Traditional Chinese medicine decoction piece cleaning method based on artificial intelligence

Abstract

The invention discloses a traditional Chinese medicine decoction piece cleaning method based on artificial intelligence, and the method comprises the following steps: S1, carrying out the image collection of the surface of a traditional Chinese medicine decoction piece through a hyperspectral imaging system, carrying out the processing of the image through a difference feature separation image enhancement algorithm, and obtaining an enhanced image which highlights the pollutant features of the surface of the decoction piece; s2, based on the enhanced image, classifying and identifying pollutants on the surfaces of the decoction pieces by utilizing a self-adaptive classification algorithm of topological feature mapping, and generating a two-dimensional distribution diagram of the pollutants; S3, according to the generated pollutant distribution diagram, combining physical characteristics of the decoction pieces, and adopting a dynamic decision tree self-adaptive optimization algorithm to obtain a two-dimensional distribution diagram of the pollutants; the water flow intensity, the cleaning time and the cleaning medium concentration parameters are dynamically adjusted according to different pollutant types and distribution conditions, so that pollutants are effectively removed and decoction pieces are protected in the cleaning process. According to the invention, image recognition and cleaning are carried out on the traditional Chinese medicine beverage through an artificial intelligence image recognition technology, so that the system can better adapt to cleaning of different pollutants of different beverages.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of traditional Chinese medicine production and processing, and particularly relates to a traditional Chinese medicine decoction piece cleaning method based on artificial intelligence. BACKGROUND

[0002] In the preparation process of traditional Chinese medicine decoction pieces, cleaning is a crucial step that directly affects the safety and quality of the traditional Chinese medicine decoction pieces. In the traditional cleaning process of traditional Chinese medicine decoction pieces, manual cleaning or semi-automatic mechanical cleaning methods are usually used. Although these methods can meet the needs in some simple scenarios, they have many problems in cleaning effect, efficiency, and automation and intelligentization of the cleaning process when facing complex and diversified pollutants.

[0003] In the prior art, the cleaning of traditional Chinese medicine decoction pieces mainly relies on traditional manual operation or semi-automatic mechanical equipment. In the manual cleaning process, workers observe and clean manually by naked eye, and remove pollutants such as dirt, impurities and insects on the surface of traditional Chinese medicine decoction pieces according to personal experience. This method is not only low in efficiency, but also greatly depends on the experience level of workers, and is prone to incomplete cleaning or over-cleaning. Although the semi-automatic mechanical cleaning equipment can improve the cleaning efficiency to some extent, it still has many problems. The mechanical equipment usually uses fixed cleaning parameters, and it is difficult to make dynamic adjustments according to the types, distribution of pollutants on the surface of traditional Chinese medicine decoction pieces and the physical properties of decoction pieces, which makes the existing technology not capable of dealing with complex and diversified pollutants and different types of decoction pieces.

[0004] The main shortcomings of the prior art are as follows: first, the identification of pollutants is not accurate enough. The traditional cleaning method relies on manual or simple mechanical screening, which cannot effectively distinguish different types of pollutants. The differences in color, shape and size of pollutants are large, which brings challenges to the cleaning process. The mechanical equipment often cannot accurately identify and classify the pollutants on the surface of decoction pieces, resulting in incomplete removal of some pollutants. Second, the setting of cleaning parameters lacks flexibility. The mechanical equipment usually presets fixed cleaning time and water flow intensity, and cannot make individualized adjustments according to specific pollutants and decoction piece characteristics. This one-size-fits-all cleaning method is prone to damage the decoction pieces due to over-cleaning, or fail to completely remove the pollutants.

[0005] Existing technologies rely on workers' experience for manual operation, but experience cannot guarantee consistency and high precision. When faced with a large number of complex Chinese herbal medicine pieces, errors are prone to occur. Although the fixed parameters of mechanical equipment improve efficiency, they lack real-time feedback and intelligent adjustment mechanisms, making it difficult to adapt to situations with uneven distribution of contaminants or diverse types of herbal medicine pieces. Traditional image processing and detection technologies have low precision and cannot accurately detect residual contaminants on the surface of herbal medicine pieces before and after cleaning, making it difficult to assess and optimize the cleaning quality. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method for cleaning Chinese herbal medicine slices based on artificial intelligence. This method uses image recognition technology based on artificial intelligence to identify and clean Chinese herbal medicine slices, thereby better adapting to the cleaning of different beverages and different contaminants.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is: a method for cleaning traditional Chinese medicine decoction pieces based on artificial intelligence, comprising the following steps:

[0008] S1. Images of the surface of Chinese herbal medicine slices are acquired using a hyperspectral imaging system, and the images are processed using an image enhancement algorithm that separates differential features to obtain an enhanced image that highlights the contaminant features on the surface of the slices.

[0009] S2. Based on the enhanced image, an adaptive classification algorithm using topological feature mapping is used to classify and identify contaminants on the surface of medicinal slices, generating a two-dimensional distribution map of the contaminants.

[0010] S3. Based on the generated pollutant distribution map and combined with the physical characteristics of the medicinal slices, a dynamic decision tree adaptive optimization algorithm is adopted to dynamically adjust the water flow intensity, cleaning time and cleaning medium concentration parameters according to different pollutant types and distribution, so as to effectively remove pollutants and protect the medicinal slices during the cleaning process.

[0011] The cleaning method further includes step S4, which involves applying a spatiotemporal feedback intelligent control algorithm during the cleaning process to monitor the cleaning progress and pollutant removal in real time, and dynamically adjusting the cleaning parameters based on the feedback results until the cleaning is completed or ended.

[0012] The cleaning method further includes step S5, which involves detecting and evaluating residual foreign matter on the medicinal slices through a multi-dimensional data fusion intelligent detection algorithm after cleaning, and automatically generating a cleaning quality report.

[0013] In step S1, a hyperspectral imaging system is used to acquire multispectral images of the surface of the Chinese herbal medicine slices. The hyperspectral imaging system acquires multidimensional image data of the surface of the Chinese herbal medicine slices in multiple different wavelength ranges, generating an image matrix containing different spectral information. ;in, Indicates wavelength. and Represents the two-dimensional spatial coordinates of the image. Indicates at wavelength Location, coordinates Image intensity value;

[0014] The acquired image data is preprocessed using a filtering function. Eliminate noise and interference signals to generate a denoised image matrix. :

[0015] ;

[0016] in, Indicates at wavelength The following noise filtering function for the image;

[0017] The image enhancement algorithm using differential feature separation in step S1 processes the image, including: processing the denoised image matrix. Spectral weighting is performed, and based on the spectral differences between pollutants and background at different wavelengths, a weighting function is used to... Enhance pollutant characteristics to generate an enhanced image matrix :

[0018] ;

[0019] in, These are the weighting coefficients for different wavelengths; based on the enhanced image matrix Generate the final enhanced image used for identification and cleaning.

[0020] Step S2 includes: generating an enhanced image matrix. The image data is processed using a topological feature mapping algorithm. This algorithm analyzes the changes in the topological structure of image pixels at different wavelengths to identify and separate contaminant regions on the surface of the medicinal slices, generating a preliminary contaminant marker matrix, denoted as... ,in Indicates coordinates The identified contaminant area;

[0021] Based on the preliminary pollutant labeling matrix An adaptive classification algorithm is applied to further classify pollutant regions. This algorithm performs multi-level classification based on the spectral characteristics, geometric shape, and edge features of pollutants, generating a high-precision pollutant classification matrix. ,in Indicates coordinates The corresponding pollutant category label;

[0022] The generated pollutant classification matrix Optimization was performed by iteratively updating the model to improve classification accuracy. A probabilistic model of pollutant features was constructed by combining Bayesian inference with the spatial distribution patterns of pollutant features. The pollutant category identification results are optimized, and a pollutant classification matrix is ​​finally generated.

[0023] The optimized pollutant classification matrix By fusing with hyperspectral information, a final two-dimensional distribution map of pollutants is generated. The two-dimensional distribution map is based on the spectral weighting function. Perform weighted processing:

[0024] ;

[0025] in, This is a two-dimensional distribution map of pollutants. This is a weighted sum for each spectral channel.

[0026] Step S3 includes using the generated pollutant distribution map. Combining the physical properties of Chinese herbal medicine slices, a dynamic decision tree adaptive optimization algorithm is used to construct a multivariate optimization model and automatically generate a cleaning plan. The optimization model comprehensively considers the type, distribution, thickness of contaminants on the surface of the slices and the material properties of the slices.

[0027] During the cleaning scheme generation process, the dynamic decision tree adaptive optimization algorithm adjusts the water flow intensity in real time through iterative calculations. The water flow intensity is dynamically adjusted based on the distribution density and thickness of pollutants, as well as the pollutant's response to the water flow.

[0028] ;

[0029] in, Indicates the pollutants in the coordinates Distribution density at that location For the thickness of the pollutants, and The weighting coefficient controls the sensitivity of water flow intensity to the characteristics of different pollutants;

[0030] Dynamically adjust cleaning time The cleaning time is based on the distribution density and thickness of contaminants and the water absorption of the medicinal slices, ensuring that the cleaning time can thoroughly remove contaminants while avoiding over-cleaning of the medicinal slices.

[0031] ;

[0032] in, For the medicinal slices in coordinates The water absorption coefficient at that location This is the cleaning rate coefficient;

[0033] The concentration of the cleaning medium is dynamically adjusted according to different types of pollutants. The concentration is optimized in real time based on the chemical properties of the pollutants and the changes in their interaction forces with the surface of the medicinal slices.

[0034] ;

[0035] in, The chemical properties of pollutants are represented on the coordinate system. The performance at the place, This represents the functional relationship between pollutant distribution density and chemical properties. This is the concentration adjustment factor;

[0036] The dynamic decision tree adaptive optimization algorithm monitors the water flow intensity, cleaning time, and medium concentration in real time during the cleaning process, so that various parameters are adjusted according to real-time data to ensure effective removal of pollutants and structural protection of Chinese herbal medicine pieces.

[0037] Step S4 includes:

[0038] During the cleaning process, several key cleaning parameters are monitored in real time, including water flow intensity. Cleaning time Pollutant removal rate and the concentration of cleaning medium It collects data through real-time sensors and performs data analysis and processing;

[0039] Adjust the water flow intensity based on the monitored cleaning progress and pollutant removal status. This ensures that the water flow intensity matches the distribution density of pollutants and the removal requirements.

[0040] ;

[0041] in, The adjusted water flow intensity, This is an incremental adjustment based on real-time feedback data;

[0042] Dynamically adjust cleaning time Combining real-time monitoring of pollutant removal rate and surface condition of medicinal slices:

[0043] ;

[0044] in, The adjusted cleaning time, This is the amount of time adjustment based on feedback from the removal rate and the condition of the medicinal slices;

[0045] Based on pollutant removal rate Based on the chemical properties of pollutants, dynamically optimize the concentration of the cleaning medium. :

[0046] ;

[0047] in, The adjusted concentration of the cleaning medium. The adjustment amount is based on the feedback of the chemical properties of the pollutants. This is the concentration adjustment factor;

[0048] By continuously monitoring the changes in various parameters during the cleaning process, and dynamically adjusting the water flow intensity, cleaning time, and medium concentration based on the feedback results, the cleaning efficiency is continuously optimized while protecting the Chinese herbal medicine pieces.

[0049] Step S5 includes performing multi-dimensional image acquisition and fusion processing on the surface of the medicinal slices after cleaning. The image acquisition is performed through multiple different angles and spectral channels to generate an image matrix of residual foreign matter containing comprehensive details. ,in, Indicates different spectral channels, and The spatial coordinates of the image;

[0050] The residual foreign object image matrix Image data matrix of contaminants before cleaning Compare the results and calculate the foreign matter removal rate matrix. :

[0051] ;

[0052] in, Indicates coordinates The closer the foreign matter removal rate is to the value, the better. The better the cleaning effect;

[0053] Based on foreign matter removal rate matrix The remaining foreign matter is classified and its distribution is analyzed to identify areas that are not thoroughly cleaned, and a distribution matrix of the remaining foreign matter is generated. ;

[0054] Based on the residual foreign matter distribution matrix and removal rate It automatically generates a cleaning quality report, which includes the types and locations of residual foreign matter, removal efficiency, and an overall assessment of the cleaning effect. :

[0055] ;

[0056] in, Indicates coordinates The quality of cleaning at the site.

[0057] Step S5 also includes automatically adjusting the relevant parameters for the next round of cleaning based on the generated cleaning quality report, performing secondary cleaning on areas that were not thoroughly cleaned, completely removing foreign matter, and storing the detection data and cleaning results in the database.

[0058] The advantages of this invention are:

[0059] (1) This invention uses hyperspectral imaging technology combined with image enhancement algorithms that separate differential features to accurately acquire multidimensional image data of the surface of Chinese herbal medicine slices. With the help of this data, the system can monitor the surface of the slices in all directions within different spectral ranges, ensuring that the characteristics of various pollutants can be effectively captured. Compared with the traditional method that relies on manual observation or mechanical cleaning, the limitation of existing technologies in accurately identifying different types of pollutants is solved, greatly improving the identification accuracy and cleaning efficiency.

[0060] (2) Based on hyperspectral image data, the topological feature mapping algorithm can further analyze the pollutant area on the surface of the medicinal slices, and combine it with the adaptive classification algorithm to accurately classify and identify the pollutants, generating a two-dimensional distribution map of the pollutants. The classification process greatly improves the accuracy of pollutant identification. It can not only effectively distinguish different types of pollutants such as mud, impurities, and insects, but also provide accurate input information for the generation of cleaning schemes. This invention overcomes the shortcomings of inaccurate pollutant identification and reliance on human experience in traditional cleaning methods, and realizes intelligent pollutant classification and treatment.

[0061] (3) The present invention adopts a dynamic decision tree adaptive optimization algorithm. Combined with the generated pollutant distribution map and the physical characteristics of Chinese herbal medicine pieces, it can automatically generate the optimal cleaning scheme. This scheme can dynamically adjust the cleaning parameters according to the type and distribution of pollutants, including water flow intensity, cleaning time and concentration of cleaning medium, to ensure that the cleaning process can be flexibly responded to according to the actual situation. This cleaning scheme solves the problem of fixed and inflexible cleaning parameters in the prior art, avoids over-cleaning or under-cleaning, and not only effectively removes pollutants, but also protects the structure and quality of Chinese herbal medicine pieces to the maximum extent.

[0062] (4) The present invention applies a spatiotemporal feedback intelligent control algorithm to monitor the cleaning progress and pollutant removal in real time, and dynamically adjusts the cleaning parameters according to the feedback results. This intelligent feedback mechanism greatly improves the automation and intelligence level of the cleaning process, solves the limitation of traditional cleaning equipment that is difficult to adjust cleaning parameters in real time. Through precise control of water flow intensity, cleaning time and cleaning medium concentration, the cleaning effect can be continuously optimized during the cleaning process, ensuring that each piece of Chinese medicine decoction piece receives the best cleaning treatment, and avoiding damage to the decoction piece due to excessive cleaning time or excessive water flow intensity.

[0063] (5) This invention uses a multi-dimensional data fusion intelligent detection algorithm to comprehensively evaluate the cleaning effect. By comparing the image data before and after cleaning, the system can accurately calculate the foreign matter removal rate, identify areas that are not thoroughly cleaned, and generate a detailed cleaning quality report. This achieves automated and intelligent evaluation of cleaning quality, overcoming the low precision problem caused by manual detection in traditional cleaning methods. Based on the cleaning quality report, the system can automatically adjust the relevant parameters for the next round of cleaning and perform secondary cleaning on residual pollutants to ensure the thoroughness and consistency of the cleaning effect. Through the feedback closed-loop mechanism, the cleaning process is continuously optimized. It can not only monitor and adjust the cleaning parameters in real time, but also gradually improve the cleaning efficiency and accuracy through learning and analysis of historical data. The data of each cleaning process will be stored and used to optimize the parameter settings for the next round of cleaning, thereby forming an adaptive intelligent cleaning process. Attached Figure Description

[0064] The following is a brief explanation of the contents of each of the accompanying drawings and the markings in the drawings:

[0065] Fig. 1 This is an overall flowchart of a method for cleaning traditional Chinese medicine decoction pieces based on artificial intelligence proposed in this invention;

[0066] Fig. 2 This is a flowchart of the identification and classification of contaminants in traditional Chinese medicine decoction pieces based on hyperspectral imaging and topological feature mapping algorithm in this invention;

[0067] Fig. 3 This is a flowchart illustrating the generation and adjustment of the cleaning scheme based on the dynamic decision tree adaptive optimization algorithm in this invention. Detailed Implementation

[0068] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and the description of the preferred embodiments.

[0069] This embodiment uses image recognition combined with artificial intelligence to identify contaminants in traditional Chinese medicine beverages and intelligently set their cleaning parameters, thereby achieving accurate, fast and automatic cleaning. It can also adjust the cleaning parameters.

[0070] like Figs. 1-3 As shown in the figure, this embodiment provides a method for cleaning traditional Chinese medicine decoction pieces based on artificial intelligence, which includes the following steps:

[0071] S1. Images of the surface of Chinese herbal medicine slices are acquired using a hyperspectral imaging system, and the images are processed using an image enhancement algorithm that separates differential features to obtain an enhanced image that highlights the contaminant features on the surface of the slices.

[0072] S2. Based on the enhanced image, an adaptive classification algorithm using topological feature mapping is used to classify and identify contaminants on the surface of medicinal slices, generating a two-dimensional distribution map of the contaminants.

[0073] S3. Based on the generated pollutant distribution map and combined with the physical characteristics of the medicinal slices, a dynamic decision tree adaptive optimization algorithm is adopted to dynamically adjust the water flow intensity, cleaning time and cleaning medium concentration parameters according to different pollutant types and distribution, so as to effectively remove pollutants and protect the medicinal slices during the cleaning process.

[0074] Step S4 includes applying a spatiotemporal feedback intelligent control algorithm during the cleaning process to monitor the cleaning progress and contaminant removal in real time, and dynamically adjusting the cleaning parameters based on the feedback results until the cleaning is completed or ended.

[0075] Step S5 includes detecting and evaluating residual foreign matter on the medicinal slices after cleaning by using a multi-dimensional data fusion intelligent detection algorithm, and automatically generating a cleaning quality report.

[0076] Based on the above steps, this application can achieve the identification and adaptive cleaning of contaminants in traditional Chinese medicine beverages. By adjusting the cleaning parameters according to the identified contaminants, intelligent and reliable cleaning is achieved. The specific solution is as follows:

[0077] like Figs. 1-3 As shown, an artificial intelligence-based method for cleaning traditional Chinese medicine decoction pieces includes the following steps:

[0078] Step 1: Acquire images of the surface of Chinese herbal medicine slices using a hyperspectral imaging system, and process the images using an image enhancement algorithm that separates differential features to highlight the contaminant characteristics on the surface of the slices;

[0079] Step 2: Based on the acquired hyperspectral images, an adaptive classification algorithm using topological feature mapping is used to classify and identify contaminants on the surface of the medicinal slices. The contaminants include different types of mud, impurities, and insects, generating a two-dimensional distribution map of the contaminants.

[0080] Step 3: Based on the generated contaminant distribution map and the physical characteristics of the medicinal slices, a dynamic decision tree adaptive optimization algorithm is used to automatically generate a cleaning plan. The water flow intensity, cleaning time, and cleaning medium concentration parameters are dynamically adjusted according to different contaminant types and distribution to ensure effective removal of contaminants and protection of the medicinal slices during the cleaning process.

[0081] Step 4: During the cleaning process, a spatiotemporal feedback intelligent control algorithm is applied to monitor the cleaning progress and contaminant removal in real time, and to dynamically adjust the cleaning parameters based on the feedback results, so as to continuously optimize the cleaning process and avoid over- or under-cleaning.

[0082] Step 5: After cleaning, the residual foreign matter in the medicinal slices is detected and evaluated using a multi-dimensional data fusion intelligent detection algorithm, and a cleaning quality report is automatically generated.

[0083] Step 1 includes the following steps:

[0084] 1. Multispectral images of the surface of Chinese herbal medicine slices are acquired using a hyperspectral imaging system. This system can obtain multidimensional image data of the surface of Chinese herbal medicine slices within multiple different wavelength ranges, generating an image matrix containing different spectral information.

[0085] ;

[0086] in, Indicates wavelength. and Represents the two-dimensional spatial coordinates of the image. Indicates at wavelength Location, coordinates Image intensity value;

[0087] 2. Preprocess the acquired image data using a filtering function. Eliminate noise and interference signals to generate a denoised image matrix. :

[0088] ;

[0089] in, Indicates at wavelength The following noise filtering function for the image;

[0090] 3. Apply the image enhancement algorithm of differential feature separation to the denoised image matrix. The algorithm performs spectral weighting processing, which calculates the spectral differences between pollutants and the background at different wavelengths using a weighting function. Enhance pollutant characteristics to generate an enhanced image matrix :

[0091] ;

[0092] in, These are weighting coefficients for different wavelengths;

[0093] 4. Based on the processed image matrix This generates an enhanced image that is ultimately used for identification and cleaning.

[0094] Step 2 includes the following steps:

[0095] 1. Based on the generation of enhanced image matrix The image data is processed using a topological feature mapping algorithm. This algorithm analyzes the changes in the topological structure of image pixels at different wavelengths to identify and separate contaminant regions on the surface of the medicinal slices, generating a preliminary contaminant marker matrix, denoted as... ,in Indicates coordinates The identified contaminant area;

[0096] 2. Based on the preliminary pollutant labeling matrix An adaptive classification algorithm is applied to further classify pollutant regions. This algorithm performs multi-level classification based on the spectral characteristics, geometric shape, and edge features of pollutants, generating a high-precision pollutant classification matrix. ,in Indicates coordinates The corresponding pollutant category label;

[0097] 3. The generated pollutant classification matrix Optimization was performed by iteratively updating the model to improve classification accuracy. A probabilistic model of pollutant features was constructed by combining Bayesian inference with the spatial distribution patterns of pollutant features. The pollutant category identification results are optimized, and a pollutant classification matrix is ​​finally generated.

[0098] 4. Optimize the pollutant classification matrix By fusing with hyperspectral information, a final two-dimensional distribution map of pollutants is generated. The two-dimensional distribution map is based on the spectral weighting function. Perform weighted processing:

[0099] ;

[0100] in, This is a two-dimensional distribution map of pollutants. This is a weighted sum for each spectral channel.

[0101] Step 3 includes:

[0102] 1. Based on the generated pollutant distribution map Combining the physical properties of Chinese herbal medicine slices, a dynamic decision tree adaptive optimization algorithm is used to construct a multivariate optimization model and automatically generate a cleaning plan. The optimization model comprehensively considers the type, distribution, thickness of contaminants on the surface of the slices and the material properties of the slices.

[0103] 2. During the cleaning plan generation process, the dynamic decision tree adaptive optimization algorithm adjusts the water flow intensity in real time through iterative calculation. The water flow intensity is dynamically adjusted based on the distribution density and thickness of pollutants, as well as the pollutant's response to the water flow.

[0104] ;

[0105] in, Indicates the pollutants in the coordinates Distribution density at that location For the thickness of the pollutants, and The weighting coefficient controls the sensitivity of water flow intensity to the characteristics of different pollutants;

[0106] 3. Dynamically adjust cleaning time The cleaning time is based on the distribution density and thickness of contaminants and the water absorption of the medicinal slices, ensuring that the cleaning time can thoroughly remove contaminants while avoiding over-cleaning of the medicinal slices.

[0107] ;

[0108] in, For the medicinal slices in coordinates The water absorption coefficient at that location This is the cleaning rate coefficient;

[0109] 4. Dynamically adjust the concentration of the cleaning medium according to different types of contaminants. The concentration is optimized in real time based on the chemical properties of the pollutants and the changes in their interaction forces with the surface of the medicinal slices.

[0110] ;

[0111] in, The chemical properties of pollutants are represented on the coordinate system. The performance at the place, This represents the functional relationship between pollutant distribution density and chemical properties. This is the concentration adjustment factor;

[0112] 5. The dynamic decision tree adaptive optimization algorithm monitors the water flow intensity, cleaning time, and medium concentration in real time during the cleaning process, so that various parameters are adjusted according to real-time data to ensure effective removal of pollutants and structural protection of Chinese herbal medicine pieces.

[0113] Step 4 includes:

[0114] 1. During the cleaning process, a spatiotemporal feedback intelligent control algorithm is applied to monitor multiple key cleaning parameters in real time, including water flow intensity. Cleaning time Pollutant removal rate and the concentration of the cleaning medium Data is collected through real-time sensors, and a spatiotemporal feedback model is constructed for data analysis and processing.

[0115] 2. Adjust the water flow intensity based on the monitored cleaning progress and pollutant removal status. This ensures that the water flow intensity matches the distribution density of pollutants and the removal requirements.

[0116] ;

[0117] in, The adjusted water flow intensity, This is an incremental adjustment based on real-time feedback data;

[0118] 3. Dynamically adjust cleaning time Combining real-time monitoring of pollutant removal rate and surface condition of medicinal slices:

[0119] ;

[0120] in, The adjusted cleaning time, This is the amount of time adjustment based on feedback from the removal rate and the condition of the medicinal slices;

[0121] 4. Based on pollutant removal rate Based on the chemical properties of pollutants, dynamically optimize the concentration of the cleaning medium. :

[0122] ;

[0123] in, The adjusted concentration of the cleaning medium. The adjustment amount is based on the feedback of the chemical properties of the pollutants. This is the concentration adjustment factor;

[0124] 5. Through a spatiotemporal feedback intelligent control algorithm, the system continuously monitors the changes in various parameters during the cleaning process and dynamically adjusts the water flow intensity, cleaning time, and medium concentration based on the feedback results to continuously optimize the cleaning efficiency while protecting the Chinese herbal medicine pieces.

[0125] Step 5 includes:

[0126] 1. After cleaning, a multi-dimensional data fusion intelligent detection algorithm is used to perform multi-dimensional image acquisition and fusion processing on the surface of the medicinal slices. The image acquisition is performed through multiple different angles and spectral channels to generate an image matrix of residual foreign matter containing comprehensive details. ,in, Indicates different spectral channels, and The spatial coordinates of the image;

[0127] 2. The residual foreign object image matrix Image data matrix of contaminants before cleaning Compare the results and calculate the foreign matter removal rate matrix. :

[0128] ;

[0129] in, Indicates coordinates The closer the foreign matter removal rate is to the value, the better. The better the cleaning effect;

[0130] 3. Based on the foreign matter removal rate matrix Intelligent detection algorithms are applied to classify and analyze the distribution of residual foreign matter, identify areas that are not thoroughly cleaned, and generate a residual foreign matter distribution matrix. ;

[0131] 4. Based on the residual foreign matter distribution matrix and removal rate It automatically generates a cleaning quality report, which includes the types and locations of residual foreign matter, removal efficiency, and an overall assessment of the cleaning effect. :

[0132] ;

[0133] in, Indicates coordinates Cleaning quality at the site;

[0134] 5. Based on the generated cleaning quality report, the system automatically adjusts the relevant parameters for the next round of cleaning, performs secondary cleaning on areas that have not been thoroughly cleaned, completely removes foreign objects, and stores the detection data and cleaning results in the database.

[0135] Based on the above solution, this application has the following technical effects and features:

[0136] This invention utilizes hyperspectral imaging technology combined with image enhancement algorithms that separate differential features to accurately acquire multidimensional image data of the surface of Chinese herbal medicine slices. Using this data, the system can comprehensively monitor the surface of the slices across different spectral ranges, ensuring that the characteristics of various contaminants are effectively captured. Compared to traditional methods relying on manual observation or mechanical cleaning, this invention overcomes the limitation of existing technologies in accurately identifying different types of contaminants, significantly improving identification accuracy and cleaning efficiency.

[0137] Based on hyperspectral image data, the topological feature mapping algorithm can further analyze the contaminant areas on the surface of medicinal slices, and combine it with an adaptive classification algorithm to accurately classify and identify contaminants, generating a two-dimensional distribution map of contaminants. The classification process greatly improves the accuracy of contaminant identification, not only effectively distinguishing different types of contaminants such as mud, impurities, and insects, but also providing accurate input information for the generation of cleaning solutions. This invention overcomes the shortcomings of traditional cleaning methods, such as inaccurate contaminant identification and reliance on human experience, and realizes intelligent contaminant classification and treatment.

[0138] This invention employs a dynamic decision tree adaptive optimization algorithm, which, combined with the generated contaminant distribution map and the physical properties of the Chinese herbal medicine slices, can automatically generate the optimal cleaning scheme. This scheme can dynamically adjust cleaning parameters, including water flow intensity, cleaning time, and concentration of the cleaning medium, according to the type and distribution of contaminants, ensuring that the cleaning process can flexibly respond to actual conditions. This cleaning scheme solves the problem of fixed and inflexible cleaning parameters in existing technologies, avoiding over-cleaning or under-cleaning. It not only effectively removes contaminants but also maximizes the protection of the structure and quality of the Chinese herbal medicine slices.

[0139] This invention applies a spatiotemporal feedback intelligent control algorithm to monitor the cleaning progress and contaminant removal in real time, and dynamically adjusts the cleaning parameters based on the feedback results. This intelligent feedback mechanism significantly improves the automation and intelligence of the cleaning process, overcoming the limitations of traditional cleaning equipment in adjusting cleaning parameters in real time. By precisely controlling the water flow intensity, cleaning time, and cleaning medium concentration, the cleaning effect can be continuously optimized during the cleaning process, ensuring that each piece of Chinese herbal medicine receives the best cleaning treatment and avoiding damage to the pieces due to excessive cleaning time or excessive water flow intensity.

[0140] This invention utilizes a multi-dimensional data fusion-based intelligent detection algorithm to comprehensively evaluate the cleaning effect. By comparing image data before and after cleaning, the system can accurately calculate the foreign matter removal rate, identify areas where cleaning is incomplete, and generate a detailed cleaning quality report. This achieves automated and intelligent evaluation of cleaning quality, overcoming the low-precision problem caused by reliance on manual inspection in traditional cleaning methods. Based on the cleaning quality report, the system can automatically adjust the relevant parameters for the next round of cleaning and perform secondary cleaning of residual contaminants to ensure the thoroughness and consistency of the cleaning effect. Through a feedback closed-loop mechanism, the cleaning process is continuously optimized. It can not only monitor and adjust cleaning parameters in real time but also gradually improve cleaning efficiency and accuracy through learning and analysis of historical data. Data from each cleaning process is stored and used to optimize the parameter settings for the next round of cleaning, thus forming an adaptive intelligent cleaning process.

[0141] Obviously, the specific implementation of this invention is not limited to the above-described manner. Any non-substantial improvements made using the inventive concept and technical solution of this invention are within the protection scope of this invention.

Claims

1. A method for cleaning traditional Chinese medicine decoction pieces based on artificial intelligence, characterized in that: Includes the following steps: S1. Images of the surface of Chinese herbal medicine slices are acquired using a hyperspectral imaging system, and the images are processed using an image enhancement algorithm that separates differential features to obtain an enhanced image that highlights the contaminant features on the surface of the slices. S2. Based on the enhanced image, an adaptive classification algorithm using topological feature mapping is used to classify and identify contaminants on the surface of medicinal slices, generating a two-dimensional distribution map of the contaminants. S3. Based on the generated pollutant distribution map and combined with the physical characteristics of the medicinal slices, a dynamic decision tree adaptive optimization algorithm is adopted to dynamically adjust the water flow intensity, cleaning time and cleaning medium concentration parameters according to different pollutant types and distribution, so as to effectively remove pollutants and protect the medicinal slices during the cleaning process.

2. The method for cleaning traditional Chinese medicine decoction pieces based on artificial intelligence as described in claim 1, characterized in that: The cleaning method further includes step S4, which involves applying a spatiotemporal feedback intelligent control algorithm during the cleaning process to monitor the cleaning progress and pollutant removal in real time, and dynamically adjusting the cleaning parameters based on the feedback results until the cleaning is completed or ended.

3. A method for cleaning traditional Chinese medicine decoction pieces based on artificial intelligence as described in claim 1 or 2, characterized in that: The cleaning method further includes step S5, which involves detecting and evaluating residual foreign matter on the medicinal slices through a multi-dimensional data fusion intelligent detection algorithm after cleaning, and automatically generating a cleaning quality report.

4. A method for cleaning traditional Chinese medicine decoction pieces based on artificial intelligence as described in claim 1 or 2, characterized in that: In step S1, a hyperspectral imaging system is used to acquire multispectral images of the surface of the Chinese herbal medicine slices. The hyperspectral imaging system acquires multidimensional image data of the surface of the Chinese herbal medicine slices in multiple different wavelength ranges, generating an image matrix containing different spectral information. ;in, Indicates wavelength. and Represents the two-dimensional spatial coordinates of the image. Indicates at wavelength Location, coordinates Image intensity value; The acquired image data is preprocessed using a filtering function. Eliminate noise and interference signals to generate a denoised image matrix. : ； in, Indicates at wavelength The following is a noise filtering function for the image.

5. The method for cleaning traditional Chinese medicine decoction pieces based on artificial intelligence as described in claim 4, characterized in that: The image enhancement algorithm using differential feature separation in step S1 processes the image, including: processing the denoised image matrix. Spectral weighting is performed, and based on the spectral differences between pollutants and background at different wavelengths, a weighting function is used to... Enhance pollutant characteristics to generate an enhanced image matrix : ； in, These are the weighting coefficients for different wavelengths; based on the enhanced image matrix Generate the final enhanced image used for identification and cleaning.

6. A method for cleaning traditional Chinese medicine decoction pieces based on artificial intelligence as described in claim 1 or 2, characterized in that: Step S2 includes: generating an enhanced image matrix. The image data is processed using a topological feature mapping algorithm. This algorithm analyzes the changes in the topological structure of image pixels at different wavelengths to identify and separate contaminant regions on the surface of the medicinal slices, generating a preliminary contaminant marker matrix, denoted as... ,in Indicates coordinates The identified contaminant area; Based on the preliminary pollutant labeling matrix An adaptive classification algorithm is applied to further classify pollutant regions. This algorithm performs multi-level classification based on the spectral characteristics, geometric shape, and edge features of pollutants, generating a high-precision pollutant classification matrix. ,in Indicates coordinates The corresponding pollutant category label; The generated pollutant classification matrix Optimization was performed by iteratively updating the model to improve classification accuracy. A probabilistic model of pollutant features was constructed by combining Bayesian inference with the spatial distribution patterns of pollutant features. The pollutant category identification results are optimized, and a pollutant classification matrix is ​​finally generated. The optimized pollutant classification matrix By fusing with hyperspectral information, a final two-dimensional distribution map of pollutants is generated. The two-dimensional distribution map is based on the spectral weighting function. Perform weighted processing: ； in, This is a two-dimensional distribution map of pollutants. This is a weighted sum for each spectral channel.

7. A method for cleaning traditional Chinese medicine decoction pieces based on artificial intelligence as described in claim 1 or 2, characterized in that: Step S3 includes using the generated pollutant distribution map. Combining the physical properties of Chinese herbal medicine slices, a dynamic decision tree adaptive optimization algorithm is used to construct a multivariate optimization model and automatically generate a cleaning plan. The optimization model comprehensively considers the type, distribution, thickness of contaminants on the surface of the slices and the material properties of the slices. During the cleaning scheme generation process, the dynamic decision tree adaptive optimization algorithm adjusts the water flow intensity in real time through iterative calculations. The water flow intensity is dynamically adjusted based on the distribution density and thickness of pollutants, as well as the pollutant's response to the water flow. ； in, Indicates the pollutants in the coordinates Distribution density at that location, For the thickness of the pollutants, and The weighting coefficient controls the sensitivity of water flow intensity to the characteristics of different pollutants; Dynamically adjust cleaning time The cleaning time is based on the distribution density and thickness of contaminants and the water absorption of the medicinal slices, ensuring that the cleaning time can thoroughly remove contaminants while avoiding over-cleaning of the medicinal slices. ； in, For the medicinal slices in coordinates The water absorption coefficient at that location This is the cleaning rate coefficient; The concentration of the cleaning medium is dynamically adjusted according to different types of pollutants. The concentration is optimized in real time based on the chemical properties of the pollutants and the changes in their interaction forces with the surface of the medicinal slices. ； in, The chemical properties of pollutants are represented on a coordinate system. The performance at the place, This represents the functional relationship between pollutant distribution density and chemical properties. This is the concentration adjustment factor; The dynamic decision tree adaptive optimization algorithm monitors the water flow intensity, cleaning time, and medium concentration in real time during the cleaning process, so that various parameters are adjusted according to real-time data to ensure effective removal of pollutants and structural protection of Chinese herbal medicine pieces.

8. The method for cleaning traditional Chinese medicine decoction pieces based on artificial intelligence as described in claim 2, characterized in that: Step S4 includes: During the cleaning process, several key cleaning parameters are monitored in real time, including water flow intensity. Cleaning time Pollutant removal rate and the concentration of cleaning medium It collects data through real-time sensors and performs data analysis and processing; Adjust the water flow intensity based on the monitored cleaning progress and pollutant removal status. This ensures that the water flow intensity matches the distribution density of pollutants and the removal requirements. ； in, The adjusted water flow intensity, This is an incremental adjustment based on real-time feedback data; Dynamically adjust cleaning time Combining real-time monitoring of pollutant removal rate and surface condition of medicinal slices: ； in, The adjusted cleaning time, This is the amount of time adjustment based on feedback from the removal rate and the condition of the medicinal slices; Based on pollutant removal rate Based on the chemical properties of pollutants, dynamically optimize the concentration of the cleaning medium. : ； in, The adjusted concentration of the cleaning medium. The adjustment amount is based on the feedback of the chemical properties of the pollutants. This is the concentration adjustment factor; By continuously monitoring the changes in various parameters during the cleaning process, and dynamically adjusting the water flow intensity, cleaning time, and medium concentration based on the feedback results, the cleaning efficiency is continuously optimized while protecting the Chinese herbal medicine pieces.

9. The method for cleaning traditional Chinese medicine decoction pieces based on artificial intelligence as described in claim 3, characterized in that: Step S5 includes performing multi-dimensional image acquisition and fusion processing on the surface of the medicinal slices after cleaning. The image acquisition is performed through multiple different angles and spectral channels to generate an image matrix of residual foreign matter containing comprehensive details. ,in, Indicates different spectral channels, and The spatial coordinates of the image; The residual foreign object image matrix Image data matrix of contaminants before cleaning Compare the results and calculate the foreign matter removal rate matrix. : ； in, Indicates coordinates The closer the foreign matter removal rate is to the value, the better. The better the cleaning effect; Based on foreign matter removal rate matrix The remaining foreign matter is classified and its distribution is analyzed to identify areas that are not thoroughly cleaned, and a distribution matrix of the remaining foreign matter is generated. ; Based on the residual foreign matter distribution matrix and removal rate It automatically generates a cleaning quality report, which includes the types and locations of residual foreign matter, removal efficiency, and an overall assessment of the cleaning effect. : ； in, Indicates coordinates The quality of cleaning at the site.

10. The method for cleaning traditional Chinese medicine decoction pieces based on artificial intelligence as described in claim 9, characterized in that: Step S5 also includes automatically adjusting the relevant parameters for the next round of cleaning based on the generated cleaning quality report, performing secondary cleaning on areas that were not thoroughly cleaned, completely removing foreign matter, and storing the detection data and cleaning results in the database.

Images