Rapid detection system and method for harmful components of traditional Chinese medicinal materials

Abstract

The invention discloses a rapid detection system and method for harmful components of traditional Chinese medicinal materials, and relates to the technical field of medicinal material analysis, full-process continuous monitoring is realized through a high-frame-rate micro spectrum probe array, and a three-dimensional component distribution map is generated in real time in combination with 5G-MEC edge calculation; an improved ant colony algorithm is adopted to construct a spectrum-process coupling matrix, a millisecond-level evaluation OPC-UA protocol is realized through a component migration index Qm, and a process adjustment parameter set delta T is ensured and fed back; a deep generative adversarial network is utilized to establish a noise feature library, wavelet packet decomposition is utilized to extract and purify spectral data, an interference compensation factor Rb is combined with a Bayesian probability model to dynamically correct detection deviation, a stochastic gradient descent algorithm is utilized to continuously optimize model parameters, the detection accuracy of different process stages reaches 95% or above, and a whole-process closed-loop quality control system is formed.

Description

Technical Field

[0001] The present invention relates to the technical field of medicinal material analysis, and in particular to a rapid detection system and method for harmful components of traditional Chinese medicine. Background Art

[0002] Traditional Chinese medicines (TCMs) are widely used in Traditional Chinese Medicine (TCM) due to their unique therapeutic benefits. However, due to the influence of cultivation, processing, and storage processes, they often contain harmful residues such as pesticides, heavy metals, and sulfur fumigants, posing a threat to drug safety and consumer health. Traditional detection methods such as gas chromatography and mass spectrometry, while highly sensitive, are complex, require long testing cycles, and require a high laboratory environment. To improve detection efficiency and accuracy, there is an urgent need to develop an integrated, intelligent system for detecting harmful components in TCMs.

[0003] However, when applied to the processing of Chinese herbal medicine slices, the following technical shortcomings often exist: 1. Most of the detection methods are random inspections, which makes it difficult to monitor the entire process; 2. Ingredient testing is disconnected from the production process, resulting in delayed data response; 3. Some complex processing techniques can easily interfere with detection accuracy. Summary of the Invention

[0004] In view of the deficiencies in the prior art, the present invention provides a rapid detection system and method for harmful components in traditional Chinese medicines, which solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a rapid detection system for harmful components of traditional Chinese medicine, including a dynamic sampling module, a multi-source data fusion module, a process interference compensation module, a risk quantification module and a real-time feedback module; The dynamic sampling module embeds a high-frame-rate micro-spectral probe array into key nodes of medicinal material processing equipment. It uses non-contact scanning to collect spectral characteristics of the surface and cross-section of medicinal materials. Combined with 5G-MEC edge computing, it generates a three-dimensional component distribution map in real time and establishes a dynamic sample library for the processing process. The multi-source data fusion module is used to integrate dynamic sampling data with production line temperature, humidity, and pressure parameters. It uses an improved ant colony algorithm to optimize the feature extraction path, construct a spectrum-process coupling matrix, and calculate the component migration index Qm to assess the diffusion trend of harmful components. The process interference compensation module uses a generative adversarial network to simulate the noise model under different processing technologies, removes the process interference frequency band through wavelet packet decomposition, generates a purified spectrum data set, and outputs the interference compensation factor Rb to correct the detection deviation; The risk quantification module establishes a Bayesian probability model based on the compensated data to dynamically calculate the probability Pc of exceeding the harmful component, triggers an early warning, and synchronously generates a set of process adjustment parameters, including the optimal correction amount of the crushing particle size and the drying temperature. The real-time feedback module directly connects with the production line PLC system through the OPC-UA protocol, pushes the risk level, component distribution thermodynamic diagram, and ΔT parameter to the central control terminal, and iteratively updates the compensation model based on historical data to realize self-optimization of detection accuracy.

[0006] Preferably, the dynamic sampling module includes a high-frame-rate miniature spectral probe unit and a dynamic sample library construction unit. The high-frame-rate miniature spectral probe unit is used to embed multiple high-frame-rate miniature spectral probe arrays into key detection node positions of the medicinal material processing equipment, adopt a non-contact scanning mode, capture multi-dimensional spectral features of the medicinal material surface and section at high speed, and automatically mark the spectral images at different detection positions; The dynamic sample library construction unit is used to combine the multi-dimensional spectral feature data of the medicinal material obtained by the high-frame-rate miniature spectral probe unit with the 5G-MEC edge computing technology, generate a three-dimensional component distribution map in the medicinal material processing process in real time, and store the detection data at different time periods and in different states in the dynamic sample library as reference data for subsequent component analysis and quality traceability.

[0007] Preferably, the source data fusion module includes a coupling matrix construction unit and a migration index calculation unit. The coupling matrix construction unit performs multi-source data fusion on the dynamic sampling data and the production line temperature, humidity, and pressure parameters, optimizes the feature extraction path through an improved ant colony algorithm, and constructs a multi-dimensional coupling matrix containing spectral features and process parameters; Preferably, the migration index calculation unit calculates the component migration index Qm of each processing node based on the coupling matrix through space-time correlation analysis, and the expression is as follows: In the formula, i represents the node index, Ci represents the component concentration of the i th node, and α and β are the space and time weight coefficients, respectively; By comparing and evaluating the migration risk threshold Qm0 and the component migration index Qm, it is determined whether there is an abnormal diffusion risk of harmful components, and the specific evaluation rules are as follows: If the component migration index Qm is greater than or equal to the migration risk threshold Qm0, it indicates that there is an abnormal diffusion risk of harmful components in the current processing link, at which time an early warning signal is sent to the risk quantification module, and a process parameter adjustment instruction is triggered; If the component migration index Qm is less than the migration risk threshold Qm0, it indicates that the harmful component diffusion in the current processing link is within the normal range, and no early warning and adjustment instruction is triggered.

[0008] Among them, the migration risk threshold Qm0 is set and adjusted according to the type of medicinal materials, processing technology and historical safety data.

[0009] Preferably, the process disturbance compensation module includes a noise modeling unit and a signal purification unit; The noise modeling unit uses a deep generative adversarial network (GAN) to establish a noise feature library under different processing technologies and generate a process interference feature spectrum; Preferably, the signal purification unit decomposes the original spectral data by wavelet packet transform, performs frequency band screening in combination with the interference characteristic spectrum, outputs the purified characteristic spectrum data, and calculates the interference compensation factor Rb: Where Sraw is the original spectrum intensity, Spure is the purified spectrum intensity; The reliability of the current detection data is determined by comparing the preset compensation validity threshold R with the interference compensation factor Rb. The specific evaluation rules are as follows: If the interference compensation factor Rb ≥ the compensation validity threshold R, it means that the current spectral data compensation effect meets the standard, the test result is valid, and the subsequent risk quantification process is allowed; If the interference compensation factor Rb is less than the compensation effectiveness threshold R, it means that the current spectral data is seriously affected by the process interference and the compensation effect is not up to standard. At this time, a re-collection instruction is sent to the dynamic sampling module, and the parameter weights of the interference compensation model are adjusted synchronously; Among them, the compensation effectiveness threshold R is optimized according to the type of medicinal materials, processing stage and historical detection data, and is set to 0.85 by default.

[0010] Preferably, the risk quantification module includes a probability model building unit and an early warning decision unit; The probability model building unit performs Bayesian network modeling on the compensated spectral feature data and dynamically calculates the probability Pc of harmful components exceeding the standard based on the type of medicinal material, processing stage and historical detection data; Preferably, the early warning decision unit is used to calculate the process adjustment parameter set ΔT by the following formula: Where K is the process sensitivity coefficient, Pc0 is the preset safety threshold, and Tstd is the standard process parameter; By comparing the preset risk level threshold Pc0 with the probability of harmful components exceeding the standard Pc, a graded early warning mechanism is triggered. The specific evaluation content is as follows: If the probability of harmful components exceeding the standard Pc ≥ the risk level threshold Pc0, it is determined that there is a safety hazard in the current processing link, and an advanced warning signal is sent to the real-time feedback module, and the process adjustment parameter set ΔT is output; If the probability of harmful components exceeding the standard Pc is less than the risk level threshold Pc0, it is determined that the current processing link is in a safe range, only the conventional data is recorded, and no early warning signal is triggered.

[0011] Preferably, the real-time feedback module establishes a real-time data channel with the production line PLC control system through the OPC-UA industrial communication protocol, synchronously transmits the risk level data, three-dimensional component distribution thermodynamic map and process adjustment parameter set ΔT generated by the risk quantification module to the central control terminal display interface, simultaneously constructs an iterative optimization model, adjusts the weight parameters of the interference compensation module based on historical detection data and actual production effect feedback, and uses the stochastic gradient descent algorithm to make the spectral detection model self-adaptive optimization, and finally, when receiving a high-level early warning signal, the current production batch is automatically locked, and a disposal scheme is pushed through the man-machine interaction interface, so that a closed-loop management is formed from risk detection to production control.

[0012] A rapid detection method for harmful components of traditional Chinese medicinal materials, comprising the following steps: Step one, embed a high-frame-rate micro-spectral probe array at key nodes of medicinal material processing equipment, use a non-contact scanning method to collect spectral characteristics of the surface and section of the medicinal material, generate a three-dimensional component distribution map in real time in combination with 5G-MEC edge computing, and establish a dynamic sample library of the processing process; Step two, integrate dynamic sampling data with production line temperature, humidity and pressure parameters, optimize the feature extraction path through an improved ant colony algorithm, construct a spectral-technology coupling matrix, and calculate a component migration index Qm to evaluate the diffusion trend of harmful components; Step three, simulate noise models under different processing technologies using a generative adversarial network, remove process interference frequency bands through wavelet packet decomposition, generate a purified spectral data set, and output an interference compensation factor Rb for correcting detection bias; Step four, based on the compensated data, establish a Bayesian probability model to dynamically calculate the probability of harmful components exceeding the standard Pc, trigger an early warning, and synchronously generate a process adjustment parameter set, including the optimal correction amount of the crushing granularity and drying temperature; Step five, directly connect with the production line PLC system through the OPC-UA protocol, push the risk level, component distribution thermodynamic map and ΔT parameter to the central control terminal, and iteratively update the compensation model based on historical data to realize self-optimization of detection accuracy.

[0013] The present application provides a rapid detection system and method for harmful components of traditional Chinese medicinal materials. (1) This rapid detection system and method for harmful components in traditional Chinese medicines addresses the problem that random sampling is the only detection method that makes it difficult to monitor the entire process. By embedding a high-frame-rate micro-spectral probe array into key processing nodes of medicinal materials and combining it with 5G-MEC edge computing to generate a three-dimensional component distribution map in real time, continuous monitoring of the entire process from raw materials to finished products is achieved. The dynamic sample library construction unit records processing data at a sampling frequency of 1-5 seconds, thereby improving monitoring coverage. (2) This rapid detection system and method for harmful components in traditional Chinese medicines addresses the problem of data response lag due to disconnection between component detection and production processes. An improved ant colony algorithm is used to construct a spectrum-process coupling matrix, and the diffusion trend of harmful components is evaluated in real time through the component migration index Qm. The OPC-UA protocol enables millisecond-level data interaction between the PLC system and the detection module, allowing the process adjustment parameter set ΔT to be fed back within 10 seconds. (3) This rapid detection system and method for harmful components in traditional Chinese medicines addresses the problem of complex processing techniques interfering with detection accuracy. A deep generative adversarial network (GAN) is used to establish a noise feature library, and spectral data is extracted and purified through wavelet packet decomposition. The interference compensation factor Rb is combined with a Bayesian probability model to dynamically correct the detection deviation, thereby increasing the detection accuracy at different process stages to more than 95%. The stochastic gradient descent algorithm updates the model parameters in each batch, forming a continuously optimized closed-loop system. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 This is a schematic diagram of the framework structure of a rapid detection system for harmful components in traditional Chinese medicines according to the present invention; Figure 2 The present invention is a schematic flow chart of the steps of a method for rapid detection of harmful components in traditional Chinese medicine. DETAILED DESCRIPTION

[0015] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0016] Example 1 See also Figure 1 , the present invention provides a rapid detection system for harmful components of traditional Chinese medicine, including a dynamic sampling module, a multi-source data fusion module, a process interference compensation module, a risk quantification module and a real-time feedback module; The dynamic sampling module embeds a high-frame-rate micro-spectral probe array at key nodes of the processing equipment for medicinal materials, uses a non-contact scanning method to collect spectral characteristics of the surface and cross-section of the medicinal materials, generates a three-dimensional component distribution map in real time through 5G-MEC edge computing, and establishes a dynamic sample library for the processing process. The multi-source data fusion module is used for integrating dynamic sampling data and production line temperature and humidity, pressure parameters, optimizing a feature extraction path through an improved ant colony algorithm, constructing a spectrum-process coupling matrix, and calculating a component migration index Qm to evaluate the diffusion trend of harmful components. The process interference compensation module uses a generative adversarial network to simulate noise models under different processing technologies, removes process interference frequency bands through wavelet packet decomposition, generates a purified spectral data set, and outputs an interference compensation factor Rb for correcting detection bias. The risk quantification module dynamically calculates the probability Pc of harmful components exceeding the standard based on the compensated data, triggers an early warning, and synchronously generates a process adjustment parameter set, including the optimal correction amount of the crushing particle size and drying temperature. The real-time feedback module directly connects with the production line PLC system through the OPC-UA protocol, pushes the risk level, component distribution thermodynamic map, and ΔT parameter to the central control terminal, and iteratively updates the compensation model based on historical data to realize self-optimization of detection accuracy.

[0017] In this embodiment, the dynamic sampling module realizes non-contact continuous collection of spectral characteristics of the surface and cross-section of medicinal materials through a high-frame-rate micro-spectral probe array (Qm0=0.75-0.95) (Ci=0.1-10mg / g), generates a three-dimensional component distribution map in real time through 5G-MEC edge computing (sampling frequency 1-5 seconds), and establishes a dynamic sample library to achieve 100% process monitoring coverage. The multi-source data fusion module uses an improved ant colony algorithm to construct a spectrum-process coupling matrix (α=0.6-0.8, β=0.2-0.4), and realizes millisecond-level evaluation (delay <50ms) of the diffusion trend of harmful components through the component migration index Qm (calculation period ≤0.5 seconds). The process interference compensation module uses a deep generative adversarial network (GAN) to establish a noise feature library, extracts purified spectral data through wavelet packet decomposition (Sraw / Spure≥0.9), and outputs the interference compensation factor Rb (R=0.85-0.95) to improve the detection accuracy to more than 95%. The risk quantification module dynamically calculates the probability of harmful components exceeding the standard based on the Bayesian probability model (Pc0=0.8-0.9), generates a process adjustment parameter set ΔT (K=1.2-2.0), and completes feedback within 10 seconds. The real-time feedback module realizes millisecond-level data interaction between the PLC system and the detection module through the OPC-UA protocol, continuously optimizes the model parameters through the stochastic gradient descent algorithm, and forms a whole-process closed-loop quality control system.

[0018] Example 2 The dynamic sampling module comprises a high-frame-rate micro-spectral probe unit and a dynamic sample library construction unit; The high-frame-rate micro-spectral probe unit is used for embedding a plurality of high-frame-rate micro-spectral probe arrays into key detection node positions of a medicinal material processing device, adopting a non-contact scanning mode, capturing multi-dimensional spectral features of a medicinal material surface and section at high speed, and automatically marking the spectral images of different detection positions; The dynamic sample library construction unit is used for combining the multi-dimensional spectral feature data of medicinal materials acquired by the high-frame-rate micro-spectral probe unit with 5G-MEC edge computing technology, generating a three-dimensional component distribution map in a medicinal material processing process in real time, and storing the detection data at different time periods and in different states in a dynamic sample library as reference data for subsequent component analysis and quality tracing.

[0019] The multi-source data fusion module comprises a coupling matrix construction unit and a migration index calculation unit; The coupling matrix construction unit performs multi-source data fusion on the dynamic sampling data and production line temperature, humidity and pressure parameters, optimizes a feature extraction path through an improved ant colony algorithm, and constructs a multi-dimensional coupling matrix containing spectral features and process parameters; The migration index calculation unit calculates a component migration index Qm of each processing node based on the coupling matrix through space-time correlation analysis, and the expression is as follows: In the formula, i represents a node index, Ci represents a component concentration of the i th node, and a and β represent space and time weight coefficients, respectively; By comparing and evaluating the component migration index Qm with a preset migration risk threshold Qm0, it is determined whether there is an abnormal diffusion risk of harmful components, and the specific evaluation rules are as follows: If the component migration index Qm is greater than or equal to the migration risk threshold Qm0, it indicates that there is an abnormal diffusion risk of harmful components in the current processing link, at this time, an early warning signal is sent to the risk quantification module, and a process parameter adjustment instruction is triggered; If the component migration index Qm is less than the migration risk threshold Qm0, it indicates that the harmful component diffusion in the current processing link is within a normal range, and no early warning and adjustment instruction is triggered.

[0020] The migration risk threshold Qm0 is set and adjusted according to medicinal material types, processing technology and historical safety data.

[0021] In the embodiment, the dynamic sampling module realizes non-contact high-speed acquisition of multi-dimensional spectral features (wavelength range 900-1700nm) of medicinal material surfaces and sections through a high-frame-rate micro-spectral probe array (frame rate 200-500fps), and a three-dimensional component distribution map (spatial resolution 0.1mm3 ), the dynamic sample library construction unit stores the whole process data in a time-stamped manner (sampling interval 1-5 seconds) to provide a complete data chain for quality tracing; the multi-source data fusion module includes a coupling matrix construction unit that optimizes the feature weight by an improved ant colony algorithm (iteration number 50-100 times) to establish a multi-dimensional coupling matrix containing spectral features (absorbance 0.1-1.5 AU) and process parameters (temperature 30-120℃ / pressure 0.1-0.5 MPa), and a migration index calculation unit that calculates the component migration index Qm (spatial weight α=0.6-0.8 / time weight β=0.2-0.4) based on the spatiotemporal correlation analysis to realize real-time early warning of harmful component diffusion risk by a preset migration risk threshold Qm0 (0.75-0.95), and automatically trigger process adjustment instructions (adjustment amplitude 5-15%) when Qm≥Qm0, which realizes dynamic correlation analysis of process parameters and component changes and improves the production process control accuracy by more than 40%; wherein the spectral probe spatial arrangement density (5-8 / cm 2 ) ensures that there is no dead angle in detection, the time weight coefficient β reflects the influence degree of different processing stages (cleaning / slicing / drying) on component migration, and the migration threshold Qm0 is set according to the differences of medicinal material types (root and stem / leaf / fruit), so that the whole system improves the harmful component detection rate to 99.5% while reducing the false positive rate to below 0.5%.

[0022] Example 3 The process interference compensation module includes a noise modeling unit and a signal purification unit. The noise modeling unit uses a deep generative adversarial network (GAN) to establish a noise feature library under different processing technologies to generate a process interference feature spectrum. The signal purification unit decomposes the original spectral data by wavelet packet transform, filters the frequency band in combination with the interference feature spectrum, outputs the purified feature spectral data, and calculates the interference compensation factor Rb: In the formula, Sraw is the original spectral intensity, and Spure is the purified spectral intensity. The preset compensation effectiveness threshold R is compared and evaluated with the interference compensation factor Rb to judge the reliability of the current detection data, and the specific evaluation rules are as follows: If the interference compensation factor Rb≥compensation effectiveness threshold R, it means that the compensation effect of the current spectral data meets the standard, the detection result is valid, and it is allowed to enter the subsequent risk quantification process; If the interference compensation factor Rb<compensation effectiveness threshold R, it means that the current spectral data is seriously affected by process interference, and the compensation effect does not meet the standard, at which time a reacquisition instruction is sent to the dynamic sampling module, and the parameter weight of the interference compensation model is adjusted synchronously. Among them, the compensation effectiveness threshold R is optimized according to the type of medicinal materials, processing stage and historical detection data, and is set to 0.85 by default.

[0023] The risk quantification module includes a probability model building unit and an early warning decision-making unit; The probability model building unit performs Bayesian network modeling on the compensated spectral feature data and dynamically calculates the probability Pc of harmful components exceeding the standard based on the type of medicinal material, processing stage and historical detection data; The early warning decision unit is used to calculate the process adjustment parameter set ΔT using the following formula: Where K is the process sensitivity coefficient, Pc0 is the preset safety threshold, and Tstd is the standard process parameter; By comparing the preset risk level threshold Pc0 with the probability of harmful components exceeding the standard Pc, a graded early warning mechanism is triggered. The specific evaluation content is as follows: If the probability of harmful components exceeding the standard Pc ≥ the risk level threshold Pc0, it is determined that there is a safety hazard in the current processing link, and an advanced warning signal is sent to the real-time feedback module, and the process adjustment parameter set ΔT is output; If the probability of harmful components exceeding the standard Pc is less than the risk level threshold Pc0, the current processing link is determined to be within a safe range, and only routine data recording is performed without triggering an early warning signal.

[0024] The real-time feedback module establishes a real-time data channel with the production line PLC control system through the OPC-UA industrial communication protocol, and synchronously transmits the risk level data, three-dimensional component distribution heat map and process adjustment parameter set ΔT generated by the risk quantification module to the central control terminal display interface; at the same time, it constructs an iterative optimization model based on historical detection data and actual production effect feedback, and uses the stochastic gradient descent algorithm to adjust the weight parameters of the interference compensation module to make the spectral detection model adaptively optimized; finally, when receiving an advanced warning signal, it automatically locks the current production batch and pushes the disposal plan through the human-computer interaction interface, so that a closed-loop management is formed from risk detection to production control.

[0025] In this embodiment, in view of the problem of insufficient accuracy of spectral detection data caused by complex process interference in the processing of traditional Chinese medicinal materials, a specific design of a process interference compensation module and a risk quantification module is proposed. The process interference compensation module includes a noise modeling unit and a signal purification unit. The noise modeling unit uses a deep generative adversarial network (generator network parameter layer number ≥ 5 layers, discriminator accuracy ≥ 90%) to establish a noise feature library (library capacity ≥ 10,000 noise modes) under different processing technologies (including cutting, crushing, drying, pressing, etc.), and automatically generates a process interference characteristic spectrum (spectral characteristic dimension ≥ 200). The signal purification unit uses wavelet packet transform (decomposition layer number ≥ 4, basic wavelet function db4) to perform multi-scale decomposition on the original spectral data (Sraw, spectral intensity range 0-1), combines the noise characteristic spectrum to select the interference frequency band, and outputs the purified characteristic spectral data (Spure, error rate ≤ 5%), and calculates the interference compensation factor Rb=Spure / Sraw. According to the preset compensation effectiveness threshold R=0.85, Rb is judged. If Rb≥R, it means that the detection data is valid and is allowed to enter the risk quantification process. If Rb<R, the dynamic sampling module is triggered to reacquire, and the parameter weight (learning rate range 0.0001-0.01) of the generative adversarial network is adjusted to enhance the model's adaptability to process interference. The risk quantification module proposes a joint design of a probability model construction unit and an early warning decision unit for potential risks after compensation. The probability model construction unit models the purified spectral feature data based on a Bayesian network (node number ≥ 50, edge number ≥ 100), fuses medicinal material types (traditional Chinese medicinal material type library ≥ 500), processing stages (subdivided into ≥ 10 processes), and historical detection data (data size ≥ 100 TB), and dynamically calculates the harmful component over-standard probability Pc (Pc value range 0-1). The early warning decision unit classifies the risk level according to Pc and the detection threshold T (default T=0.75). If Pc≥T, the early warning process is triggered, and the process adjustment parameter set is automatically generated, including the crushing granularity Δd (adjustment range 0.1mm-2mm), the drying temperature ΔT (adjustment range 2℃-20℃), and the processing time Δt (adjustment range 10s-600s). The real-time feedback module designs a real-time communication mechanism based on the OPC-UA protocol to meet the rapid response requirement of the detection result. The module is directly connected with the production line PLC system through the OPC-UA protocol (data refresh period ≤ 1 s, communication delay ≤ 50 ms), and synchronously pushes the risk level (divided into three levels of low, medium and high), the component distribution thermal map (image resolution ≥ 1080P) and the process adjustment parameters Δd, ΔT and Δt to the central control terminal. At the same time, the interference compensation model is iteratively updated (model update period ≤ 24 hours) according to the historical detection data and the latest detection result, so as to realize the dynamic self-optimization of the detection system and effectively guarantee the real-time, accuracy and stability of the detection of harmful components in traditional Chinese medicinal materials.

[0026] Embodiment 4 Please refer to Figure 2 A rapid detection method of harmful components in traditional Chinese medicinal materials, comprising the following steps: Step one, embed the high-frame-rate micro-spectrum probe array into the key nodes of the processing equipment of medicinal materials, adopt a non-contact scanning mode to collect the spectral characteristics of the surface and section of the medicinal materials, combine the 5G-MEC edge calculation to generate a three-dimensional component distribution map in real time, and establish a dynamic sample library of the processing process; Step two, integrate the dynamic sampling data with the temperature, humidity and pressure parameters of the production line, optimize the feature extraction path through the improved ant colony algorithm, construct a spectrum-process coupling matrix, and calculate the component migration index Qm to evaluate the diffusion trend of harmful components; Step three, simulate the noise model under different processing technologies by using the generative adversarial network, remove the process interference frequency band through wavelet packet decomposition, generate a purified spectrum data set, and output the interference compensation factor Rb for correcting the detection deviation; Step four, based on the compensated data, establish a Bayesian probability model to dynamically calculate the harmful component exceedance probability Pc, and trigger an early warning, and synchronously generate a set of process adjustment parameters, including the optimal correction amount of the crushing granularity and the drying temperature; Step five, directly connect with the production line PLC system through the OPC-UA protocol, push the risk level, component distribution thermal map and ΔT parameter to the central control terminal, and iteratively update the compensation model based on the historical data to realize the self-optimization of the detection accuracy.

[0027] Although embodiments of the present application have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A traditional Chinese medicine harmful ingredient rapid detection system, characterized in that: The dynamic sampling module, the multi-source data fusion module, the process interference compensation module, the risk quantification module and the real-time feedback module are comprised. ​ The dynamic sampling module embeds high-frame-rate micro-spectral probe arrays into key nodes of the processing equipment of medicinal materials, adopts a non-contact scanning mode to collect spectral characteristics of the surface and section of the medicinal materials, combines 5G-MEC edge computing to generate a three-dimensional component distribution map in real time, and establishes a dynamic sample library of the processing process. The multi-source data fusion module is used for integrating the dynamic sampling data and the temperature and humidity and pressure parameters of the production line, optimizing a feature extraction path through an improved ant colony algorithm, constructing a spectral-process coupling matrix, and calculating a component migration index Qm to evaluate the diffusion trend of harmful components. The process interference compensation module simulates noise models under different processing technologies through a generative adversarial network, removes process interference frequency bands through wavelet packet decomposition, generates a purified spectral data set, and outputs an interference compensation factor Rb for correcting detection bias. The risk quantification module establishes a Bayesian probability model based on the compensated data, dynamically calculates a harmful component exceeding probability Pc, triggers an early warning, and synchronously generates a process adjustment parameter set, including the optimal correction amount of the crushing granularity and the drying temperature. The real-time feedback module directly connects with the PLC system of the production line through an OPC-UA protocol, pushes the risk level, the component distribution thermodynamic map and the ΔT parameter to the central control terminal, and iteratively updates the compensation model based on historical data to realize self-optimization of detection accuracy.

2. The traditional Chinese medicine harmful ingredient rapid detection system according to claim 1, characterized in that: The dynamic sampling module comprises a high-frame-rate micro-spectral probe unit and a dynamic sample library construction unit. The high-frame-rate micro-spectral probe unit is used for embedding multiple high-frame-rate micro-spectral probe arrays into key detection node positions of the processing equipment of medicinal materials, adopting a non-contact scanning mode to capture multi-dimensional spectral characteristics of the surface and section of the medicinal materials at a high speed, and automatically labeling spectral images at different detection positions. The dynamic sample library construction unit is used for combining 5G-MEC edge computing technology to generate a three-dimensional component distribution map in the processing process of medicinal materials in real time based on multi-dimensional spectral characteristic data of medicinal materials acquired by the high-frame-rate micro-spectral probe unit, and storing detection data at different time periods and in different states in the dynamic sample library as reference data for subsequent component analysis and quality tracing.

3. The traditional Chinese medicine harmful ingredient rapid detection system according to claim 2, characterized in that: The multi-source data fusion module comprises a coupling matrix construction unit and a migration index calculation unit. The coupling matrix construction unit performs multi-source data fusion on the dynamic sampling data and the temperature and humidity and pressure parameters of the production line, optimizes a feature extraction path through an improved ant colony algorithm, and constructs a multi-dimensional coupling matrix containing spectral characteristics and process parameters.

4. The traditional Chinese medicine material harmful ingredient rapid detection system according to claim 3, characterized in that: The migration index calculation unit calculates a component migration index Qm of each processing node based on the coupling matrix through spatio-temporal correlation analysis, and the expression is as follows: In the formula, i represents a node index, Ci represents a component concentration of the i th node, and α and β are spatial and temporal weight coefficients, respectively. By comparing and evaluating the preset migration risk threshold Qm0 and the component migration index Qm, it is determined whether there is an abnormal diffusion risk of harmful components, and the specific evaluation rules are as follows: If the component migration index Qm is greater than or equal to the migration risk threshold Qm0, it indicates that there is an abnormal diffusion risk of harmful components in the current processing link, at this time, a warning signal is sent to the risk quantification module, and a process parameter adjustment instruction is triggered; If the component migration index Qm is less than the migration risk threshold Qm0, it indicates that the harmful component diffusion of the current processing link is within a normal range, and no warning and adjustment instructions are triggered. The migration risk threshold Qm0 is set and adjusted according to the type of medicinal materials, processing technology and historical safety data.

5. The traditional Chinese medicine harmful ingredient rapid detection system according to claim 4, characterized in that: The process interference compensation module includes a noise modeling unit and a signal purification unit. The noise modeling unit uses a deep generative adversarial network (GAN) to establish a noise feature library under different processing technologies, and generates a process interference feature spectrum.

6. The traditional Chinese medicine harmful ingredient rapid detection system according to claim 5, characterized in that: The signal purification unit decomposes the original spectral data through wavelet packet transform, and performs frequency band screening in combination with the interference feature spectrum to output purified feature spectral data, and calculates the interference compensation factor Rb: In the formula, Sraw is the original spectral intensity, and Spure is the purified spectral intensity. By comparing the interference compensation factor Rb with the preset compensation effectiveness threshold R, the reliability of the current detection data is evaluated, and the specific evaluation rules are as follows: If the interference compensation factor Rb is greater than or equal to the compensation effectiveness threshold R, it indicates that the compensation effect of the current spectral data meets the standard, the detection result is valid, and it is allowed to enter the subsequent risk quantification process; If the interference compensation factor Rb is less than the compensation effectiveness threshold R, it indicates that the current spectral data is seriously affected by process interference, and the compensation effect does not meet the standard, at this time, a reacquisition instruction is sent to the dynamic sampling module, and the parameter weight of the interference compensation model is adjusted synchronously. The compensation effectiveness threshold R is optimized according to the type of medicinal materials, processing stage and historical detection data, and is set to 0.85 by default.

7. The traditional Chinese medicine harmful ingredient rapid detection system according to claim 6, characterized in that: The risk quantification module includes a probability model construction unit and a warning decision unit. The probability model construction unit models the compensated spectral feature data through a Bayesian network, and dynamically calculates the harmful component over-limit probability Pc according to the type of medicinal materials, processing stage and historical detection data.

8. The traditional Chinese medicine material harmful ingredient rapid detection system according to claim 7, characterized in that: The warning decision unit is used to calculate the process adjustment parameter set AT through the following formula: In the formula, K is the process sensitivity coefficient, Pc0 is the preset safety threshold, and Tstd is the standard process parameter. By comparing the harmful component over-limit probability Pc with the preset risk level threshold Pc0, a hierarchical warning mechanism is triggered, and the specific evaluation content is as follows: If the harmful component over-limit probability Pc is greater than or equal to the risk level threshold Pc0, it is determined that there is a safety hazard in the current processing link, a high-level warning signal is sent to the real-time feedback module, and the process adjustment parameter set AT is outputted; If the harmful component over-limit probability Pc is less than the risk level threshold Pc0, it is determined that the current processing link is within a safe range, and only normal data recording is performed, without triggering a warning signal.

9. The traditional Chinese medicine material harmful ingredient rapid detection system according to claim 8, characterized in that: The real-time feedback module establishes a real-time data channel with the production line PLC control system through the OPC-UA industrial communication protocol, synchronously transmits the risk level data, three-dimensional component distribution thermodynamic diagram and process adjustment parameter set AT generated by the risk quantification module to the central control terminal display interface, At the same time, an iterative optimization model is constructed, based on historical detection data and actual production effect feedback, and the random gradient descent algorithm is used to adjust the weight parameters of the interference compensation module, so that the spectral detection model is self-adaptive and optimized; finally, when receiving the high-level early warning signal, the current production batch is automatically locked, and the disposal scheme is pushed through the man-machine interface, so that the risk detection and production control form a closed-loop management.

10. A method for rapid detection of harmful components in Chinese medicinal materials, according to any one of the rapid detection systems for harmful components in Chinese medicinal materials in claims 1-9, characterized in that: The method comprises the following steps: Step one, embed the key nodes of processing equipment of medicinal materials through high frame rate micro spectral probe array, collect the spectral characteristics of the surface and section of medicinal materials by using non-contact scanning method, generate three-dimensional component distribution map in real time combined with 5G-MEC edge calculation, and establish dynamic sample library in processing; Step two, integrate dynamic sampling data and production line temperature and humidity, pressure parameters, optimize feature extraction path through improved ant colony algorithm, construct spectral-technology coupling matrix, and calculate component migration index Qm to evaluate the diffusion trend of harmful components; Step three, simulate noise models under different processing technologies by using generative adversarial network, remove process interference frequency band by wavelet packet decomposition, generate purified spectral data set, and output interference compensation factor Rb for correcting detection deviation; Step four, based on the compensated data, a Bayesian probability model is established to dynamically calculate the probability Pc of harmful components exceeding the standard, and trigger an early warning, and generate a set of process adjustment parameters, including the optimal correction amount of crushing particle size and drying temperature; Step five, through OPC-UA protocol, the risk level, component distribution thermodynamic diagram and ΔT parameter are pushed to the central control terminal by directly connecting with the production line PLC system, and the compensation model is iteratively updated based on historical data, so as to realize the self-optimization of detection accuracy.

Images