tracking and records its subsequent continuous values. The time series of crystal morphological heterogeneity index in the frame image is denoted as the crystal growth sequence to be tested. The crystal growth sequence to be tested is dynamically time-warped and matched with the typical abnormal crystal growth sequences pre-stored in the expert knowledge base. The abnormal state of the candidate particle is confirmed in time only when the matching degree between the crystal growth sequence to be tested and a typical abnormal crystal growth sequence is higher than the average matching similarity.

[0017] Preferably, the abnormal state of candidate particles confirmed through time-series analysis is globally confirmed, including: For candidate particles confirmed through time-series analysis, the system further performs real-time statistics on the global distribution of crystal morphology heterogeneity indices of all crystal particles within the current field of view, and calculates the 95th percentile of all crystal particles; then, based on the absolute difference between the crystal morphology heterogeneity index and the 95th percentile, the relative significance of the candidate particle is calculated; when the relative significance of the candidate particle exceeds a certain multiple, the abnormal state of the candidate particle is globally confirmed.

[0018] This invention determines a dynamic standard for judging whether crystal particles are abnormal based on the overall distribution of all crystal particles. This standard will be adjusted accordingly at different process stages to avoid using a fixed standard to judge crystal particles at different stages. This ensures that minor abnormalities can be detected in time at the beginning of the process and that serious abnormalities that really need attention can be accurately identified in the later stages.

[0019] Preferably, the process of comparing globally confirmed candidate particles with the final warning threshold and issuing a warning includes: Only when a candidate particle Only after passing the temporal growth fingerprint matching verification and global confirmation of abnormal states can the system confirm it as a high-confidence warning target. The system then compares its final crystal comprehensive warning value with a fixed final decision threshold. Once the condition that the crystal comprehensive warning value exceeds the final decision threshold is met, the system immediately triggers the highest-level warning and highlights the candidate particle. The location, image, and all computational features of the identified anomalous crystal nuclei.

[0020] Secondly, the present invention provides a real-time monitoring system for the crystallization process of Cistanche deserticola, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned real-time monitoring method for the crystallization process of Cistanche deserticola is implemented.

[0021] By adopting the above technical solution, a computer program for real-time monitoring of the crystallization process of Cistanche deserticola is generated and stored in a memory for loading and execution by a processor. A terminal device is then created based on the memory and processor for convenient use.

[0022] The beneficial effects of this invention are as follows: It can monitor the crystallization process in real time and accurately, promptly detect any abnormalities, and help staff take measures in the early stages of problems to prevent them from escalating and causing the entire batch of products to fail to meet quality standards, thus reducing production losses. Simultaneously, the entire monitoring process is highly automated and easy to operate, reducing errors and workload from manual operation and improving production efficiency. Furthermore, a stable crystallization process ensures consistent product quality, enhances the product's competitiveness in the market, brings better economic benefits to enterprises, and also promotes the development of monitoring technology in the production of Cistanche deserticola active ingredients. Attached Figure Description

[0023] Figure 1 This is a flowchart illustrating a real-time monitoring method for the crystallization process of Cistanche deserticola according to the present invention; Figure 2 This is a schematic diagram showing the actual situation inside the Cistanche deserticola pot in this invention; Figure 3 This is a schematic diagram illustrating the crystallization and extraction of Cistanche deserticola in this invention. Detailed Implementation

[0024] This invention discloses a method for real-time monitoring of the crystallization process of Cistanche deserticola, referring to... Figure 1 This includes steps S1-S4: S1: Perform segmentation model processing and image analysis on the acquired Cistanche deserticola image to obtain the crystal particle geometric parameters, including particle area, outline perimeter, and the length and width of the minimum bounding rectangle.

[0025] It should be noted that the monitoring method of this invention is based on data derived from continuous observation of the microscopic world inside the crystallization vessel. First, it requires acquiring and standardizing this raw image information from the data source to generate the basic data needed for subsequent calculations, such as... Figure 2 Actual picture of the inside of the pot containing Cistanche deserticola. Figure 3 Image of Cistanche deserticola crystallization extraction. The semi-continuous crystallization process of Cistanche deserticola's active ingredients aims to maintain the stable growth of normal needle-like crystals within the reactor. However, traditional offline sampling and testing methods not only suffer from significant time lag, but temperature and pressure changes during the sampling process can also lead to crystal dissolution or secondary nucleation, failing to reflect the true state within the reactor. This invention automates and standardizes the acquisition of in-situ, real-time, full-volume particle image data within the reactor, establishing a clean and well-organized data foundation for subsequent intelligent analysis.

[0026] Specifically, the acquired Cistanche deserticola images are processed using a segmentation model and analyzed to obtain the crystal grain geometric parameters, including grain area, contour perimeter, and the length and width of the minimum bounding rectangle, including: Using an immersion high-resolution microscopic imaging probe, real-time image streams of the solid-liquid two-phase mixture inside the crystallization vessel are continuously acquired at a preset frame rate. Each acquired image frame is processed using an instance segmentation model to accurately identify and segment each individual crystal particle in the image. For each identified particle, image analysis algorithms are used to obtain the crystal particle's geometric parameters, including: particle area, outline perimeter, and the length and width of the minimum bounding rectangle. A screening algorithm is then applied to identify and exclude crystal agglomerates that are too large or have irregular shapes based on the crystal particle's geometric parameters.

[0027] Thus, the geometric parameters of the crystal grains were obtained.

[0028] S2: Calculate the theoretical area of ​​the crystal based on the perimeter of the crystal grains, and compare the theoretical area with the actual area of ​​the crystal to obtain the crystal morphological isomerism index.

[0029] It should be noted that in real crystallization scenarios, an ideal crystal particle exhibits a specific needle-like or rod-like shape, with a regular and highly anisotropic morphology. In contrast, the metastable nucleus of an abnormal crystal particle often appears as a blocky or flocculent mass, with a more compact and isotropic morphology. Therefore, this invention constructs an evaluation criterion that allows the system to automatically identify and calculate the degree of morphological degradation from needle-like to blocky, thereby enabling the identification of individual abnormal particles at the starting point of the analysis.

[0030] Specifically, the theoretical area of ​​the crystal is calculated based on the perimeter of the crystal grains, and the theoretical area is compared with the actual area of ​​the crystal to obtain the crystal morphological isomerism index, including: The crystal morphological isomerism index satisfies the following expression: ; In the formula, Indicates the first The crystal morphological isomerism index of each crystal grain, dimensionless; Indicates the first The perimeter of the outline of a crystal grain; Indicates the first The length of the smallest bounding rectangle of a crystal grain; Indicates the first The width of the minimum bounding rectangle of a crystal grain; Represents the natural exponential function; Pi is the mathematical constant of a circle.

[0031] In the formula, Indicates the first The theoretical crystal area of ​​a given crystal grain with a given perimeter, denominator For the first The actual area of ​​the crystal occupied by each crystal particle; The imbalance between the theoretical area and the actual area of ​​the crystal was calculated. This ratio is extremely large for needle-shaped crystals and relatively small for bulk crystals. This means that by using a natural exponential function for nonlinear mapping, extremely large ratios are mapped to values ​​close to 0, and small ratios are mapped to values ​​significantly greater than 0, thereby achieving nonlinear amplification of the morphology from needle-like to any non-needle-like form.

[0032] For example, suppose there exists an ideal needle-shaped crystal grain with the following geometric characteristics: micrometer, micrometers, its perimeter is approximately For micrometers, its crystal morphological isomerism index is: This value is extremely close to 0; there is also a case where an abnormal crystal nucleus, approximately square in shape, appears, with the following geometric characteristics: micrometer, micrometers, its circumference For micrometers, its crystal morphological isomerism index is: The above calculation results, To retain three decimal places, To retain eight decimal places.

[0033] S3: Using a single crystal grain as the center crystal, a circular neighborhood is predefined. The crystal morphology isomerism index and the Euclidean distance from the center crystal are calculated for all crystal grains in the circular neighborhood except for the center crystal. Based on the sparse distribution of the single crystal grain and the neighboring crystal grains and the crystal morphology isomerism index, the crystal neighborhood perturbation degree is calculated.

[0034] It should be noted that in a real crystallization vessel, the abnormal morphology of a single crystal particle may sometimes be due to image noise or accidental overlap and orientation of particles, and does not necessarily represent a real abnormal crystal nucleus event. A real, newly formed abnormal crystal nucleus is the result of a sudden change in the local physicochemical environment. Such a sudden change often has a spatial aggregation effect. Therefore, this invention introduces a crystal neighborhood perturbation index. This index not only considers the morphology of neighboring particles, but also weights the values ​​based on the distance of neighboring crystal particles from the center of the single crystal particle, so as to more accurately calculate the overall stability of the microenvironment in which a single crystal particle is located.

[0035] Specifically, a circular neighborhood is predefined with a single crystal grain as the central crystal. The crystal morphological isomerism index and Euclidean distance from the central crystal are calculated for all crystal grains within the circular neighborhood except for the central crystal. Based on the sparse distribution of the single crystal grain and the neighboring crystal grains and the crystal morphological isomerism index, the crystal neighborhood perturbation is calculated, including: The crystal neighborhood perturbation degree satisfies the following expression: ; In the formula, Indicates the first The crystal neighborhood perturbation degree of each crystal grain, dimensionless; Indicates the first The set of all crystal particles in the circular neighborhood of a crystal particle, excluding the central crystal. Indicates the first In the circular neighborhood of the nth crystal grain Crystal morphological isomerism index of individual crystal grains; Indicates the first The first crystal grain and the first In the circular neighborhood of the nth crystal grain Euclidean distance between individual crystal grains; It is a very small positive number, used to prevent the inner denominator from being zero; It is a very small positive number used to prevent the outer denominator from being zero.

[0036] In the formula, Indicates the first In the circular neighborhood of the nth crystal grain The crystal morphology isomerism index of the first crystal grain, the crystal morphology isomerism index being related to the crystal morphology isomerism index of the first crystal grain. The contribution of the distance between individual crystal grains; the closer the distance, the greater the influence. The smaller the value, the higher the contribution. The larger; This represents the summation of the contribution values ​​corresponding to all neighboring crystal particles; A weighted average was calculated such that anomalous crystal particles closer to the central particle contribute significantly more to the perturbation of the crystal neighborhood than anomalous crystal particles farther away.

[0037] For example, suppose the first There are two anomalous crystal particles surrounding the first crystal particle. and Their crystal morphological isomerism indices are all ,but and Adjacent, distance micrometer; and The distance is far. Micrometers, using a distance-weighted formula. and Negligible Its crystal neighborhood perturbation degree However, if a simple arithmetic average is used, the result is Assuming The anomaly is higher. ,and Low anomaly At this point, the perturbation degree of the neighborhood of the weighted crystal is This is significantly higher than the arithmetic mean. This accurately reflects the strong influence of high-risk neighboring states on the current microenvironment of crystal particles. The above calculation results... To retain one decimal place, To retain three decimal places.

[0038] S4: Calculate the crystal signal-to-noise ratio (SNR) based on the difference between the crystal morphology heterogeneity index and the crystal neighborhood perturbation degree of the current crystal particles; calculate the comprehensive early warning value of the crystal by combining the numerical correlation and dynamic adjustment relationship between the crystal SNR and the crystal morphology heterogeneity index; perform time-series growth fingerprint matching verification on candidate crystal particles that first exceed the average comprehensive early warning value of all crystals, and perform time-series confirmation on candidate particles with high matching degree; perform global confirmation of abnormal state on candidate particles that have passed the time-series confirmation based on the global crystal particle distribution of the current process stage; compare the globally confirmed candidate particles with the final early warning threshold and issue an early warning.

[0039] It should be noted that the ultimate goal of this invention is to achieve ultra-high sensitivity early warning for rare and abnormal crystal nuclei through a computational framework that is entirely data-driven and constrained by physical processes. Thus, this invention has completed the progressive calculation from the standardization of raw image data to the calculation of the crystal morphology heterogeneity index of individual particles and the crystal neighborhood perturbation degree of their microenvironment. The following sections propose a peak amplification calculation mode, which compares and amplifies the crystal morphology heterogeneity index representing individual anomalies with the crystal neighborhood perturbation degree representing the background environment.

[0040] Specifically, based on the difference between the crystal morphological isomerism index and the crystal neighborhood perturbation degree of the current crystal particles, the crystal signal-to-noise ratio is calculated, including: The crystal morphological heterogeneity index of the current crystal particle is compared with the perturbation degree of the crystal neighborhood, and the calculation result is used as the crystal signal-to-noise ratio. The crystal signal-to-noise ratio is used to evaluate the degree of anomaly of the current crystal particle relative to its surrounding background.

[0041] It should be noted that the ultimate goal of this invention is to achieve ultra-high sensitivity early warning for rare and abnormal crystal nuclei. To this end, this invention has obtained, through a series of calculations, a crystal morphology heterogeneity index characterizing the degree of anomalousness of a single crystal particle, and a crystal signal-to-noise ratio characterizing the clarity of the anomalous signal relative to its local background. Now, this invention faces a final core problem: how to fuse the crystal morphology heterogeneity index, representing signal intensity, and the crystal signal-to-noise ratio, representing signal clarity, into a final, decision-level early warning value. A simple linear combination or product cannot reflect the dynamic decision-making needs of complex industrial scenarios. For example, in an extremely stable system, even a clear, minute anomaly should be greatly amplified; while in a chaotic system where the system has generally deteriorated, only those truly outstanding anomalous signals deserve attention. Therefore, this invention proposes a calculation method that can dynamically adjust the amplification based on the global environment.

[0042] Specifically, by combining the numerical correlation and dynamic adjustment relationship between the crystal signal-to-noise ratio and the crystal morphological isomerism index, a comprehensive early warning value for the crystal is calculated, including: The overall early warning value for crystals satisfies the following expression: ; In the formula, Indicates the first The overall early warning value of each crystal particle is dimensionless. Indicates the first Crystal morphological isomerism index of individual crystal grains; Indicates the first The signal-to-noise ratio of each crystal particle; This represents the mean crystal morphological isomerism index of all crystal particles; It is a very small dimensionless positive number used to prevent the denominator from being zero; It is a natural exponential function.

[0043] In the formula, Indicates the first The degree of deviation of the crystal morphological isomerism index of an individual crystal grain from the mean of the crystal morphological isomerism indices of all crystal grains. This indicates that when the crystallization process is stable, When I was very young, even Smaller It will rapidly approach the upper limit of 2, making the system extremely sensitive to small anomalies; when the crystallization process deteriorates, When it rises, only Only crystal grains far exceeding the average level can make Maintain a high value; It achieves a combination of clarity and dynamic sensitivity for abnormal signals, so that the enhancement of abnormal signals depends on both the clarity of the signal itself and the sensitivity adjustment under the current environment. Yes The magnification factor, through The structure retains The information is further enhanced with dynamic effects; By using individual crystal particles and Multiplication enables dynamic and adaptive early warning of risks to crystal particles, that is, it is extremely sensitive to minor anomalies in stable environments and focuses on the most prominent risks in deteriorating environments.

[0044] For example, when the overall crystallization environment is stable, an isolated anomalous crystal grain appears. Its own crystal morphological isomerism index is relatively high, which is However, the surrounding crystal particles are all normal crystal particles, leading to abnormal crystal particles. The crystal neighborhood perturbation is extremely low, which is... Then its signal-to-noise ratio At this point, the system is stable, and the average crystal morphology isomerism index of all crystal particles is very low, at [value missing]. , , At this point, the overall early warning value for the crystal is: When the overall crystallization environment has deteriorated to a certain extent, there exists an abnormal crystal grain with the same anomaly. Its own crystal morphological isomerism index and anomalous crystal grains Same, for However, other anomalous crystal particles also exist around it, resulting in a high degree of perturbation in its crystal neighborhood. Then its signal-to-noise ratio At this point, the system has generally deteriorated, and the average crystal morphology isomerism index of all crystal particles is high. , At this point, the overall early warning value for the crystal is: As can be seen from the comparison, this invention introduces a global average value. As a dynamic benchmark, the system automatically becomes extremely sensitive in stable environments and automatically becomes less sensitive in deteriorating environments to focus on the most prominent risks. The above calculation results... , , All values ​​are rounded to three decimal places.

[0045] It should be noted that the crystal comprehensive warning value calculated above is only a snapshot of the spatial information of a single frame image. This can easily lead to the misjudgment of transient artifacts caused by image noise, accidental particle overlap, etc., as real threats. A real, anomalous crystal nucleus with technological risks will inevitably follow specific physicochemical laws in its morphological evolution, exhibiting a continuous or specific growth pattern over time. Therefore, this step aims to improve the accuracy of the warning by verifying the life cycle of crystal particles, filtering out isolated anomalies without historical origins.

[0046] Preferably, candidate crystal particles that first exceed the average of all crystal comprehensive warning values ​​are subjected to time-series growth fingerprint matching verification, and candidate particles with high matching degrees are confirmed time-series, including: For candidate crystal particles whose comprehensive early warning value exceeds the average of all comprehensive early warning values ​​for the first time, the system initiates tracking and records its subsequent continuous values. The time series of crystal morphological heterogeneity index in the frame image is denoted as the crystal growth sequence to be tested. The crystal growth sequence to be tested is dynamically time-warped and matched with the typical abnormal crystal growth sequences pre-stored in the expert knowledge base. The abnormal state of the candidate particle is confirmed in time only when the matching degree between the crystal growth sequence to be tested and a typical abnormal crystal growth sequence is higher than the average matching similarity.

[0047] It should be noted that whether a crystal particle is abnormal or not is a relative concept, and the judgment criteria should be dynamically adjusted according to the overall state of the crystallization process. For example, in the early stage of crystallization, the entire system environment is relatively stable, and a slight abnormality in the crystal morphology of a single crystal particle is worth noting. However, in the later stage of crystallization, the crystal particle population itself may exhibit a certain degree of morphological generalization, at which point a higher abnormality threshold is required. This step aims to establish an adaptive reference system to ensure that the system only issues warnings for truly significant anomalies within the current process context.

[0048] Preferably, based on the global crystal particle distribution at the current process stage, the abnormal state of candidate particles confirmed through time-series analysis is globally confirmed, including: For candidate particles confirmed through time-series analysis, the system further performs real-time statistics on the global distribution of crystal morphology heterogeneity indices of all crystal particles within the current field of view, and calculates the 95th percentile of all crystal particles; then, based on the absolute difference between the crystal morphology heterogeneity index and the 95th percentile, the relative significance of the candidate particle is calculated; when the relative significance of the candidate particle exceeds a certain multiple, the abnormal state of the candidate particle is globally confirmed.

[0049] It should be noted that the 95th percentile is a statistical indicator. Specifically, it refers to the value at the 95th percentile when all crystal particles in the current field of view are sorted from low to high according to their crystal morphology heterogeneity index. In this invention, it represents the statistical upper limit of the morphology exhibited by the vast majority of crystal particles at the current moment, and is therefore used as a dynamic anomaly detection baseline that can automatically follow changes in process conditions.

[0050] It should be noted that after spatial correlation analysis, temporal fingerprint verification, and global confirmation of abnormal states, the system has accumulated multi-dimensional evaluation criteria for each candidate crystal particle. This step is the final link in the decision-making process, and its purpose is to integrate all evaluation criteria. Only those candidate crystal particles that have passed the triple tests of space, time, and global status are identified as high-confidence warning targets and trigger the final warning action.

[0051] Specifically, the candidate particles that have received global confirmation are compared with the final warning threshold, and a warning is issued, including: Only when a candidate particle Simultaneously, after passing the temporal growth fingerprint matching verification and the global confirmation of the abnormal state, the system can confirm it as a high-confidence warning target and compare its final crystal comprehensive warning value with a fixed final decision threshold. Once the condition that the crystal comprehensive warning value is greater than the final decision threshold is met, the system immediately triggers the highest level of warning and highlights the location, image and all calculated features of the abnormal crystal nucleus.

[0052] It should be noted that the fixed final decision threshold is the optimal risk cutoff value determined by offline statistical analysis of production data from historical successful and failed batches, after achieving a balance between a high detection rate and an acceptable false alarm rate.

[0053] This completes the real-time monitoring and early warning of the crystallization process of Cistanche deserticola.

[0054] This invention also discloses a real-time monitoring system for the crystallization process of Cistanche deserticola, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, a real-time monitoring method for the crystallization process of Cistanche deserticola according to this invention is implemented.

[0055] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0056] While this specification has shown and described numerous embodiments of the invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of this invention.

Claims

1. A method for real-time monitoring of the crystallization process of Cistanche deserticola, characterized in that, include: The acquired Cistanche deserticola images were processed using a segmentation model and analyzed to obtain the crystal particle geometric parameters, including particle area, outline perimeter, and the length and width of the minimum bounding rectangle. The theoretical area of ​​the crystal is calculated based on the perimeter of the crystal grains. The theoretical area is then compared with the actual area of ​​the crystal to obtain the crystal morphological isomerism index. A circular neighborhood is pre-defined with a single crystal grain as the center crystal. The crystal morphological isomerism index and Euclidean distance of all crystal grains in the circular neighborhood except the center crystal are calculated. Based on the sparse distribution of the single crystal grain and the neighboring crystal grains and the crystal morphological isomerism index, the perturbation degree of the crystal neighborhood is calculated. Based on the difference between the crystal morphological heterogeneity index and the crystal neighborhood perturbation degree of the current crystal particles, the crystal signal-to-noise ratio is calculated. Combining the numerical correlation and dynamic adjustment relationship between the crystal signal-to-noise ratio and the crystal morphological heterogeneity index, the comprehensive early warning value of the crystal is calculated. Candidate crystal particles that first exceed the average comprehensive early warning value of all crystals are subjected to time-series growth fingerprint matching verification, and candidate particles with high matching degree are confirmed time-series. Based on the global crystal particle distribution at the current process stage, the candidate particles that have passed the time-series confirmation are globally confirmed for abnormal states. The globally confirmed candidate particles are compared with the final early warning threshold and an early warning is issued.

2. The method for real-time monitoring of the crystallization process of Cistanche deserticola according to claim 1, characterized in that, The process of obtaining crystal grain geometric parameters, including grain area, perimeter, and the length and width of the minimum bounding rectangle, includes: Using an immersion high-resolution microscopic imaging probe, real-time image streams of the solid-liquid two-phase mixture inside the crystallization vessel are continuously acquired at a preset frame rate. Each acquired image frame is processed using an instance segmentation model to accurately identify and segment each individual crystal particle in the image. For each identified particle, image analysis algorithms are used to obtain the crystal particle's geometric parameters, including: particle area, outline perimeter, and the length and width of the minimum bounding rectangle. A screening algorithm is then applied to identify and exclude crystal agglomerates that are too large or have irregular shapes based on the crystal particle's geometric parameters.

3. The method for real-time monitoring of the crystallization process of Cistanche deserticola according to claim 1, characterized in that, The crystal morphological isomerism index satisfies the following expression: ； In the formula, Indicates the first The crystal morphological isomerism index of each crystal grain, dimensionless; Indicates the first The perimeter of the outline of a crystal grain; Indicates the first The length of the smallest bounding rectangle of a crystal grain; Indicates the first The width of the minimum bounding rectangle of a crystal grain; Represents the natural exponential function; Pi is the mathematical constant of a circle.

4. The method for real-time monitoring of the crystallization process of Cistanche deserticola according to claim 1, characterized in that, The crystal neighborhood perturbation degree satisfies the following expression: ； In the formula, Indicates the first The crystal neighborhood perturbation degree of each crystal grain, dimensionless; Indicates the first The set of all crystal particles in the circular neighborhood of a crystal particle, excluding the central crystal. Indicates the first In the circular neighborhood of the nth crystal grain Crystal morphological isomerism index of individual crystal particles; Indicates the first The first crystal grain and the first In the circular neighborhood of the nth crystal grain Euclidean distance between individual crystal grains; It is a very small positive number, used to prevent the inner denominator from being zero; It is a very small positive number used to prevent the outer denominator from being zero.

5. The method for real-time monitoring of the crystallization process of Cistanche deserticola according to claim 1, characterized in that, The calculation of the crystal signal-to-noise ratio includes: The crystal morphological heterogeneity index of the current crystal particle is compared with the perturbation degree of the crystal neighborhood, and the calculation result is used as the crystal signal-to-noise ratio. The crystal signal-to-noise ratio is used to evaluate the degree of anomaly of the current crystal particle relative to its surrounding background.

6. The method for real-time monitoring of the crystallization process of Cistanche deserticola according to claim 1, characterized in that, The calculated crystal-based comprehensive early warning value satisfies the following expression: ； In the formula, Indicates the first The overall early warning value of each crystal particle is dimensionless. Indicates the first Crystal morphological isomerism index of individual crystal particles; Indicates the first The signal-to-noise ratio of each crystal particle; This represents the mean crystal morphological isomerism index of all crystal particles; It is a very small dimensionless positive number used to prevent the denominator from being zero; It is a natural exponential function.

7. The method for real-time monitoring of the crystallization process of Cistanche deserticola according to claim 1, characterized in that, The step of performing time-series verification on candidate particles with high matching degree includes: For candidate crystal particles whose comprehensive early warning value first exceeds the average of all comprehensive early warning values, the system initiates tracking and records its subsequent continuous values. The time series of crystal morphological heterogeneity index in the frame image is denoted as the crystal growth sequence to be tested. The crystal growth sequence to be tested is dynamically time-warped and matched with the typical abnormal crystal growth sequences pre-stored in the expert knowledge base. The abnormal state of the candidate particle is confirmed in time only when the matching degree between the crystal growth sequence to be tested and a certain typical abnormal crystal growth sequence is higher than the average matching similarity value.

8. The method for real-time monitoring of the crystallization process of Cistanche deserticola according to claim 1, characterized in that, The global confirmation of abnormal states for candidate particles confirmed through time-series analysis includes: For candidate particles confirmed through time-series analysis, the system further performs real-time statistics on the global distribution of crystal morphology heterogeneity indices of all crystal particles within the current field of view, and calculates the 95th percentile of all crystal particles; then, based on the absolute difference between the crystal morphology heterogeneity index and the 95th percentile, the relative significance of the candidate particle is calculated; when the relative significance of the candidate particle exceeds a certain multiple, the abnormal state of the candidate particle is globally confirmed.

9. A method for real-time monitoring of the crystallization process of Cistanche deserticola according to claim 1, characterized in that, The step of comparing globally confirmed candidate particles with the final warning threshold and issuing a warning includes: Only when a candidate particle Only after passing the temporal growth fingerprint matching verification and global confirmation of abnormal states can the system confirm it as a high-confidence warning target. The system then compares its final crystal comprehensive warning value with a fixed final decision threshold. Once the condition that the crystal comprehensive warning value exceeds the final decision threshold is met, the system immediately triggers the highest-level warning and highlights the candidate particle. The location, image, and all computational features of the identified anomalous crystal nuclei.

10. A real-time monitoring system for the crystallization process of Cistanche deserticola, characterized in that, include: A processor and a memory, the memory storing computer program instructions that, when executed by the processor, implement a real-time monitoring method for the crystallization process of Cistanche deserticola according to any one of claims 1-9.

Images