Near-infrared online detection method for volatile oil component in medicinal material drying process

Abstract

The invention relates to the technical field of chemical or physical analysis, in particular to a near-infrared online detection method for volatile oil components in a medicinal material drying process. The technical problem that it is difficult to determine the optimal drying time of the medicinal materials based on a static sampling method is solved. The method comprises the following steps: determining abnormal wave data of a target monitoring point at a plurality of monitoring moments based on sensor time sequence data; determining a first state mutation moment of the target monitoring point based on the medicinal material data and the drying abnormity possibility; determining an optimal drying index of the target monitoring point based on the abnormal wave data and the first state sudden change moment; and determining a target monitoring moment in the plurality of monitoring moments based on the optimal drying index, wherein the target monitoring moment is used for determining the medicinal material drying duration. The method is used for chemical or physical analysis scenes of medicinal material drying.

Description

Technical Field

[0001] This invention relates to the field of infrared light analysis of material composition technology, specifically to a near-infrared online detection method for volatile oil components during the drying process of medicinal materials. Background Technology

[0002] After harvesting, medicinal herbs typically contain high moisture content, making them susceptible to microbial and pest growth. Therefore, drying processes are necessary to reduce moisture for preservation while minimizing the loss of volatile oils and other active ingredients. Near-infrared spectroscopy, as a rapid and non-destructive detection method, has been explored for monitoring volatile oil components during the drying process, aiming to accurately determine the drying endpoint. Currently, existing near-infrared detection methods primarily involve randomly selecting static samples from batches of medicinal herbs for spectral acquisition and quantitative analysis based on the reflectance of characteristic volatile oil bands to assess content changes. However, medicinal herb drying is inherently a dynamic process, accompanied by moisture evaporation and continuous changes in the physical state of the herbs. This static sampling method struggles to reflect the true state of the entire batch in real time, resulting in unsatisfactory monitoring of volatile oil components and difficulty in determining the optimal drying time. Summary of the Invention

[0003] To address the technical problems of existing methods based on static sampling, which struggle to reflect the true state of a batch of medicinal materials in real time, resulting in unsatisfactory monitoring of volatile oil components and difficulty in determining the optimal drying time, this invention aims to provide a near-infrared online detection method for volatile oil components during the drying process of medicinal materials. The specific technical solution adopted is as follows: In a first aspect, the present invention provides a near-infrared online detection method for volatile oil components during the drying process of medicinal materials. The method includes: determining abnormal wave data of a target monitoring point at multiple monitoring times based on sensor time-series data; the sensor time-series data includes: infrared spectral data and medicinal material data; the abnormal wave data includes at least one of the following: the number of abnormal waves at the monitoring time, and the probability of drying anomalies at the monitoring time; determining a first state change moment of the target monitoring point based on the medicinal material data and the probability of drying anomalies; the first state change moment is the moment when the state of volatile oil in the medicinal material undergoes a sudden change during the drying process; determining an optimal drying index for the target monitoring point based on the abnormal wave data and the first state change moment; the optimal drying index is used to characterize the advantage of stopping drying at the target moment; determining a target monitoring moment among multiple monitoring times based on the optimal drying index; the target monitoring moment is used to determine the drying duration of the medicinal material.

[0004] In conjunction with the first aspect mentioned above, in one possible implementation, the target monitoring point is any one of multiple monitoring points. The method specifically includes: determining the temporal variation difference of reflectance at at least one spectral wavelength among multiple monitoring points based on infrared spectral data; determining the drying anomaly probability at at least one spectral wavelength based on the deviation degree of the reflectance temporal variation difference from a preset reflectance temporal variation difference; determining the number of spectral wavelengths among multiple monitoring points that meet preset conditions as the number of anomaly waves; the preset conditions include at least one of the following: the result after normalization of the drying anomaly probability is greater than a first preset threshold.

[0005] In conjunction with the first aspect mentioned above, in one possible implementation, the method specifically includes: determining the reflectance difference values ​​of multiple monitoring points at adjacent monitoring times of at least one spectral wavelength; determining a set of cooperative monitoring points based on the reflectance difference values; ensuring that the reflectance difference values ​​of the monitoring points in the cooperative monitoring point set have the same sign; and determining the temporal variation difference of reflectance based on the absolute value of the difference between the reflectance difference value of the target monitoring point and the reflectance difference value of each monitoring point in the cooperative monitoring point set.

[0006] In conjunction with the first aspect mentioned above, in one possible implementation, the infrared spectral data includes: characteristic wavelengths of volatile oils and reference wavelengths of moisture. The method specifically includes: determining the shrinkage rate of the medicinal material at the target monitoring point based on image data; determining the weight change rate of the medicinal material at the target monitoring point based on weight data; determining a spectral correlation change index based on the reflectance changes of the characteristic wavelengths of volatile oils and the reference wavelengths of moisture; determining a state change probability value based on the temporal changes of the shrinkage rate, weight change rate, spectral correlation change index, and the possibility of drying anomalies; and determining the moment when the state change probability value is greater than a second preset threshold as the first state change moment.

[0007] In conjunction with the first aspect above, in one possible implementation, the method specifically includes: determining the time series of reflectance changes of the characteristic wavelength of volatile oil and the time series of reflectance changes of the reference wavelength of moisture; determining the correlation coefficient between the time series of reflectance changes of the characteristic wavelength of volatile oil and the time series of reflectance changes of the reference wavelength of moisture; and determining the spectral correlation change index based on the mean of multiple correlation coefficients.

[0008] In conjunction with the first aspect mentioned above, in one possible implementation, the method specifically includes: determining a mutation time priority index based on the duration from the first drying start time to the first state mutation time and the duration from the second drying start time to the second state mutation time; the second state mutation time is the state mutation time of a historically similar drying process; the historically similar drying process is a historical drying process whose similarity index with the target monitoring point's drying is greater than a third preset threshold; determining the difference in volatile oil content between the target monitoring point and the end of a historically similar drying process; determining a drying effectiveness index based on the mutation time priority index, the number of abnormal waves, the probability of drying anomalies, and the difference in volatile oil content; and determining at least one high-efficiency drying monitoring point based on the drying effectiveness index of each monitoring time among multiple monitoring times. The optimal drying index is determined based on the drying effectiveness index of high-efficiency drying monitoring points, the historical drying effectiveness index of high-efficiency drying monitoring points, and the number of high-efficiency drying monitoring points.

[0009] In conjunction with the first aspect mentioned above, in one possible implementation, the method specifically includes: normalizing the optimal drying index at multiple monitoring times; and determining the first monitoring time where the normalization result is greater than a fourth preset threshold as the target monitoring time based on the result of the normalization of the optimal drying index at multiple monitoring times.

[0010] In conjunction with the first aspect mentioned above, in one possible implementation, the method further includes: acquiring historical drying data; the historical drying data includes: the initial spectral baseline reflectance of the historical drying process, the initial reflectance of the moisture reference wavelength, and the time-series curve of environmental parameters; determining the comprehensive similarity between the historical drying process and the target drying process based on the initial spectral baseline reflectance of the historical drying process, the initial reflectance of the moisture reference wavelength, and the time-series curve of environmental parameters; determining historically similar drying processes in the historical drying process; the historically similar drying processes are historical drying processes with a comprehensive similarity greater than a fifth preset threshold.

[0011] In conjunction with the first aspect mentioned above, in one possible implementation, the method further includes: installing multiple monitoring points in the target drying area; installing spectral detection equipment, image acquisition equipment, and weighing equipment at the multiple monitoring points; and collecting infrared spectral data and medicinal material data based on a preset cycle.

[0012] In conjunction with the first aspect mentioned above, in one possible implementation, the method further includes: displaying or transmitting the target monitoring time and the optimal drying index for each monitoring point via an output device.

[0013] The present invention has the following beneficial effects: This invention constructs a comprehensive monitoring system, from identifying abnormal wave data to determining the moment of state change, and then to calculating the optimal drying index, enabling dynamic tracking and quantitative evaluation of volatile oil components during the drying process of medicinal materials. This method effectively overcomes the limitations of traditional static detection, significantly improving the accuracy and reliability of volatile oil state judgment through multi-dimensional data fusion analysis. This provides a scientific basis for determining the optimal drying time, avoiding volatile oil loss or energy waste caused by stopping drying too early or too late. It solves the technical problems of existing methods based on static sampling, which struggle to reflect the true state of the entire batch of medicinal materials in real time, resulting in unsatisfactory monitoring of volatile oil components and difficulty in determining the optimal drying time. Attached Figure Description

[0014] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a schematic flowchart of a near-infrared online detection method for volatile oil components during the drying process of medicinal materials, provided in one embodiment of the present invention. Figure 2 This is a flowchart illustrating another near-infrared online detection method for volatile oil components during the drying process of medicinal materials, provided in an embodiment of the present invention. Figure 3 This is a flowchart illustrating another near-infrared online detection method for volatile oil components during the drying process of medicinal materials, provided in an embodiment of the present invention. Figure 4 This is a flowchart illustrating another near-infrared online detection method for volatile oil components during the drying process of medicinal materials, provided in an embodiment of the present invention. Figure 5 This is a flowchart illustrating another near-infrared online detection method for volatile oil components during the drying process of medicinal materials, provided in an embodiment of the present invention. Figure 6 This is a flowchart illustrating another near-infrared online detection method for volatile oil components during the drying process of medicinal materials, provided in an embodiment of the present invention. Figure 7 This is a flowchart illustrating another near-infrared online detection method for volatile oil components during the drying process of medicinal materials, provided in an embodiment of the present invention. Figure 8 This is a flowchart illustrating another near-infrared online detection method for volatile oil components during the drying process of medicinal materials, provided in an embodiment of the present invention. Figure 9 This is a flowchart illustrating another near-infrared online detection method for volatile oil components during the drying process of medicinal materials, provided in an embodiment of the present invention. Figure 10 This is a flowchart illustrating another near-infrared online detection method for volatile oil components during the drying process of medicinal materials, provided as an embodiment of the present invention. Detailed Implementation

[0016] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a near-infrared online detection method for volatile oil components during the drying process of medicinal materials according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0017] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0018] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0019] The terms "first" and "second," etc., used in the specification and drawings of this invention are used to distinguish different objects or to distinguish different treatments of the same object, rather than to describe a specific order of objects.

[0020] Furthermore, the terms "comprising" and "having," and any variations thereof, used in the description of this invention are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include other steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices. It should be noted that in embodiments of this invention, words such as "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in embodiments of this invention should not be construed as preferred or advantageous over other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0021] In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0022] The following description, in conjunction with the accompanying drawings, details the specific scheme of a near-infrared online detection method for volatile oil components during the drying process of medicinal materials provided by the present invention.

[0023] Please see Figure 1 The present invention illustrates a near-infrared online detection method for volatile oil components during the drying process of medicinal materials according to an embodiment of the present invention. The method includes the following steps S101-S104, which will be described in detail below.

[0024] S101. Determine the abnormal wave data of the target monitoring point at multiple monitoring times based on sensor time series data.

[0025] The sensor time-series data includes: infrared spectral data and medicinal material data; the abnormal wave data includes at least one of the following: the number of abnormal waves at the monitoring time and the probability of drying abnormalities at the monitoring time.

[0026] One possible implementation involves acquiring time-series sensor data, including infrared spectral data and medicinal material data, as the processing object. Abnormal signals deviating from the normal variation patterns of the population are identified. When the variation trend of a specific monitoring point at a certain wavelength differs significantly from that of most monitoring points, and this difference cannot be explained by the known characteristic wavelength variation patterns of volatile oils, that wavelength can be determined to be an abnormal wave.

[0027] Among possible implementations, other methods exist for identifying anomalous wave data. These include, but are not limited to: using principal component analysis to extract the main variation patterns from the entire spectral data and identifying residual components orthogonal to the main variation patterns as anomalous signals; using anomaly detection algorithms such as isolated forests to directly identify outliers in the spectral characteristics of each monitoring point; or constructing an autoencoder neural network to identify wavelength components with large reconstruction errors as anomalous waves. These alternative solutions all achieve the core function of identifying anomalous wavelengths that deviate from normal variation patterns through a data-driven approach.

[0028] S102. Based on the medicinal material data and the probability of drying anomalies, determine the first state change moment of the target monitoring point.

[0029] The first state mutation moment is the moment when the state of the volatile oil in the medicinal material undergoes a sudden change during the drying process.

[0030] In one possible implementation, medicinal material data and the probability of drying anomalies are used as input data. Based on multi-dimensional information such as changes in the physical state of medicinal materials, changes in the correlation between the spectral characteristics of volatile oil and moisture, and abnormal fluctuations in the system, the critical moment when the state of volatile oil changes abruptly. Among these, changes in the physical state of medicinal materials reflect the evolution process of the macroscopic characteristics of medicinal materials, changes in the spectral correlation index characterize the changes in the coupling relationship between volatile oil volatilization and moisture evaporation, and the temporal changes in the probability of drying anomalies reflect the stability characteristics of the system monitoring data.

[0031] S103. Determine the optimal drying index of the target monitoring point based on abnormal wave data and the moment of first state change.

[0032] The optimal drying index is used to characterize the advantage of stopping drying at the target time.

[0033] One possible implementation involves acquiring monitoring data, including anomalous wave data and the moment of the first state abrupt change, as the processing object. A multi-dimensional quality assessment strategy is employed, establishing a quantitative evaluation system for drying effect by comprehensively considering the timeliness and stability of the drying process. When a specific monitoring point achieves optimal comprehensive evaluation results across the three dimensions of timeliness, stability, and product quality, the drying process at that monitoring point can be determined to have reached its optimal state.

[0034] Understandably, timeliness evaluation can effectively reflect the efficiency level of the drying process, stability evaluation reflects the level of control over the degree of disturbance to the drying process, and product quality evaluation is directly related to the quality standards of the final product. Together, the three constitute a complete drying quality assessment framework.

[0035] S104. Determine the target monitoring time among multiple monitoring times based on the optimal drying index.

[0036] Among them, the target monitoring time is used to determine the drying time of medicinal materials.

[0037] One possible implementation involves acquiring an optimal drying index sequence encompassing multiple monitoring times as the processing object. A strategy combining dynamic threshold judgment and first-time satisfaction of preset conditions is employed. By analyzing the temporal variation pattern of the optimal drying index, the optimal drying termination time is determined. When the optimal drying index first reaches a preset satisfactory level and meets stability requirements at a specific monitoring time, that time can be identified as the target monitoring time.

[0038] In one possible implementation, the optimal drying index at each monitoring time is first normalized and converted to a unified evaluation scale. Then, the optimal drying index after normalization is judged in chronological order to see if it exceeds a preset first threshold. When the first monitoring time that meets the preset condition is detected, it is determined as the target monitoring time and used as the optimal drying endpoint for the current batch of medicinal materials.

[0039] Understandably, normalization can eliminate the influence of data dimensions. For example, normalization by maximum and minimum values ​​can make the drying processes of different batches comparable. The judgment mechanism that first meets the preset conditions can ensure that drying is terminated in time when the quality requirements are met, thus avoiding the loss of volatile oils caused by over-drying.

[0040] The technical solution provided in the above embodiments can bring at least the following beneficial effects: This embodiment, by constructing a full-process monitoring system from abnormal wave data identification to the judgment of the moment of state change, and then to the calculation of the optimal drying index, realizes the dynamic tracking and quantitative evaluation of volatile oil components during the drying process of medicinal materials. This method can effectively overcome the limitations of traditional static detection, and significantly improve the accuracy and reliability of volatile oil state judgment through multi-dimensional data fusion analysis, thereby providing a scientific basis for determining the optimal drying time, avoiding volatile oil loss or energy waste caused by stopping drying too early or too late, thus solving the technical problems of existing methods based on static sampling that are difficult to reflect the real state of the entire batch of medicinal materials in real time, have unsatisfactory monitoring effects on volatile oil components, and are difficult to determine the optimal drying time.

[0041] In one possible implementation, the target monitoring point is any one of multiple monitoring points; please refer to [link / reference]. Figure 2 The present invention illustrates a near-infrared online detection method for volatile oil components during the drying process of medicinal materials according to an embodiment of the present invention. The method includes the following steps S201-S203, which will be described in detail below.

[0042] S201. Based on infrared spectral data, determine the temporal variation differences of reflectance at at least one spectral wavelength among multiple monitoring points.

[0043] One possible implementation involves acquiring infrared spectral data from multiple monitoring points as the processing object. An analysis strategy combining differential and synergistic approaches is employed, comparing the reflectance changes of each monitoring point at the same spectral wavelength to identify anomalous signals deviating from the overall pattern. When the reflectance change trend of a specific monitoring point at a certain wavelength is significantly inconsistent with that of most monitoring points, it can be determined that the monitoring point exhibits an anomalous change at that wavelength.

[0044] In one possible implementation, the reflectance difference value of each monitoring point at at least one spectral wavelength between adjacent monitoring times is calculated. This value reflects the magnitude and direction of reflectance change over time. Then, a set of co-monitoring points is determined based on the sign consistency of the reflectance difference values. This set consists of monitoring points with the same direction of change as the target monitoring point. If the set of co-monitoring points is not empty, the temporal variation difference of reflectance is determined by calculating the absolute value of the difference between the reflectance difference value of the target monitoring point and the reflectance difference values ​​of each monitoring point in the set of co-monitoring points. This difference value is used to quantify the degree of deviation of the target monitoring point from the group of co-monitoring points at a specific wavelength. If the set of co-monitoring points is empty, it indicates that the change trend of all other monitoring points at the current wavelength is opposite to that of the target monitoring point, which itself constitutes a significant anomaly. In this case, the temporal variation difference of reflectance is determined by averaging the absolute values ​​of the differences between the reflectance difference value of the target monitoring point and the reflectance difference values ​​of all other monitoring points. This difference value is used to quantify the degree of reverse deviation between the target monitoring point and all other monitoring points.

[0045] Understandably, when the target monitoring point changes in the opposite direction to all other monitoring points, it constitutes an extremely significant anomaly. Quantifying the degree of this anomaly by calculating its difference from all other monitoring points is technically logically sound.

[0046] S202. Based on the degree of deviation between the time-series variation difference of reflectance and the preset time-series variation difference of reflectance, determine the possibility of drying anomalies at at least one spectral wavelength.

[0047] One possible implementation involves acquiring data on temporal variations in reflectance as the processing object. An analytical strategy combining benchmark comparison and anomaly assessment is employed. By comparing the temporal variations in reflectance at each monitoring point with a preset reference benchmark, the probability of anomalies at specific spectral wavelengths is quantified. When the reflectance variation at a specific monitoring point at a certain wavelength deviates significantly from the reference benchmark, and this deviation cannot be explained by normal volatile oil variation patterns, it can be determined that there is a high probability of drying anomalies at that wavelength.

[0048] Understandably, comparing the difference index with the characteristic wavelength of volatile oil can effectively distinguish between normal component changes and abnormal interference, while the quantitative assessment of the degree of deviation reflects the significance of the abnormal signal. The greater the degree of deviation, the higher the possibility that the wavelength is interfered with by non-target factors.

[0049] For example, firstly, the temporal variation of reflectance at the target monitoring point at the current wavelength is compared with the temporal variation of reflectance at the same monitoring point across all known characteristic wavelengths of volatile oils, and the overall deviation is calculated. Then, this temporal variation of reflectance is compared with the average temporal variation of reflectance at the same wavelength across all monitoring points. Finally, based on a comprehensive evaluation of the two comparison results, the probability of drying anomalies at the current wavelength is determined. This probability value reflects the degree of risk of spectral signal distortion due to oxidation, derivative formation, or other anomalies.

[0050] For example, the wavelength difference index Satisfy the following formula 1: Formula 1 in, For monitoring points The difference value at wavelength j is used to characterize the rate and direction of reflectivity change. The number of monitoring points For monitoring point i and The absolute value of the difference in reflectance variation amplitude at wavelength j up to the k-th detection is used to characterize the difference in the degree of change between the two monitoring points; The coefficient of cooperation is the difference value at point i at wavelength j, which has the same positive and negative sign as the difference value at wavelength j. Therefore, the coefficient of cooperation at that point is denoted as [coefficient of cooperation]. Otherwise, it is 0, used to determine the monitoring point i and At wavelength of Whether the changing trends are consistent over time; It is a positive correction factor.

[0051] Understandably, in all division and logarithmic operations involved in this application, a smoothing mechanism is employed to prevent the computer program from crashing or generating invalid values ​​due to a zero denominator or zero input. Specifically, a positive correction factor ε (e.g., 0.001) is superimposed on the denominator term of the division operation or the argument term of the logarithmic function, thereby ensuring the robustness and feasibility of the algorithm under extreme conditions.

[0052] Understandably, the numerator, by weighting and summing the differences in reflectance changes among the co-monitoring points, reflects the overall degree of deviation between the target monitoring point and a group of points with similar trends. It serves a filtering function, including only monitoring points with consistent trends and excluding interference from points with opposite trends; denominator It is a normalization factor that ensures the results are not affected by the number of monitoring points, thus guaranteeing the comparability of the difference index; the wavelength difference index. This quantitatively characterizes the degree of deviation between the spectral variation behavior of the target monitoring point at a specific wavelength and time and other monitoring points in the same batch. The larger the value, the more abnormal the variation pattern of the monitoring point at that wavelength, and the higher the possibility of interference.

[0053] For example, the probability of drying abnormalities satisfies the following formula 2: Formula 2 in, For target monitoring points At wavelength No. The difference index of wavelength during the second detection; For the first During the second test, all monitoring points were at wavelength Difference index The average value; For target monitoring points In the Characteristic wavelengths of volatile oils In the Difference index at each test; A positive correction factor; The number of characteristic wavelengths of the effective volatile oil.

[0054] Understandably, when hour, Only population relative significance is calculated. (Molecular) The denominator represents the magnitude of the change in the target monitoring point at the current wavelength; This represents the average amplitude (or a minimum baseline value) of the variation across the entire batch at that wavelength. The ratio of the two directly measures the degree of anomalousness of the target point's variation amplitude relative to the batch average.

[0055] Understandably, when hour, The target wavelength difference index was calculated. Difference index with each known volatile oil characteristic wavelength The absolute deviation is calculated and averaged. This quantitatively characterizes the overall degree to which the variation pattern of the target wavelength deviates from the variation pattern of the standard volatile oil. By using absolute values ​​to sum rather than direct summation, it effectively avoids the problem of positive and negative values ​​of deviations from different characteristic wavelengths canceling each other out.

[0056] Understandably, the probability of drying anomalies comprehensively assesses the risk of interference with a specific wavelength by non-volatile oil factors. The higher the value, the higher the probability of interference with that wavelength due to abnormal factors such as oxidation and derivative formation, and the greater the necessity to label it as an interfering wavelength.

[0057] S203. Determine the number of spectral wavelengths that meet the preset conditions among multiple monitoring points as the number of abnormal waves.

[0058] Among them, the preset conditions include at least one of the following: the result after normalization of the probability of drying abnormality is greater than the first preset threshold.

[0059] For example, the first preset threshold can be specifically, for instance, the preset condition is that the result after normalization of the probability of drying abnormality is greater than 0.5.

[0060] One possible implementation involves acquiring drying anomaly probability data for each monitoring point at each spectral wavelength as the processing object. An analysis strategy combining threshold judgment and statistical analysis is employed. By setting reasonable judgment conditions, the abnormal state at each wavelength is binarized for judgment, thereby statistically determining the overall scale of the interfered wavelengths. When the drying anomaly probability at a specific monitoring point at a certain wavelength exceeds a preset judgment standard, that wavelength can be included in the statistical range of abnormal waves.

[0061] Understandably, normalization can convert the probability of drying anomalies into a standard dimension, making it easier to set a unified judgment threshold; while the number of wavelengths that meet the conditions reflects the extent of interference with the spectral data of the monitoring point. The more wavelengths that meet the conditions, the more common the abnormal situation in the drying process of the monitoring point.

[0062] For example, firstly, the probability of drying anomalies at each spectral wavelength is normalized so that its values ​​fall within a uniform numerical range; then, a preset condition is set as a judgment threshold, which can be a fixed numerical threshold or a dynamic threshold based on overall distribution characteristics; finally, the number of spectral wavelengths at the monitoring point that meet the preset condition is counted, and this number is determined as the number of anomalous waves. This value, as an important component of the anomalous wave data, reflects the overall quality of the spectral data at the monitoring point.

[0063] The technical solution provided in the above embodiments can bring at least the following beneficial effects: This embodiment, by analyzing the temporal differences in reflectance changes at multiple monitoring points at specific spectral wavelengths, can systematically identify spectral interference caused by non-target components during the drying process. This method utilizes a lateral comparison mechanism of monitoring point data within a batch, enhancing the robustness of abnormal wavelength detection and effectively avoiding misjudgments caused by localized deterioration of medicinal materials or environmental anomalies, thus providing a purer and more reliable data foundation for subsequent analysis.

[0064] Please see Figure 3 The diagram shows a near-infrared online detection method for volatile oil components during the drying process of medicinal materials according to an embodiment of the present invention. The method includes the following steps S301-S303, which will be described in detail below.

[0065] S301. Determine the reflectance difference value of multiple monitoring points at at least one adjacent monitoring time of a spectral wavelength.

[0066] One possible implementation involves acquiring time-series reflectance data from multiple monitoring points at at least one spectral wavelength as the processing object. An analysis strategy of calculating the difference between adjacent time points is employed to quantify the rate and direction of change of the spectral signal over time by calculating the reflectance change between adjacent monitoring times. When the reflectance of a specific monitoring point at a certain wavelength changes significantly between adjacent time points, the corresponding difference value can effectively capture this change characteristic.

[0067] Understandably, differential calculation can eliminate the influence of spectral baseline drift and highlight the dynamic changes in reflectance; while the selection of adjacent time points determines the temporal resolution of change detection. The shorter the interval, the more refined the capture of the change process.

[0068] For example, firstly, a reflectance sequence of each monitoring point at the at least one spectral wavelength is obtained in chronological order; then, a difference operation is performed on the reflectance values ​​of adjacent monitoring times in the sequence, that is, the reflectance value of the later time is subtracted from the reflectance value of the previous time; finally, the calculated difference is determined as the reflectance difference value. This value is a vector parameter, whose absolute value represents the magnitude of change and whose sign represents the direction of change.

[0069] For example, taking the detection of monitoring point i at wavelength j as an example, the reflectance of wavelength j obtained from each detection of this point is arranged in time sequence, and its first-order difference is obtained. For example, in the k-th detection, the difference value corresponding to wavelength j is: ,in, Let i be the reflectance of monitoring point i at wavelength j during the k-th detection. The reflectance of monitoring point i at wavelength j during the (k-1)th detection.

[0070] S302. Determine the set of collaborative monitoring points based on reflectivity difference values.

[0071] Among them, the reflectance difference values ​​of the monitoring points in the collaborative monitoring point set have the same sign.

[0072] One possible implementation involves acquiring reflectance difference values ​​from multiple monitoring points at at least one spectral wavelength as the processing object. An analysis strategy based on trend consistency identification is employed, comparing the sign characteristics of the reflectance difference values ​​from each monitoring point to filter out a group of monitoring points with the same trend as the target monitoring point. When the sign of the reflectance difference value between a specific monitoring point and the target monitoring point is consistent at the same wavelength, it indicates that they exhibit coordinated spectral change behavior at that wavelength.

[0073] Understandably, the sign of the reflectance difference value can effectively characterize the directional characteristics of spectral changes, with a positive sign indicating increased reflectance and a negative sign indicating decreased reflectance. The judgment of the consistency of the signs reflects the degree of coordination of the spectral change trends of each monitoring point. The higher the consistency, the more uniform the change pattern of each monitoring point at this wavelength.

[0074] For example, firstly, the reflectance difference between the target monitoring point and all other monitoring points at a specific spectral wavelength is obtained; then, the sign of the reflectance difference between each monitoring point and the target monitoring point is compared one by one; finally, all monitoring points with the same sign are grouped into a co-monitoring point set. This set represents a group of monitoring points that have the same trend of change as the target monitoring point at the current wavelength.

[0075] For example, up to the k-th detection, the monitoring points If the growth trend is the same at wavelength j, meaning the difference value is the same sign as the difference value at wavelength j for point i, then the synergy coefficient at that point is marked. Otherwise, it is 0.

[0076] S303. Based on the absolute value of the difference between the reflectance difference value of the target monitoring point and the reflectance difference value of each monitoring point in the collaborative monitoring point set, determine the difference in reflectance over time.

[0077] One possible implementation involves obtaining the reflectance difference values ​​of both the target monitoring point and each monitoring point in the co-monitoring point set as the processing object. An analytical strategy combining difference quantification and statistical analysis is employed. By calculating the degree of difference in reflectance variation between the target monitoring point and the co-monitoring point group, the deviation of the target monitoring point from the group's variation pattern is quantified. When the absolute value of the difference between the reflectance difference value of the target monitoring point and the reflectance difference values ​​of each monitoring point in the co-monitoring point set is large, it indicates that the variation behavior of that monitoring point at the current wavelength differs significantly from the group pattern.

[0078] For example, firstly, the reflectance difference value of the target monitoring point at a specific wavelength is obtained, and simultaneously, the reflectance difference values ​​of all monitoring points in the collaborative monitoring point set at the same wavelength are obtained. Then, the absolute value of the difference between the reflectance difference value of the target monitoring point and that of each monitoring point in the collaborative monitoring point set is calculated. Finally, all absolute difference values ​​are statistically processed, and their mean is determined as the reflectance temporal variation difference. This difference value reflects the degree of deviation of the target monitoring point from the variation amplitude of the collaborative monitoring point group at a specific wavelength.

[0079] The technical solution provided in the above embodiments can bring at least the following beneficial effects: This embodiment can accurately quantify the spectral change differences between target monitoring points and cooperating points by calculating the reflectance difference value and screening the set of cooperating monitoring points based on sign consistency. This method effectively captures the cooperating change patterns that should exist in the same batch of medicinal materials under similar drying conditions, thereby accurately identifying abnormal fluctuations that deviate from the population trend and improving the sensitivity and specificity of interference wavelength identification.

[0080] In one possible implementation, the infrared spectral data includes: characteristic wavelengths of volatile oils and reference wavelengths of moisture; please refer to [link to relevant documentation]. Figure 4 The present invention illustrates a near-infrared online detection method for volatile oil components during the drying process of medicinal materials according to an embodiment of the present invention. The method includes the following steps S401-S405, which will be described in detail below.

[0081] S401. Based on image data, determine the shrinkage rate of medicinal materials at the target monitoring point.

[0082] One possible implementation involves acquiring time-series image data of the medicinal materials at the target monitoring point as the processing object. An analytical strategy combining image analysis and rate of change calculation is employed. By identifying and quantifying the relative changes in the apparent area of ​​the medicinal materials between adjacent monitoring times, the degree of shrinkage is determined. When the medicinal materials shrink during the drying process, their apparent area decreases accordingly. By calculating the area change ratio between adjacent times, the dynamic process of shrinkage can be effectively characterized.

[0083] It is understandable that image grayscale processing can simplify the subsequent image analysis process, and accurate identification of the medicinal material region is a prerequisite for calculating the apparent area; the calculation of the area ratio directly reflects the relative magnitude of the shrinkage of the medicinal material between adjacent time points, and the smaller the ratio, the greater the degree of shrinkage.

[0084] For example, the images of medicinal materials collected at each monitoring time are first converted to grayscale; then, the medicinal material regions are identified and extracted using an image segmentation algorithm, and the number of pixels in the region is calculated as a quantitative indicator of the apparent area of ​​the medicinal material; finally, the ratio of the medicinal material area at the current monitoring time to the medicinal material area at the previous monitoring time is calculated. This parameter quantitatively characterizes the relative change in the shrinkage of the medicinal material between adjacent monitoring times.

[0085] Understandably, this step, by establishing an objective and quantitative calculation mechanism for the ratio of medicinal material area, provides key physical change parameters for accurately determining the moment of abrupt change in the state of volatile oils, thereby enhancing the system's ability to integrate and analyze the multi-dimensional characteristics of the drying process.

[0086] S402. Based on the weight data, determine the weight change rate of the medicinal materials at the target monitoring point.

[0087] One possible implementation involves acquiring time-series weight data of medicinal materials at the target monitoring point as the processing object. An analysis strategy based on the ratio of adjacent time points is employed to determine the weight loss rate of the medicinal materials by quantifying the relative change in material quality between adjacent monitoring times. As moisture evaporates during the drying process, the weight of the medicinal materials decreases accordingly. By calculating the weight change ratio between adjacent time points, the dehydration efficiency of the drying process can be effectively characterized.

[0088] Understandably, continuous monitoring of weight data can capture dynamic quality changes during the drying process, while the calculation of the weight ratio between adjacent moments directly reflects the relative rate of moisture evaporation from the medicinal materials. The smaller the ratio, the faster the moisture evaporates and the more significant the drying process.

[0089] For example, firstly, the weight data of the medicinal materials collected at the target monitoring point at the current monitoring time k, and the weight data of the medicinal materials collected at the previous monitoring time k-1 are obtained; then, the ratio of the weight at the current time to the weight at the previous time is calculated. This parameter quantitatively characterizes the degree of relative change in the weight of the medicinal materials between adjacent monitoring times and is one of the key physical parameters reflecting the drying process.

[0090] S403. Determine the spectral correlation change index based on the reflectance changes of the characteristic wavelength of volatile oil and the reference wavelength of moisture.

[0091] One possible implementation involves acquiring time-series reflectance data containing the characteristic wavelengths of volatile oils and a moisture reference wavelength. An analytical strategy combining correlation analysis and statistical evaluation is employed to determine the changing coupling relationship between moisture interference and volatile oil volatilization by quantifying the correlation between reflectance variation trends among different characteristic wavelengths. When the reflectance variation trends of the volatile oil characteristic wavelength and the moisture reference wavelength are highly correlated, it indicates that volatile oil detection is still affected by moisture interference; when the correlation weakens, it indicates that volatile oil volatilization is beginning to dominate.

[0092] Understandably, the construction of a time series of reflectance changes is the basis for correlation analysis, and the calculation of correlation coefficients can effectively quantify the degree of synchronization between the changes of two wavelength sequences; the statistical processing of multiple correlation coefficients can improve the stability of correlation assessment and avoid interference from local fluctuations.

[0093] For example, firstly, time series sequences of reflectance changes for the characteristic wavelength of volatile oil and the reference wavelength of moisture are constructed separately; then, the Pearson correlation coefficient between each pair of time series sequences of reflectance changes for the characteristic wavelength of volatile oil and the reference wavelength of moisture is calculated; finally, the absolute value of all correlation coefficients is averaged, and the absolute value of this average is determined as the spectral correlation change index. This index comprehensively characterizes the degree of correlation between volatile oil volatilization behavior and moisture evaporation; the lower the value, the more independent the volatile oil volatilization is from moisture changes.

[0094] S404. Based on the temporal changes in shrinkage rate, weight change rate, spectral correlation change index, and the possibility of drying anomalies, determine the probability value of state change.

[0095] One possible implementation involves acquiring data on the temporal changes in shrinkage rate, weight change rate, spectral correlation index, and the likelihood of drying anomalies. An analytical strategy combining multi-parameter fusion and mutation feature identification is employed. By comprehensively evaluating characteristics across multiple dimensions, including changes in physical state, mass, spectral features, and system stability, the likelihood of a mutation in the volatile oil state is determined. When the physical state of the medicinal material tends to stabilize, mass change slows, spectral correlation decreases, and abnormal system fluctuations increase, it indicates that the volatile oil state may be at a critical point of mutation.

[0096] Understandably, the shrinkage change rate and weight change rate reflect the stability of the macroscopic physical state of the medicinal materials, and their values ​​approaching 1 indicate that the physical changes tend to be gradual; the spectral correlation change index characterizes the degree of reduction of moisture interference, and its decrease indicates that the volatilization of volatile oils begins to dominate; the temporal change of the probability of drying anomalies reflects the stability of the system monitoring data, and its increase may indicate the arrival of the state transition period.

[0097] For example, the probability value of a state change The following formula 3 is satisfied: Formula 3 in, Let be the mean of the reflectance changes at the k-th detection for all reference wavelengths. The difference index between the characteristic wavelength and the reference wavelength of each volatile oil up to the kth detection. The mean of the pairwise Pearson coefficients of a time series. For this point at various wavelengths Compared to the previous monitoring The average growth rate; The weight ratio is the ratio of the weight of monitoring point i in the kth measurement to the weight in the (k-1)th measurement, which is used to characterize the rate of water evaporation. The ratio is the area ratio, which is the ratio of the area of ​​the medicinal material detected at monitoring point i in the kth test to the area detected in the (k-1)th test, and represents the degree of shrinkage. The mean Pearson correlation coefficient of the time series of the difference index between the characteristic wavelength and the reference wavelength of volatile oil is used to characterize the correlation between changes in volatile oil and changes in moisture content. The absolute value of the mean of the reflectance difference values ​​for all reference wavelengths at the k-th detection is used to characterize the overall intensity of moisture change; The average increase in the probability of drying anomalies at each wavelength compared to the previous detection is used to characterize system instability; 0.001 is a smoothing term to prevent the denominator from being zero. 0.001 is much smaller than the absolute value of the mean of the reflectance difference, so its impact on the calculation results is negligible.

[0098] Among them, the molecular part This comprehensively reflects changes in physical state. A ratio close to 1 indicates a slowdown in physical change, potentially predicting a state transition. It's understandable that, in a drying scenario, the weight measured in the k-th test is less than or equal to the weight measured in the (k-1)-th test, and the area of ​​the medicinal material measured in the k-th test is less than or equal to the area measured in the (k-1)-th test. and The maximum value of each fraction is 1; the denominator part middle, A decrease indicates a reduced correlation between volatile oil evaporation and water evaporation. A decrease indicates a weakening of moisture change, and both suggest that volatile oils will dominate the change process. Growth indicates an increase in system uncertainty, which usually occurs during state transitions; the formula as a whole reflects the common indication of "slower physical changes, reduced moisture influence, decreased correlation, and increased system uncertainty" as a sudden change in state.

[0099] Understandably, the probability value of a state change. The probability of significant changes in volatile oil volatilization behavior at a specific monitoring time was quantitatively assessed. This is a dimensionless comprehensive evaluation index, designed to provide a relative evaluation index rather than an absolute quantity with definite physical units. A peak value indicates that a state transition has indeed occurred, providing a key criterion for determining the optimal drying stage. The probability of a sudden state change is a dimensionless comprehensive evaluation index, and its magnitude is used to compare the quality of drying conditions at different times or monitoring points.

[0100] S405. The moment when the probability value of state change is greater than the second preset threshold is determined as the first state change moment.

[0101] One possible implementation involves acquiring time-series data containing the probability values ​​of state mutations as the processing object. An analysis strategy combining extreme value identification and threshold judgment is employed. By analyzing the temporal variation patterns of the probability values ​​of state mutations, the key time points where significant changes in the state of the volatile oil occur are determined. When the probability value of a state mutation reaches a local maximum at a specific monitoring time and exceeds a preset judgment criterion, that time can be identified as the first state mutation moment.

[0102] For example, the moment when the probability of a state change is greater than 0.6 is determined as the first state change moment. It is understood that the exemplary value of 0.6 is a reference value determined based on historical drying experience.

[0103] For example, firstly, the probability values ​​of state abrupt changes at each monitoring time are arranged chronologically to form a sequence; then, local maxima points in the sequence are identified; next, these local maxima are compared with historical average levels; finally, the first monitoring time that simultaneously satisfies the local maximum condition and exceeds the historical average level is determined as the first state abrupt change time. This time characterizes the critical point in time when the volatile oil volatilization behavior undergoes a fundamental change during the drying process of medicinal materials.

[0104] The technical solution provided in the above embodiments can bring at least the following beneficial effects: This embodiment, by integrating images of medicinal materials, weight data, and spectral correlation changes of characteristic wavelengths of volatile oils and moisture, can comprehensively characterize the synergistic evolution of the physical state and chemical composition of medicinal materials during the drying process. This method utilizes the complementary advantages of multi-source data, overcoming the deficiency of single spectral data being easily affected by moisture, thereby more accurately capturing the key turning points of volatile oil volatilization and providing a clear basis for the division of drying stages.

[0105] Please see Figure 5 The diagram shows a near-infrared online detection method for volatile oil components during the drying process of medicinal materials according to an embodiment of the present invention. The method includes the following steps S501-S503, which will be described in detail below.

[0106] S501. Determine the time series of reflectance changes at the characteristic wavelengths of volatile oils and the time series of reflectance changes at the reference wavelengths of moisture.

[0107] One possible implementation involves acquiring time-series reflectance data containing characteristic wavelengths of volatile oils and reference wavelengths of moisture as the processing object. An analytical strategy combining time-series sequence construction and feature extraction is employed. By systematically organizing reflectance variation data at specific wavelengths, a feature sequence usable for correlation analysis is formed. Obtaining a complete and time-aligned reflectance variation sequence provides a reliable data foundation for subsequent spectral correlation analysis.

[0108] Understandably, the calculation of reflectance variation is the basis for constructing time series, which reflects the relative change of reflectance at each wavelength relative to the initial state or the previous state; while the construction of time series ensures the continuity and comparability of data in the time dimension, providing the necessary conditions for analyzing the trend of change.

[0109] S502. Determine the correlation coefficient between the time series of reflectance changes at the characteristic wavelength of volatile oil and the time series of reflectance changes at the reference wavelength of moisture.

[0110] One possible implementation involves acquiring time-series sequences of reflectance changes at characteristic wavelengths of volatile oils and reflectance changes at a moisture reference wavelength. A statistical correlation analysis strategy is employed to determine the coupling relationship between volatile oil evaporation and moisture evaporation by quantifying the degree of linear correlation between the two time-series sequences. When the trends of the two sequences are highly consistent, it indicates that volatile oil detection is still significantly affected by moisture changes; when the correlation weakens, it indicates that volatile oil evaporation begins to exhibit characteristics independent of moisture changes.

[0111] S503. Determine the spectral correlation change index based on the mean of multiple correlation coefficients.

[0112] One possible implementation involves obtaining correlation coefficients between multiple characteristic wavelengths of volatile oils and a reference wavelength for moisture as the processing object. A central tendency statistical analysis strategy is employed, and the overall correlation between volatile oil volatilization and moisture evaporation is determined by calculating the mean of multiple correlation coefficients. A high mean indicates that the changes in the characteristic wavelengths of volatile oils generally maintain a strong correlation with the reference wavelength for moisture; a low mean indicates that the volatilization of volatile oils is beginning to decouple from the influence of moisture interference.

[0113] For example, firstly, the correlation coefficients of all volatile oil characteristic wavelengths and moisture reference wavelengths are obtained; then, the arithmetic mean of these correlation coefficients is calculated; finally, this arithmetic mean is determined as the spectral correlation change index. This index comprehensively characterizes the average correlation strength between all volatile oil characteristic wavelengths and moisture reference wavelengths on reflectance changes.

[0114] The technical solution provided in the above embodiments can bring at least the following beneficial effects: This embodiment establishes a quantitative index of the correlation between moisture and volatile oil volatilization behavior by calculating the time series correlation between the reflectance of the characteristic wavelength of volatile oil and the reference wavelength of moisture. This method can keenly capture the critical state where moisture interference is weakened and volatile oil dominates volatilization, significantly improving the stability and accuracy of state judgment.

[0115] Please see Figure 6 The diagram shows a near-infrared online detection method for volatile oil components during the drying process of medicinal materials according to an embodiment of the present invention. The method includes the following steps S601-S605, which will be described in detail below.

[0116] S601. Based on the duration from the first start of drying time to the first state change time, and the duration from the second start of drying time to the second state change time, determine the change time priority index.

[0117] Among them, the second state change moment is the state change moment of a similar historical drying process; the similar historical drying process is a historical drying process in which the similarity index of the medicinal material drying at the target monitoring point is greater than the third preset threshold.

[0118] One possible implementation involves acquiring the first state transition moment of the current drying process and the second state transition moment of similar historical drying processes as the processing objects. A timeliness comparison analysis strategy is employed to evaluate the efficiency advantage of the current drying process by comparing the time difference required for the current drying process to reach the state transition with that of the historical baseline process. When the time required for the current process to reach the state transition is significantly shorter than that of the historical baseline, it indicates that the current drying process has higher efficiency, and its transition moment priority index is correspondingly higher.

[0119] For example, the mutation time priority index Satisfy the following formula 4: Formula 4 in, Indicates monitoring point In the The higher the value of the mutation time priority index during the second detection, the higher the efficiency of the current drying process. This indicates the duration of the sudden change in the current drying process, i.e., from the initial drying time to the current monitoring point. The duration of the first state transition moment; This represents the average duration of mutations in historically similar processes. From the historical database, samples are selected that are similar to the current monitoring point. The similarity index of the drying process is greater than a third preset threshold for all historical similar drying processes; the duration of each historical similar process from its start drying time to its state change time is calculated; finally, the arithmetic mean of these durations is obtained. , It is a positive correction factor.

[0120] For example, a historically similar drying process is defined as a time when the similarity index between the historical drying process and the target monitoring point is greater than 0.5. It is understood that the exemplary value of 0.5 is a reference value determined based on historical drying experience.

[0121] S602. Determine the difference in volatile oil content between the target monitoring point and the end of a similar historical drying process.

[0122] One possible implementation involves acquiring current volatile oil content data for the target monitoring point, as well as volatile oil content data from the end of similar historical drying processes. A difference quantification and comprehensive analysis strategy is employed to evaluate the effectiveness of the current drying process in retaining volatile oils by comparing the deviation of the current volatile oil content from historical baseline values. A small difference between the current volatile oil content and the historical baseline indicates that the current drying process has achieved a relatively ideal state in terms of volatile oil retention.

[0123] For example, firstly, the volatile oil content of the target monitoring point at the current monitoring time is calculated using near-infrared spectroscopy analysis; simultaneously, the volatile oil content at the end of all historical similar drying processes is obtained; then, the difference between the current volatile oil content and each historical volatile oil content is calculated; finally, these differences are summed to obtain the comprehensive volatile oil content difference. This difference reflects the degree of deviation of the current drying state from the historical optimal drying state at the end.

[0124] S603. Based on the priority index of abrupt change time, the number of abnormal waves, the probability of drying anomalies, and the difference in volatile oil content, the effective drying index is determined.

[0125] One possible implementation involves using indicators such as the priority index of abrupt change times, the number of abnormal waves, the average probability of drying anomalies, and the difference in volatile oil content as processing objects. A multi-dimensional comprehensive evaluation strategy is employed, integrating indicators such as timeliness, process stability, and final product quality to quantitatively assess the overall effectiveness of the drying process. When the drying process simultaneously possesses characteristics such as high efficiency, low interference, and good product quality, its corresponding drying effectiveness index will reach a high level.

[0126] Understandably, the mutation time priority index reflects the timeliness advantage of the drying process, and the larger the value, the higher the efficiency; the number of abnormal waves and the mean of the drying anomaly probability reflect the stability of the drying process, and the smaller the value, the lower the degree of interference in the process; the difference in volatile oil content characterizes the quality level of the final product, and the smaller the value, the closer the volatile oil retention effect is to the historical best level.

[0127] For example, the drying efficiency index The following formula 5 is satisfied: Formula 5 in, This represents the number of abnormal wavelengths of the medicinal material present at monitoring point i up to the k-th detection during this drying process. This represents the arithmetic mean of the drying anomalies corresponding to the abnormal wavelengths marked at monitoring point i during the k-th detection. The absolute value of the sum of the differences between the volatile oil content at monitoring point i up to the kth monitoring time and the volatile oil content at the end of all historical similar drying processes is the closest to the optimal drying state. It is a priority index for abrupt changes, reflecting the efficiency of the current drying process relative to historical processes; A positive correction factor with the same dimensions as the volatile oil content; , , To prevent the parameter tuning coefficient from returning to zero, it is a preset positive correction factor (e.g., 0.001). Its function is to ensure that when... , or Even when the actual value is zero or extremely small, the corresponding penalty factor remains a meaningful positive number, preventing individual defects from being ignored or causing computational instability.

[0128] What is understandable is the molecular part. Characterizing the advantage of timeliness, a ratio greater than 1 indicates that the volatile oil undergoes a sudden change earlier in the current process, resulting in higher drying efficiency; the denominator part The more abnormal wavelengths there are, the higher the probability of an anomaly, and the greater the degree of interference. The formula characterizes the retention of volatile oils; a smaller difference indicates that the content is closer to the historical optimum. The overall formula reflects an evaluation philosophy that encourages timeliness, penalizes interference, and requires content close to the ideal level. (Drying Effective Index) It is a comprehensive score of the drying quality of a single monitoring point. The Drying Effective Index is a dimensionless comprehensive evaluation index. Its value is used to compare the drying status at different times or different monitoring points. Its purpose is to obtain a relative evaluation index, rather than an absolute quantity with a clear physical unit. The higher the value, the better the drying process at that monitoring point performs in the three dimensions of efficiency, stability and final product quality, and the more worthy it is as a decision-making reference.

[0129] 604. Determine at least one high-efficiency drying monitoring point based on the drying effectiveness index at each of the multiple monitoring times.

[0130] One possible implementation involves acquiring the drying effectiveness index across multiple monitoring times as the processing object. An analysis strategy combining threshold screening and state assessment is employed. By setting reasonable judgment criteria, the drying effect of each monitoring point is graded, identifying monitoring points with excellent drying effects. When the drying effectiveness index of a specific monitoring point exceeds the preset excellent standard at the current monitoring time, that monitoring point can be classified as a high-efficiency drying monitoring point.

[0131] Understandably, the normalization of the drying effectiveness index can eliminate the dimensional differences between different monitoring batches, making the judgment criteria universally applicable; while continuous evaluation at multiple monitoring times can ensure that the efficient drying state has a certain degree of continuity, avoiding misjudgment caused by random fluctuations.

[0132] For example, firstly, the drying effectiveness index of each monitoring point at the current monitoring time is normalized so that its value falls within the standard range of 0 to 1; then, a third preset threshold of 0.7 is set as the criterion for efficient drying. This preset threshold is understood to be obtained through statistical analysis of multiple batches of historical data or through calibration experiments, and can be optimized and adjusted based on subsequent actual drying results; finally, all monitoring points whose normalized drying effectiveness index exceeds this threshold are identified as efficient drying monitoring points. These monitoring points represent the spatial locations where the drying effect reaches an excellent level at the current monitoring time.

[0133] S605. Determine the optimal drying index based on the effective drying index of high-efficiency drying monitoring points, the historical effective drying index of high-efficiency drying monitoring points, and the number of high-efficiency drying monitoring points.

[0134] One possible implementation involves acquiring the current and historical drying effectiveness indices, as well as the number of high-efficiency drying monitoring points, as the processing objects. A strategy combining group decision-making and growth stability analysis is employed. By comprehensively evaluating the number of high-efficiency monitoring points, the current drying effect, and the growth trend of the effect, the optimal drying level for the entire batch is determined. When the number of high-efficiency monitoring points is sufficient, the drying effect is significant, and the growth tends to stabilize, it indicates that the entire batch of medicinal materials has reached its optimal drying state.

[0135] For example, the optimal drying index Satisfy the following formula 6: Formula 6 in, This represents the number of monitoring points currently marked as high-efficiency drying points. Corresponding to each high-efficiency drying point The mean; The instantaneous growth rate of the drying effectiveness index at monitoring point i during the k-th detection is determined by comparing the drying effectiveness index at the current moment with the drying effectiveness index at the previous detection moment. The absolute difference between the mean and 1 represents the stability of the increase in drying effect. To smooth out the terms and prevent the denominator from being zero, 0.001 is a physical quantity much smaller than the magnitude of the stability of the increase in drying effect, and therefore its impact on the calculation results is negligible; The numerator part... The larger the number of monitoring points that represent the high efficiency standard, the better the drying effect of the entire batch of medicinal materials; the denominator is... The difference represents the stability of the drying effect growth. The closer the difference is to 0, the slower the growth and the closer to the drying endpoint. The formula as a whole reflects the termination judgment criterion of "most of the time reaching high efficiency and the effect growth slowing down", which balances the drying effect and drying efficiency.

[0136] Understandably, the optimal drying index The suitability of each monitoring point as the drying endpoint is evaluated at the batch level. The larger the value, the more the drying efficiency can be maximized while ensuring the retention of volatile oils. It is a quantitative indicator of the optimal drying time.

[0137] The technical solution provided in the above embodiments can bring at least the following beneficial effects: This embodiment establishes a comprehensive drying quality evaluation system by introducing comparative analysis of similar historical drying processes, combined with abnormal wave data evaluation and volatile oil content difference calculation. This method effectively eliminates the influence of batch-to-batch initial state differences, making the drying effect evaluation results more comparable and objective, and providing a reliable local basis for accurate decision-making.

[0138] Please see Figure 7 The present invention illustrates a near-infrared online detection method for volatile oil components during the drying process of medicinal materials according to an embodiment of the present invention. The method includes the following steps S701-S702, which will be described in detail below.

[0139] S701. Normalize the optimal drying index at multiple monitoring times.

[0140] One possible implementation involves obtaining the optimal drying index sequence encompassing multiple monitoring times as the processing object. A data standardization strategy is employed, performing scaling on the time-series data to eliminate the influence of different dimensions on data values ​​at different monitoring times, ensuring they fall within a uniform numerical range. After normalization, the optimal drying indices at different times will be comparable, facilitating subsequent unified threshold determination and time-series analysis.

[0141] S702. Based on the result of normalization of the optimal drying index at multiple monitoring times, the first monitoring time when the normalization result is greater than the fourth preset threshold is determined as the target monitoring time.

[0142] One possible implementation involves obtaining the normalized result of the optimal drying index across multiple monitoring points as the processing object. An analysis strategy combining time-series scanning and threshold judgment is employed. By sequentially detecting the normalized optimal drying index sequence over time, the first monitoring point reaching a satisfactory level is identified. When the normalized optimal drying index is first found to exceed a preset satisfactory standard during the time-series scanning process, that point can be determined as the optimal termination point of the drying process.

[0143] It is understandable that scanning in chronological order can ensure that the first moment that meets the conditions is identified, which is in line with the process requirement that the drying process needs to be terminated in time to retain volatile oils; while setting a reasonable fourth preset threshold can balance the drying effect and efficiency. If the threshold is too high, it may lead to over-drying, and if the threshold is too low, it may terminate the drying process too early.

[0144] For example, will Normalize to (0,1), and record the first time when the normalization result is greater than 0.7 as the target monitoring time.

[0145] The technical solution provided in the above embodiments can bring at least the following beneficial effects: This embodiment realizes the intelligent transformation from discrete monitoring points to overall batch decision-making through the normalization processing of the optimal drying index and threshold judgment. This method can identify the critical moment when the entire batch of medicinal materials reaches the optimal drying state in a timely and accurate manner, effectively avoiding the loss of volatile oils caused by over-drying while ensuring drying quality.

[0146] Please see Figure 8 The present invention illustrates a near-infrared online detection method for volatile oil components during the drying process of medicinal materials according to an embodiment of the present invention. The method includes the following steps S801-S803, which will be described in detail below.

[0147] S801, Obtain historical drying data.

[0148] The historical drying data includes: the initial spectral baseline reflectance of the historical drying process, the initial reflectance of the moisture reference wavelength, and the time series curves of environmental parameters.

[0149] One possible implementation involves retrieving relevant data from a historical database system, including historical drying process data, as the processing object. A strategy combining data retrieval and feature extraction is employed, systematically collecting and organizing historical drying records to provide comparable benchmark data for the current drying process. When complete and standardized historical drying data is obtained, a reliable reference system can be established for subsequent similarity analysis and effectiveness evaluation.

[0150] Understandably, the completeness of historical drying data directly affects the accuracy of similarity analysis, requiring the inclusion of multiple data types that can characterize the key features of the drying process; while data standardization is an important prerequisite for ensuring comparability between different batches of data.

[0151] S802. Determine the overall similarity between the historical drying process and the target drying process based on the initial spectral baseline reflectance of the historical drying process, the initial reflectance of the moisture reference wavelength, and the time series curve of environmental parameters.

[0152] One possible implementation involves acquiring the initial spectral baseline reflectance of the historical drying process and the target drying process, the initial reflectance of the moisture reference wavelength, and the time-series curves of environmental parameters as the processing objects. A multi-dimensional similarity fusion analysis strategy is employed to determine the overall similarity of the two drying processes in their initial states and external conditions by comprehensively evaluating the degree of similarity across different feature dimensions. When two drying processes exhibit high similarity across multiple dimensions, their overall similarity is correspondingly high.

[0153] For example, comprehensive similarity The following formula 7 is satisfied: Formula 7 in, For the first The The difference in reflectance between the initial spectral baseline of monitoring point i during the next drying process and the current drying process. The difference in initial reflectance between the two monitoring points at the reference wavelength for moisture content. The mean of the DTW distance between the temperature and humidity time series curves corresponding to the two monitoring points mentioned above; , , These are preset weighting coefficients, all positive numbers; used to adjust the relative contribution importance of the three difference measures to the overall similarity; for example, =0.2, =0.5, =0.3, , , The relative contribution importance of the three difference measures to the overall similarity is set, and subsequent optimization and correction can be made based on the actual effect. This invention does not limit this. , , These are preset scaling factors, all positive, used to eliminate differences in the dimensions and typical orders of magnitude of the three difference measures, enabling them to be weighted and summed fairly. , , These are the standard deviations, means, or empirically set constants of each difference measure in historical data; this invention does not limit these. It is understood that the three difference terms are combined through multiplication to reflect the joint requirements of multi-dimensional similarity; excessive difference in any dimension will significantly reduce the similarity index. It is understood that when conducting… Before function normalization, it is necessary to apply a standardization factor. , ,as well as Standardization is performed to unify the units of measurement; for example, the standardized factor. The standard deviation of the initial spectral baseline reflectance difference can be taken from historical data. The standard deviation of the initial reflectance difference at the reference wavelength of water can be used. The standard deviation of the dynamic time-warped distance of the environmental parameters can be taken as an example. Standardization ensures that parameters with different physical dimensions can be comprehensively calculated within a unified mathematical framework. After standardization, parameters with different dimensions are transformed into dimensionless relative differences, thus meeting the requirements of the exponential function for dimensionless input. Mapping the difference values ​​to the (0,1] interval, the similarity is 1 when the difference is 0, and the similarity approaches 0 as the difference increases; dynamic time-normalized distance. It can handle the scaling and offset of time-series curves on the time axis, and is more suitable for comparing the similarity of drying processes than Euclidean distance; comprehensive similarity This study comprehensively quantifies the similarity between the current drying process and historical processes in terms of initial state and external conditions. The closer the value is to 1, the more suitable the historical process is as a benchmark for the current process, and the higher the reliability of the evaluation results.

[0154] S803. Identify historically similar drying processes in the historical drying process.

[0155] Among them, historically similar drying processes are those with a comprehensive similarity greater than the fifth preset threshold.

[0156] One possible implementation involves acquiring comprehensive similarity data from multiple historical drying processes as the processing object. An analysis strategy combining threshold filtering and ranking is employed. By setting reasonable similarity criteria, reference processes highly similar to the current target drying process are selected from the historical data. When the comprehensive similarity of a historical drying process exceeds a preset qualification standard, it can be included in the set of historical similar drying processes.

[0157] For example, firstly, the overall similarity between each historical drying process and the current target drying process is calculated; then, a fifth preset threshold of 0.5 is set as the similarity qualification standard; finally, all historical drying processes with an overall similarity greater than 0.5 are identified as historically similar drying processes. These selected historical processes constitute a set of reference benchmarks for subsequent drying effect evaluation and quality comparison.

[0158] The technical solution provided in the above embodiments can bring at least the following beneficial effects: This embodiment establishes a precise historical data screening mechanism by constructing a multi-dimensional similarity evaluation model based on the initial spectral baseline, moisture reference wavelength reflectance, and environmental parameters. This method significantly improves the historical referenceability of state judgment and content assessment, and enhances the system's adaptability and assessment accuracy under different drying conditions.

[0159] Please see Figure 9 The present invention illustrates a near-infrared online detection method for volatile oil components during the drying process of medicinal materials according to an embodiment of the present invention. The method includes the following steps S901-S903, which will be described in detail below.

[0160] S901. Install the multiple monitoring points in the target drying area.

[0161] One possible implementation involves planning and deploying monitoring points within the target drying area. A strategy combining spatially uniform distribution with representative sampling is employed, establishing a comprehensive monitoring network by placing monitoring equipment at key locations within the drying area. Once the monitoring points are installed according to the predetermined plan, they provide a spatially distributed monitoring foundation for subsequent data collection.

[0162] Understandably, the uniformity of the spatial distribution of monitoring points can ensure the representativeness of the collected data for the entire drying area and avoid the existence of monitoring blind spots; while the reasonable selection of monitoring point locations needs to take into account environmental characteristics such as airflow organization and temperature distribution in the drying room to ensure that the monitoring data can truly reflect the actual situation of the drying process.

[0163] S902. Install spectral detection equipment, image acquisition equipment, and weighing equipment at multiple monitoring points.

[0164] One possible implementation involves integrating and installing multi-sensor devices at predetermined monitoring locations. Employing a strategy of collaborative deployment and functional complementarity, different types of data acquisition equipment are configured at each monitoring point to establish multi-dimensional monitoring capabilities. Once the three types of equipment are installed according to design requirements, simultaneous monitoring of the optical properties, morphological characteristics, and quality changes of the medicinal materials can be achieved.

[0165] S903: Collect infrared spectral data and medicinal material data based on a preset cycle.

[0166] One possible implementation involves using pre-installed multi-sensor devices to collect data at predetermined time intervals. A strategy combining synchronous triggering and periodic inspection is employed, establishing a continuous and consistent time-series dataset of the drying process by activating all sensors at fixed time intervals and acquiring monitoring data. Upon completion of a full collection cycle, multi-dimensional time-series information reflecting the dynamic changes in the state of the medicinal materials can be obtained.

[0167] The technical solution provided in the above embodiments can bring at least the following beneficial effects: This embodiment establishes a comprehensive and unified data acquisition system for the drying process through standardized monitoring point deployment and multi-sensor synchronous acquisition scheme. This method ensures the consistency and integrity of multi-source data in the spatiotemporal dimensions, guarantees the quality and reliability of monitoring data from the source, and provides a solid data foundation for subsequent analysis.

[0168] Please see Figure 10 The diagram illustrates a near-infrared online detection method for volatile oil components during the drying process of medicinal materials, according to an embodiment of the present invention. The method includes the following steps S1001, which will be described in detail below. S1001. Display or transmit the target monitoring time and the optimal drying index of each monitoring point through the output device.

[0169] One possible implementation involves using the target monitoring time and the optimal drying index for each monitoring point obtained from system analysis as output. A strategy combining visualization and data communication is employed, providing decision support information on the drying process to operators or higher-level systems through a human-machine interface or data interface. Once the data output is complete, it provides intuitive and quantitative evidence for determining the termination of drying operations and evaluating the process.

[0170] The technical solution provided in the above embodiments can bring at least the following beneficial effects: This embodiment provides operators with intuitive and quantitative decision support by outputting the target monitoring time and the optimal drying index of each monitoring point in real time. This method realizes transparent management of the drying process status, which facilitates timely intervention and process optimization, and at the same time provides complete data support for quality traceability and production improvement.

[0171] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0172] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A near-infrared online detection method for volatile oil components during the drying process of medicinal materials, characterized in that, The method includes: Abnormal wave data of the target monitoring point at multiple monitoring times are determined based on sensor time-series data; the sensor time-series data includes: infrared spectral data and medicinal material data; the abnormal wave data includes at least one of the following: the number of abnormal waves at the monitoring time, and the probability of drying abnormalities at the monitoring time; Based on the medicinal material data and the probability of drying anomalies, the first state change time of the target monitoring point is determined; the first state change time is the moment when the state of the volatile oil of the medicinal material changes abruptly during the drying process. Based on the abnormal wave data and the optimal drying index of the target monitoring point determined based on the first state change time, the optimal drying index is used to characterize the advantage of stopping drying at the target time. The target monitoring time is determined among the multiple monitoring times based on the optimal drying index; the target monitoring time is used to determine the drying time of the medicinal materials.

2. The near-infrared online detection method for volatile oil components during the drying process of medicinal materials according to claim 1, characterized in that, The target monitoring point is any one of multiple monitoring points; the step of determining the abnormal wave data of the target monitoring point at multiple monitoring times based on sensor time-series data includes: Based on the infrared spectral data, the temporal variation differences of reflectance at at least one spectral wavelength of the multiple monitoring points are determined; Based on the degree of deviation between the time-series variation difference of reflectance and the preset time-series variation difference of reflectance, the possibility of drying anomaly at the at least one spectral wavelength is determined. The number of spectral wavelengths that meet the preset conditions among the multiple monitoring points is determined as the number of abnormal waves; the preset conditions include at least one of the following: the result of the normalized processing of the drying abnormality probability is greater than a first preset threshold.

3. The near-infrared online detection method for volatile oil components during the drying process of medicinal materials according to claim 2, characterized in that, Based on the infrared spectral data, the temporal variation differences in reflectance at at least one spectral wavelength of the multiple monitoring points are determined, including: Determine the reflectance difference values ​​of the plurality of monitoring points at adjacent monitoring times of at least one spectral wavelength; A set of collaborative monitoring points is determined based on the reflectance difference value; the reflectance difference values ​​of the monitoring points in the collaborative monitoring point set have the same sign; The time-series variation difference of reflectance is determined based on the absolute value of the difference between the reflectance difference value of the target monitoring point and the reflectance difference value of each monitoring point in the set of collaborative monitoring points.

4. The near-infrared online detection method for volatile oil components during the drying process of medicinal materials according to claim 3, characterized in that, The infrared spectral data includes: multiple characteristic wavelengths of volatile oils and multiple moisture reference wavelengths; the medicinal material data includes at least one of the following: image data and weight data; the determination of the first state change moment of the target monitoring point based on the medicinal material data and the probability of drying anomalies includes: Based on the image data, the shrinkage rate of the medicinal material at the target monitoring point is determined; Based on the weight data, the weight change rate of the medicinal materials at the target monitoring point is determined; The spectral correlation change index is determined based on the reflectance changes of the characteristic wavelength of the volatile oil and the reference wavelength of the moisture. Based on the shrinkage change rate, the weight change rate, the spectral correlation change index, and the temporal change of the drying anomaly probability, the state change probability value is determined. The moment when the probability value of the state change is greater than the second preset threshold is determined as the first state change moment.

5. The near-infrared online detection method for volatile oil components during the drying process of medicinal materials according to claim 4, characterized in that, The determination of spectral correlation change indicators based on the reflectance changes of the characteristic wavelength of the volatile oil and the reference wavelength of the moisture includes: Determine the time series of reflectance changes at the characteristic wavelengths of the volatile oil and the time series of reflectance changes at the reference wavelengths of the moisture. Determine the correlation coefficient between the time series of reflectance changes at the characteristic wavelength of the volatile oil and the time series of reflectance changes at the reference wavelength of the moisture. The spectral correlation change index is determined based on the mean of multiple correlation coefficients.

6. The near-infrared online detection method for volatile oil components during the drying process of medicinal materials according to claim 5, characterized in that, The determination of the optimal drying index for the target monitoring point based on the abnormal wave data and the first state abrupt change time includes: Based on the duration from the first start of drying to the first state change time, and the duration from the second start of drying to the second state change time, a change time priority index is determined; the second state change time is the state change time of a historically similar drying process; the historically similar drying process is a historical drying process whose similarity index with the drying of medicinal materials at the target monitoring point is greater than a third preset threshold; the first start of drying time is the time when the target monitoring point begins drying; the second start of drying time is the time when the historically similar drying process begins drying; Determine the difference in volatile oil content between the target monitoring point and the point at the end of the historically similar drying process; The drying effectiveness index is determined based on the mutation time priority index, the number of abnormal waves, the probability of drying anomalies, and the difference in volatile oil content. At least one high-efficiency drying monitoring point is determined based on the drying effectiveness index of each of the multiple monitoring times; The optimal drying index is determined based on the drying effectiveness index of the high-efficiency drying monitoring point, the historical drying effectiveness index of the high-efficiency drying monitoring point, and the number of the high-efficiency drying monitoring points.

7. The near-infrared online detection method for volatile oil components during the drying process of medicinal materials according to claim 6, characterized in that, Determining the target monitoring time among the plurality of monitoring times based on the optimal drying index includes: The optimal drying index at the multiple monitoring times is normalized. Based on the result of the normalization of the optimal drying index at the multiple monitoring times, the first monitoring time with a normalization result greater than the fourth preset threshold is determined as the target monitoring time.

8. The near-infrared online detection method for volatile oil components during the drying process of medicinal materials according to claim 7, characterized in that, The method further includes: Acquire historical drying data; the historical drying data includes: the initial spectral baseline reflectance of the historical drying process, the initial reflectance of the moisture reference wavelength, and the time series curve of environmental parameters; The overall similarity between the historical drying process and the target drying process is determined based on the initial spectral baseline reflectance, the initial reflectance of the moisture reference wavelength, and the time series curves of environmental parameters. Identify historically similar drying processes within the historical drying process; the historically similar drying processes are those with a comprehensive similarity greater than a fifth preset threshold.

9. The near-infrared online detection method for volatile oil components during the drying process of medicinal materials according to claim 8, characterized in that, The method further includes: Install the aforementioned multiple monitoring points in the target drying area; Spectroscopic detection equipment, image acquisition equipment, and weighing equipment are installed at the multiple monitoring points; The infrared spectral data and the medicinal material data are collected at a preset period.

10. The near-infrared online detection method for volatile oil components during the drying process of medicinal materials according to claim 9, characterized in that, The method further includes: The target monitoring time and the optimal drying index of each monitoring point are displayed or transmitted through the output device.

Images