Intelligent screening method and system for traditional Chinese medicine decoction pieces

Abstract

The invention relates to the technical field of material determination, and particularly discloses an intelligent screening method and system for traditional Chinese medicine decoction pieces, and the method comprises the steps: S1, obtaining decoction piece standard data of to-be-detected traditional Chinese medicine decoction pieces, collecting decoction piece physical data and decoction piece chemical data of the to-be-detected traditional Chinese medicine decoction pieces, and obtaining historical detection data of the to-be-detected traditional Chinese medicine decoction pieces; s2, calculating a feature deviation value of each feature in the physical feature vector of the to-be-detected traditional Chinese medicine decoction piece and a physical deviation value of the to-be-detected traditional Chinese medicine decoction piece; s3, constructing a chemical detection model, and determining chemical deviation data of the to-be-detected traditional Chinese medicine decoction pieces; and S4, determining the traditional Chinese medicine decoction piece grade of the to-be-detected traditional Chinese medicine decoction pieces and performing intelligent screening. According to the method, a comprehensive and accurate decoction piece standard system can be constructed, multi-dimensional and quantitative evaluation of the quality of the decoction pieces is realized, the grade of the decoction pieces is comprehensively and accurately judged, subjective errors and quality misjudgment are avoided, the screening efficiency and accuracy are improved, and a scientific solution is provided for quality control of traditional Chinese medicines.

Description

Technical Field

[0001] The present invention relates to the technical field of data retrieval, and particularly to an intelligent screening method and system for traditional Chinese medicine decoction pieces. Background Art

[0002] The quality control and screening of traditional Chinese medicine decoction pieces have always been important topics in the traditional Chinese medicine industry. Traditional methods mainly rely on experienced pharmacists to make judgments through the senses, combined with simple physical and chemical detection means, such as character observation, moisture determination, component content detection, etc. However, this method has problems such as strong subjectivity, low efficiency, and difficulty in quantification, and it is difficult to meet the standardized and large-scale requirements of modern traditional Chinese medicine production. In recent years, with the development of computer technology, image processing technology, and chemical analysis technology, intelligent detection methods have gradually been introduced into the quality control of traditional Chinese medicine decoction pieces. For example, machine vision technology is used to analyze the appearance characteristics of decoction pieces, and spectral and chromatographic technologies are used to quantitatively analyze chemical components, providing the possibility for accurate screening of traditional Chinese medicine decoction pieces. However, there is no comprehensive intelligent screening method in the existing technology that combines appearance characteristics, chemical components, and historical detection data, and it is impossible to fully utilize multi-dimensional data to comprehensively evaluate decoction pieces.

[0003] Therefore, the present invention proposes an intelligent screening method and system for traditional Chinese medicine decoction pieces. Summary of the Invention

[0004] The present invention provides an intelligent screening method and system for traditional Chinese medicine decoction pieces. By analyzing the decoction piece standard data, historical detection data of the traditional Chinese medicine decoction pieces to be detected, as well as the decoction piece physical data and decoction piece chemical data collected from the traditional Chinese medicine decoction pieces to be detected, calculating the physical deviation value of the traditional Chinese medicine decoction pieces to be detected, determining the chemical deviation data of the traditional Chinese medicine decoction pieces to be detected, and determining the grade of the traditional Chinese medicine decoction pieces of the traditional Chinese medicine decoction pieces to be detected, the intelligent screening of the traditional Chinese medicine decoction pieces to be detected is realized. A comprehensive and accurate decoction piece standard system can be constructed to realize multi-dimensional and quantitative evaluation of the quality of decoction pieces, comprehensively and accurately judge the grade of decoction pieces, avoid subjective errors and quality misjudgments, improve the screening efficiency and accuracy, and provide a scientific solution for traditional Chinese medicine quality control.

[0005] The present invention provides an intelligent screening method for traditional Chinese medicine decoction pieces, including: S1: Obtain the decoction piece standard data of the traditional Chinese medicine decoction pieces to be detected, collect the decoction piece physical data of the traditional Chinese medicine decoction pieces to be detected, collect the decoction piece chemical data of the traditional Chinese medicine decoction pieces to be detected, and obtain the historical detection data of the traditional Chinese medicine decoction pieces to be detected; S2: Analyze the decoction piece physical data of the traditional Chinese medicine decoction pieces to be detected, determine the appearance feature vector, appearance feature value vector, and physical feature vector of the traditional Chinese medicine decoction pieces to be detected, calculate the feature deviation value of each feature in the physical feature vector of the traditional Chinese medicine decoction pieces to be detected, and calculate the physical deviation value of the traditional Chinese medicine decoction pieces to be detected; S3: Construct a chemical detection model based on the chemical data, historical detection data, and standard data of the Chinese herbal pieces to be detected, and determine the chemical deviation data of the Chinese herbal pieces to be detected based on the chemical data of the pieces and the chemical detection model; S4: Determine the grade of the Chinese herbal pieces to be detected based on the physical deviation value and chemical deviation data of the Chinese herbal pieces to be detected, and perform intelligent screening based on the grade of the Chinese herbal pieces to be detected.

[0006] Preferably, based on an intelligent screening method for a type of Chinese herbal pieces, obtain the standard data of the Chinese herbal pieces to be detected, including: Extract the first standard data of the Chinese herbal pieces to be detected from the pharmacopoeia, and extract the second standard data of the Chinese herbal pieces to be detected from the industry standard database. Among them, the first standard data includes the first standard sub-data of multiple sub-standards, and the second standard data includes the second standard sub-data of multiple sub-standards; Traverse each sub-standard in the second standard data. If the first standard data includes the first standard sub-data of the sub-standard currently traversed in the second standard data, determine the standard sub-range of each parameter of each sub-standard in the second standard data currently traversed based on the first standard sub-data, and mark the existence of the sub-standard currently traversed in the second standard data; For each sub-standard in the second standard data with an existence mark, determine the standard sub-range of each parameter of each sub-standard with an existence mark based on the second standard sub-data of each sub-standard in the second standard data; For each sub-standard in the second standard data without an existence mark, determine the standard sub-range of each parameter of each sub-standard without an existence mark based on the second standard sub-data of each sub-standard in the second standard data; For each sub-standard in the first standard data without an existence mark, determine the standard sub-range of each parameter of each sub-standard without an existence mark based on the first standard sub-data of each sub-standard in the first standard data; Determine the standard parameter vector based on all the parameters of all the sub-standards with existence marks and all the parameters of all the sub-standards without existence marks; Determine the standard data of the Chinese herbal pieces to be detected based on the standard parameter vector, the standard sub-range of all the parameters of all the sub-standards with existence marks, the standard sub-range, and the standard sub-range of all the parameters of all the sub-standards without existence marks.

[0007] Preferably, based on an intelligent screening method for a type of Chinese herbal pieces, collect the physical data of the Chinese herbal pieces to be detected, collect the chemical data of the Chinese herbal pieces to be detected, and obtain the historical detection data of the Chinese herbal pieces to be detected, including: Based on the appearance image data of the traditional Chinese medicine decoction pieces to be detected collected by an image acquisition device, wherein the appearance image data includes the appearance images of multiple collection positions of the traditional Chinese medicine decoction pieces to be detected; Based on the texture data of the traditional Chinese medicine decoction pieces to be detected collected by a texture detection device, wherein the texture data includes multiple texture characteristics of the traditional Chinese medicine decoction pieces to be detected and the texture characteristic values of each texture characteristic; Based on the weight data of the traditional Chinese medicine decoction pieces to be detected collected by a weight detection device, wherein the weight data includes multiple weight characteristics of the traditional Chinese medicine decoction pieces to be detected and the weight characteristic values of each weight characteristic; Based on the appearance image data, texture data and weight data of the traditional Chinese medicine decoction pieces to be detected, determine the physical data of the traditional Chinese medicine decoction pieces to be detected; Based on the chemical data of the traditional Chinese medicine decoction pieces to be detected collected by a chemical detection device, wherein the chemical data of the decoction pieces includes multiple component characteristics of the traditional Chinese medicine decoction pieces to be detected and the component characteristic values of each component characteristic; Obtain the historical detection data of the traditional Chinese medicine decoction pieces to be detected, wherein the historical detection data includes the historical physical detection vectors, historical chemical detection vectors and historical chemical detection grades of multiple historical traditional Chinese medicine decoction pieces.

[0008] Preferably, based on an intelligent screening method for a traditional Chinese medicine decoction piece, analyze the physical data of the traditional Chinese medicine decoction piece to be detected, determine the appearance feature vector, appearance feature value vector and physical feature vector of the traditional Chinese medicine decoction piece to be detected, calculate the feature deviation value of each feature in the physical feature vector of the traditional Chinese medicine decoction piece to be detected, and calculate the physical deviation value of the traditional Chinese medicine decoction piece to be detected, including: Perform image preprocessing on the appearance image of each collection position in the appearance image data of the traditional Chinese medicine decoction piece to be detected; Perform binarization processing on the appearance images of all collection positions in the preprocessed appearance image data of the traditional Chinese medicine decoction piece to be detected and extract the contour of the traditional Chinese medicine decoction piece to be detected, perform feature extraction on the extracted contour of the traditional Chinese medicine decoction piece to be detected, and determine the shape feature vector and shape feature value vector of the traditional Chinese medicine decoction piece to be detected, wherein the shape feature vector includes multiple shape features, and the shape feature value vector includes the shape feature values of each shape feature in the shape feature vector; Convert the appearance images of all collection positions in the preprocessed appearance image data of the traditional Chinese medicine decoction piece to be detected to the HSV color space, perform color feature extraction on the converted appearance images of all collection positions, and determine the color feature vector and color feature value vector of the traditional Chinese medicine decoction piece to be detected, wherein the color feature vector includes multiple color features, and the color feature value vector includes the color feature values of each color feature in the color feature vector; Grayscale the appearance images at all acquisition positions in the appearance image data of the traditional Chinese medicine cut pieces to be detected after preprocessing, extract texture features from the appearance images at all acquisition positions after grayscale processing, and determine the texture feature vector and texture eigenvalue vector of the traditional Chinese medicine cut pieces to be detected. Among them, the texture feature vector includes multiple texture features, and the texture eigenvalue vector includes the texture eigenvalues of each texture feature in the texture feature vector; Generate the appearance feature vector of the traditional Chinese medicine cut pieces to be detected based on the shape feature vector, color feature vector, and texture feature vector of the traditional Chinese medicine cut pieces to be detected. At the same time, determine the appearance eigenvalue vector of the traditional Chinese medicine cut pieces to be detected based on the shape eigenvalue vector, color eigenvalue vector, texture eigenvalue vector, and appearance feature vector of the traditional Chinese medicine cut pieces to be detected.

[0009] Preferably, based on an intelligent screening method for traditional Chinese medicine cut pieces, analyze the physical data of the traditional Chinese medicine cut pieces to be detected, determine the appearance feature vector, appearance eigenvalue vector, and physical feature vector of the traditional Chinese medicine cut pieces to be detected, calculate the feature deviation value of each feature in the physical feature vector of the traditional Chinese medicine cut pieces to be detected, and calculate the physical deviation value of the traditional Chinese medicine cut pieces to be detected. It also includes: Determine the physical feature vector of the traditional Chinese medicine cut pieces to be monitored based on all texture features in the texture data, all weight features in the weight data, and the appearance feature vector of the traditional Chinese medicine cut pieces to be monitored; Calculate the feature deviation value of each feature in the physical feature vector of the traditional Chinese medicine cut pieces to be detected based on the physical feature vector, historical detection data, and cut piece standard data of the traditional Chinese medicine cut pieces to be monitored; Calculate the physical deviation value of the traditional Chinese medicine cut pieces to be detected based on the feature deviation values of all features in the physical feature vector of the traditional Chinese medicine cut pieces to be detected.

[0010] Preferably, based on an intelligent screening method for traditional Chinese medicine cut pieces, construct a chemical detection model based on the chemical data, historical detection data, and cut piece standard data of the traditional Chinese medicine cut pieces to be detected, and determine the chemical deviation data of the traditional Chinese medicine cut pieces to be detected based on the chemical data and the chemical detection model, including: Construct a chemical detection model for cut pieces based on the cut piece standard data, historical chemical detection vectors, and historical chemical detection grades of all historical traditional Chinese medicine cut pieces in the historical detection data; Input the chemical data of the traditional Chinese medicine cut pieces to be detected into the constructed chemical detection model for cut pieces, and determine the chemical deviation data and chemical detection grade of the traditional Chinese medicine cut pieces to be detected based on the output result of the chemical detection model for cut pieces. Among them, the chemical deviation data includes multiple component features with deviations and the deviation values of each component feature with deviations, and the chemical detection grade includes excellent, good, medium, and poor.

[0011] Preferably, based on an intelligent screening method for traditional Chinese medicine decoction pieces, the grade of the traditional Chinese medicine decoction pieces to be detected is determined based on the physical deviation value and chemical deviation data of the traditional Chinese medicine decoction pieces to be detected, including: Based on the physical deviation value of the traditional Chinese medicine decoction pieces to be detected, determine the physical detection grade of the traditional Chinese medicine decoction pieces to be detected; Based on the physical detection grade of the traditional Chinese medicine decoction pieces to be detected and the chemical detection grade in the chemical deviation data, determine the grade of the traditional Chinese medicine decoction pieces to be detected.

[0012] The present invention provides an intelligent screening system for traditional Chinese medicine decoction pieces, which is used to execute the intelligent screening method for traditional Chinese medicine decoction pieces described in any one of Embodiments 1 to 7.

[0013] Acquisition module: Obtain the standard data of the decoction pieces of the traditional Chinese medicine decoction pieces to be detected, collect the physical data of the decoction pieces of the traditional Chinese medicine decoction pieces to be detected, collect the chemical data of the decoction pieces of the traditional Chinese medicine decoction pieces to be detected, and obtain the historical detection data of the traditional Chinese medicine decoction pieces to be detected; Physical module: Analyze the physical data of the decoction pieces of the traditional Chinese medicine decoction pieces to be detected, determine the appearance feature vector, appearance eigenvalue vector and physical feature vector of the traditional Chinese medicine decoction pieces to be detected, calculate the feature deviation value of each feature in the physical feature vector of the traditional Chinese medicine decoction pieces to be detected, and calculate the physical deviation value of the traditional Chinese medicine decoction pieces to be detected; Chemical module: Construct a chemical detection model based on the chemical data of the decoction pieces of the traditional Chinese medicine decoction pieces to be detected, historical detection data and standard data of the decoction pieces, and determine the chemical deviation data of the traditional Chinese medicine decoction pieces to be detected based on the chemical data of the decoction pieces and the chemical detection model; Screening module: Based on the physical deviation value and chemical deviation data of the traditional Chinese medicine decoction pieces to be detected, determine the grade of the traditional Chinese medicine decoction pieces to be detected, and perform intelligent screening based on the grade of the traditional Chinese medicine decoction pieces to be detected.

[0014] The beneficial effects of the present invention compared with the prior art are as follows: By analyzing the obtained standard data of the decoction pieces of the traditional Chinese medicine decoction pieces to be detected, historical detection data, and the collected physical data and chemical data of the decoction pieces of the traditional Chinese medicine decoction pieces to be detected, calculate the physical deviation value of the traditional Chinese medicine decoction pieces to be detected, determine the chemical deviation data of the traditional Chinese medicine decoction pieces to be detected, and determine the grade of the traditional Chinese medicine decoction pieces to be detected to realize the intelligent screening of the traditional Chinese medicine decoction pieces to be detected. A comprehensive and accurate decoction piece standard system can be constructed to realize multi-dimensional and quantitative evaluation of the quality of decoction pieces, comprehensively and accurately judge the grade of decoction pieces, avoid subjective errors and quality misjudgments, improve the screening efficiency and accuracy, and provide a scientific solution for the quality control of traditional Chinese medicine.

[0015] Other features and advantages of the present invention will be described in the following specification, and partly will become apparent from the specification, or will be understood by implementing the present invention. The objectives and other advantages of the present invention can be achieved and obtained by the structures specifically pointed out in this application document.

[0016] The technical solution of the present invention will be further described in detail below through the drawings and embodiments. Description of the Drawings

[0017] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention, and do not constitute a limitation to the present invention. In the drawings: Figure 1 It is a flowchart of an intelligent screening method for a traditional Chinese medicine decoction piece in an embodiment of the present invention; Figure 2 It is a schematic diagram of an intelligent screening system for a traditional Chinese medicine decoction piece in an embodiment of the present invention. Detailed Embodiments

[0018] The following describes the preferred embodiments of the present invention with reference to the drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention. Embodiment 1:

[0019] The present invention provides an intelligent screening method for traditional Chinese medicine decoction pieces. Referring to Figure 1 , it includes: S1: Obtain the decoction piece standard data of the traditional Chinese medicine decoction piece to be detected, collect the physical data of the traditional Chinese medicine decoction piece to be detected, collect the chemical data of the traditional Chinese medicine decoction piece to be detected, and obtain the historical detection data of the traditional Chinese medicine decoction piece to be detected; S2: Analyze the physical data of the traditional Chinese medicine decoction piece to be detected, determine the appearance feature vector, appearance eigenvalue vector and physical feature vector of the traditional Chinese medicine decoction piece to be detected, calculate the feature deviation value of each feature in the physical feature vector of the traditional Chinese medicine decoction piece to be detected, and calculate the physical deviation value of the traditional Chinese medicine decoction piece to be detected; S3: Construct a chemical detection model based on the chemical data, historical detection data and decoction piece standard data of the traditional Chinese medicine decoction piece to be detected, and determine the chemical deviation data of the traditional Chinese medicine decoction piece to be detected based on the chemical data of the decoction piece and the chemical detection model; S4: Based on the physical deviation value and chemical deviation data of the traditional Chinese medicine decoction piece to be detected, determine the grade of the traditional Chinese medicine decoction piece to be detected, and perform intelligent screening based on the grade of the traditional Chinese medicine decoction piece to be detected.

[0020] In this embodiment, the past detection records of the same variety of decoction pieces are retrieved from the enterprise database.

[0021] In this embodiment, after preprocessing the appearance image, operations such as binarization and contour extraction are performed. Taking chrysanthemum cut pieces as an example, shape features such as the shape of its petals and the diameter of the flower, color features such as hue and saturation, and texture features such as the degree of texture roughness are extracted to form an appearance feature vector; the specific numerical values of each feature (such as the flower diameter of 3 cm and the main hue H value of 30°) constitute an appearance feature value vector.

[0022] In this embodiment, the historical chemical detection vectors, historical chemical detection grades in the historical detection data are integrated with the cut piece standard data, and a model is trained using a machine learning algorithm (such as the random forest algorithm). For example, training on a large amount of historical data of ginseng cut pieces enables the model to identify the quality grades corresponding to different chemical detection data.

[0023] In this embodiment, the chemical data of the cut piece to be detected is input into the model, the model outputs the chemical detection grade, and the component features and deviation values with deviations are determined by comparing with the standard data. For example, when detecting a batch of ginseng cut pieces, the model determines that the chemical detection grade is "medium", and it is found that the content of ginsenoside Rg1 is lower than the standard, thus obtaining the chemical deviation data.

[0024] In this embodiment, the physical detection grade is determined according to the physical deviation value (such as "excellent" for a low physical deviation value), combined with the chemical detection grade in the chemical deviation data, and the cut piece grade is comprehensively evaluated according to the preset rules. For example, if the physical detection grade is "good" and the chemical detection grade is "good", then the cut piece grade is "first grade".

[0025] Intelligent screening is carried out: In this embodiment, through devices such as robotic arms and pneumatic sorting devices, sorting is performed according to the cut piece grade. The "super grade" and "first grade" cut pieces are sorted to the high-quality product packaging line, and the "unqualified" cut pieces are sorted to the waste treatment line.

[0026] The beneficial effects of the above technology are as follows: By analyzing the cut piece standard data, historical detection data of the cut piece of traditional Chinese medicine to be detected, as well as the cut piece physical data and cut piece chemical data collected, calculating the physical deviation value of the cut piece of traditional Chinese medicine to be detected, determining the chemical deviation data of the cut piece of traditional Chinese medicine to be detected, and determining the cut piece grade of the cut piece of traditional Chinese medicine to be detected, the intelligent screening of the cut piece of traditional Chinese medicine to be detected is realized. A comprehensive and accurate cut piece standard system can be constructed to achieve multi-dimensional and quantitative evaluation of the cut piece quality, comprehensively and accurately judge the cut piece grade, avoid subjective errors and quality misjudgments, improve the screening efficiency and accuracy, and provide a scientific solution for the quality control of traditional Chinese medicine. Embodiment 2:

[0027] Based on Embodiment 1, an intelligent screening method for cut pieces of traditional Chinese medicine, obtaining the cut piece standard data of the cut piece of traditional Chinese medicine to be detected, including: Extract the first standard data of the Chinese herbal medicine pieces to be detected from the pharmacopoeia, and extract the second standard data of the Chinese herbal medicine pieces to be detected from the industry standard database. Among them, the first standard data includes the first standard sub-data of multiple sub-standards, and the second standard data includes the second standard sub-data of multiple sub-standards; Traverse each sub-standard in the second standard data. If the first standard data includes the first standard sub-data of the sub-standard in the currently traversed second standard data, determine the piece standard sub-range of each parameter of each sub-standard in the currently traversed second standard data based on the first standard sub-data, and mark the existence of the sub-standard in the currently traversed second standard data; For each sub-standard with an existence mark in the second standard data, determine the piece standard secondary range of each parameter of each sub-standard with an existence mark based on the second standard sub-data of each sub-standard in the second standard data; For each sub-standard without an existence mark in the second standard data, determine the piece standard sub-range of each parameter of each sub-standard without an existence mark based on the second standard sub-data of each sub-standard in the second standard data; For each sub-standard without an existence mark in the first standard data, determine the piece standard sub-range of each parameter of each sub-standard without an existence mark based on the first standard sub-data of each sub-standard in the first standard data; Determine the standard parameter vector based on all the parameters of all the sub-standards with existence marks and all the parameters of all the sub-standards without existence marks; Determine the piece standard data of the Chinese herbal medicine pieces to be detected based on the standard parameter vector, the piece standard sub-range of all the parameters of all the sub-standards with existence marks, the piece standard secondary range, and the piece standard sub-range of all the parameters of all the sub-standards without existence marks.

[0028] In this embodiment, extract the first standard data of the Chinese herbal medicine pieces to be detected from the pharmacopoeia. For example, when detecting Astragalus membranaceus pieces, the pharmacopoeia stipulates that its shape is round or oval thick slices, with a diameter of 0.8 - 3.5 cm and a moisture content not exceeding 13.0%. These data are the first standard data; extract the second standard data from the industry standard database. For example, the industry standard may have additional regulations on the color uniformity and slice thickness uniformity of Astragalus membranaceus pieces. These requirements constitute the second standard data.

[0029] In this embodiment, each parameter in the second standard data is traversed. For example, when traversing the pharmacopoeia regulations of the sub-standard of the color of Astragalus membranaceus in the industry standard, it is specified that the periphery is light brownish-yellow to light brownish-brown, the cortex of the cut surface is yellowish-white, the xylem is light yellow, there are radial textures and fissures, fibrous, and powdery; the industry standard is that the periphery of the cut pieces is light brownish-yellow to light brownish-brown, the cortex of the cut surface is yellowish-white, the xylem is light yellow, and the color is uniform. Then, the first standard sub-data of the first standard data regarding the color sub-standard is considered. At this time, based on the standards for colors in the pharmacopoeia, the standard sub-range of the cut pieces for the color sub-standard is determined.

[0030] In this embodiment, for each parameter with an existence mark in the second standard data, such as the marked color sub-standard, based on the second standard sub-data of this parameter in the industry standard, its standard sub-range of the cut pieces is further determined. For example, "the color difference value of Astragalus membranaceus cut pieces in the same batch is within a certain RGB range, and the proportion of the main color is within a certain range".

[0031] In this embodiment, for each sub-standard without an existence mark in the second standard data, such as the unique sub-standard of the initial microbial content in the industry standard, based on the regulations of the industry standard for this sub-standard, the standard sub-range of the initial microbial content is determined. For example, the total number of colonies of Astragalus membranaceus cut pieces before packaging shall not exceed X CFU / g.

[0032] In this embodiment, for each sub-standard without an existence mark in the first standard data, such as the pharmacodynamic sub-standard stipulated in the pharmacopoeia but not covered by the industry standard, at this time, based on the pharmacopoeia regulations, the standard sub-range of the content parameter of astragaloside IV in the pharmacodynamic sub-standard is determined (the content of astragaloside IV shall not be less than 0.08%).

[0033] In this embodiment, all parameters of all sub-standards with existence marks and all parameters of all sub-standards without existence marks are summarized to form a standard parameter vector, comprehensively covering various indicators affecting the quality of Astragalus membranaceus cut pieces.

[0034] The beneficial effects of the above technology are as follows: Obtain the standard data of the cut pieces of the traditional Chinese medicine to be detected, which can dynamically integrate the parameter ranges and construct a more comprehensive and accurate cut piece standard system. Avoid standard conflicts and omissions, and improve the applicability of the standards. Example 3:

[0035] Based on Example 1, an intelligent screening method for traditional Chinese medicine cut pieces, which collects the physical data of the cut pieces of the traditional Chinese medicine to be detected, collects the chemical data of the cut pieces of the traditional Chinese medicine to be detected, and obtains the historical detection data of the traditional Chinese medicine to be detected, including: Collect the appearance image data of the traditional Chinese medicine cut pieces to be detected based on an image acquisition device. Among them, the appearance image data includes the appearance images of multiple collection positions of the traditional Chinese medicine cut pieces to be detected. Collect the texture data of the Chinese herbal medicine pieces to be detected based on a texture detection device, where the texture data includes multiple texture characteristics of the Chinese herbal medicine pieces to be detected and the texture characteristic values of each texture characteristic; Collect the weight data of the Chinese herbal medicine pieces to be detected based on a weight detection device, where the weight data includes multiple weight characteristics of the Chinese herbal medicine pieces to be detected and the weight characteristic values of each weight characteristic; Determine the physical data of the Chinese herbal medicine pieces to be detected based on the appearance image data, texture data, and weight data of the Chinese herbal medicine pieces to be detected; Collect the chemical data of the Chinese herbal medicine pieces to be detected based on a chemical detection device, where the chemical data of the pieces includes multiple component characteristics of the Chinese herbal medicine pieces to be detected and the component characteristic values of each component characteristic; Obtain the historical detection data of the Chinese herbal medicine pieces to be detected, where the historical detection data includes the historical physical detection vectors, historical chemical detection vectors, and historical chemical detection grades of multiple historical Chinese herbal medicine pieces.

[0036] In this embodiment, the image acquisition device can be a multispectral high-definition camera, a high-resolution industrial camera, etc. For example: Use an industrial camera (such as a 5-megapixel CMOS camera) with a ring light source to take pictures of the Chinese herbal medicine pieces to be detected from multiple angles (such as the top, 45° side, and bottom) to ensure complete coverage of the appearance characteristics. The resolution of each image is ≥2048×1536 pixels, and the frame rate is ≥10 frames per second to avoid motion blur.

[0037] In this embodiment, for example: When detecting Angelica sinensis pieces, set 3 groups of cameras on the conveyor belt to collect images of the top surface (observing the color of the cut surface), side surface (measuring the thickness), and bottom surface (identifying impurity attachment) of the pieces respectively. Each group of cameras is equipped with a white diffuse reflection light source to eliminate the reflection interference and ensure that the color of the pieces in the image is consistent with the actual situation.

[0038] In this embodiment, the texture detection device can be a pressure sensor, a vibration analyzer, a texture analyzer, etc. For example: Apply physical stress to the pieces through a texture detector (such as a TA.XT Plus type texture analyzer), record the force-displacement curve, and extract characteristics such as hardness, brittleness, and toughness. Each sample is detected 3 times repeatedly, and the average value is taken to reduce the error.

[0039] In this embodiment, for example: When detecting Gastrodia elata pieces, use a probe with a diameter of 2 mm to press down on the pieces at a speed of 1 mm / s, and record the maximum breaking force (hardness) and the displacement at the time of fracture (brittleness). If the standard requires the hardness to be ≥50 N, and the three detection values of a certain sample are 52 N, 50 N, and 49 N, then the average hardness value is 50.3 N, which meets the standard.

[0040] In this embodiment, the texture characteristics can be the hardness, toughness, brittleness, etc. of the Chinese herbal medicine pieces to be detected.

[0041] In this embodiment, the weight detection device can be a rapid moisture analyzer, an oven, a high-precision electronic scale, etc. For example, the rapid moisture analyzer quickly measures the moisture content through the principles of heating and weighing. The oven calculates the moisture content by drying to a constant weight at 105°C.

[0042] In this embodiment, the weight characteristics can be the weight, water content, etc. of the traditional Chinese medicine decoction pieces to be detected.

[0043] In this embodiment, the chemical detection device can be a spectral analysis instrument, a chromatographic analysis instrument, etc. For example, when using a near-infrared spectrometer, the traditional Chinese medicine decoction pieces to be detected are placed in a sample cell. After the near-infrared light irradiates the decoction pieces, the spectral information of the reflected or transmitted light is collected, and the relationship between the spectrum and the component content is analyzed through a chemometric model (such as partial least squares method) to predict the component characteristic values of each component feature; the Raman spectrometer utilizes the Raman scattering effect to obtain the characteristic spectrum of the molecular vibration and rotation of the decoction pieces to identify the types of effective components; the high-performance liquid chromatograph injects the extract of the decoction pieces into the chromatographic column, separates according to the difference in the distribution coefficients of different components in the stationary phase and the mobile phase, and then detects the peak areas of each component through a detector (such as an ultraviolet detector, a fluorescence detector) to calculate the component characteristic values.

[0044] In this embodiment, for example, when detecting licorice decoction pieces, the content of glycyrrhizic acid determined by HPLC is 2.2% (the pharmacopoeia standard ≥ 2.0%), the lead content detected by AAS is 0.3 mg / kg (the standard ≤ 5.0 mg / kg), and 33 kinds of banned pesticides are not detected by GC-MS, and the chemical data meet the standards.

[0045] In this embodiment, the component characteristics can be effective components (such as alkaloids, flavonoids, glycosides, etc.), heavy metal components (such as lead, mercury, cadmium, arsenic, etc.), pesticide residue components (such as various pesticides such as organophosphorus and organochlorine), or aflatoxin components, other harmful substances (such as sulfur dioxide residue), etc.

[0046] The beneficial effects of the above technologies are as follows: By collecting the physical data of the traditional Chinese medicine decoction pieces to be detected, collecting the chemical data of the traditional Chinese medicine decoction pieces to be detected, and obtaining the historical detection data of the traditional Chinese medicine decoction pieces to be detected, a multi-dimensional three-dimensional data system can be constructed to enhance the accuracy and adaptability of screening. Embodiment 4:

[0047] Based on Embodiment 3, an intelligent screening method for traditional Chinese medicine decoction pieces analyzes the physical data of the traditional Chinese medicine decoction pieces to be detected, determines the appearance feature vector, appearance eigenvalue vector, and physical feature vector of the traditional Chinese medicine decoction pieces to be detected, calculates the feature deviation value of each feature in the physical feature vector of the traditional Chinese medicine decoction pieces to be detected, and calculates the physical deviation value of the traditional Chinese medicine decoction pieces to be detected, including: Perform image preprocessing on the appearance images at each acquisition position in the appearance image data of the traditional Chinese medicine decoction pieces to be detected; Perform binarization processing on the appearance images at all acquisition positions in the appearance image data of the preprocessed traditional Chinese medicine decoction pieces to be detected and extract the contour of the traditional Chinese medicine decoction pieces to be detected. Extract features from the extracted contour of the traditional Chinese medicine decoction pieces to be detected to determine the shape feature vector and the shape eigenvalue vector of the traditional Chinese medicine decoction pieces to be detected. Among them, the shape feature vector includes multiple shape features, and the shape eigenvalue vector includes the shape eigenvalues of each shape feature in the shape feature vector; Convert the appearance images at all acquisition positions in the appearance image data of the preprocessed traditional Chinese medicine decoction pieces to be detected into the HSV color space, perform color feature extraction on the converted appearance images at all acquisition positions, and determine the color feature vector and the color eigenvalue vector of the traditional Chinese medicine decoction pieces to be detected. Among them, the color feature vector includes multiple color features, and the color eigenvalue vector includes the color eigenvalues of each color feature in the color feature vector; Perform grayscale processing on the appearance images at all acquisition positions in the appearance image data of the preprocessed traditional Chinese medicine decoction pieces to be detected, perform texture feature extraction on the grayscale processed appearance images at all acquisition positions, and determine the texture feature vector and the texture eigenvalue vector of the traditional Chinese medicine decoction pieces to be detected. Among them, the texture feature vector includes multiple texture features, and the texture eigenvalue vector includes the texture eigenvalues of each texture feature in the texture feature vector; Generate the appearance feature vector of the traditional Chinese medicine decoction pieces to be detected based on the shape feature vector, color feature vector, and texture feature vector of the traditional Chinese medicine decoction pieces to be detected. At the same time, determine the appearance eigenvalue vector of the traditional Chinese medicine decoction pieces to be detected based on the shape eigenvalue vector, color eigenvalue vector, texture eigenvalue vector, and appearance feature vector of the traditional Chinese medicine decoction pieces to be detected.

[0048] In this embodiment, an image preprocessing algorithm is used to remove noise in the appearance image, adjust parameters such as the brightness and contrast of the image, and improve the image quality. For example, the median filtering algorithm is used to remove noise, and histogram equalization is used to enhance the contrast of the image, making the details of the decoction piece image clearer.

[0049] In this embodiment, for example: taking Angelica sinensis decoction pieces as an example, when collecting the appearance image, due to uneven light, part of the image is darker and the details are blurred. The random noise in the image is removed by median filtering, and then histogram equalization is performed, making the features such as the section texture and color of the Angelica sinensis decoction pieces easier for subsequent analysis.

[0050] In this embodiment, the preprocessed appearance image is converted into a binary image, so that there is an obvious black-and-white contrast between the cut pieces of Chinese herbal medicine and the background in the image, and then the contour of the cut pieces of Chinese herbal medicine is extracted. The contour is analyzed to extract shape features such as perimeter, area, circularity, and aspect ratio, and a shape feature vector is formed; the specific numerical value corresponding to each feature is the shape feature value vector.

[0051] In this embodiment, for example: for the cut pieces of Astragalus membranaceus, after binarization, its contour is clearly presented. It is calculated that its perimeter is 12 cm, the area is 8 cm², the circularity is 0.8 (the closer to 1, the rounder), and the aspect ratio is 1.2. These features form a shape feature vector, and the corresponding numerical values form a shape feature value vector, which can be used to judge whether the cut pieces of Astragalus membranaceus meet the standard shape specifications.

[0052] In this embodiment, the appearance image is converted from the RGB color space to the HSV color space. The HSV color space is more in line with the human perception characteristics of colors and is convenient for extracting color features. By analyzing the distribution of hue (H), saturation (S), and value (V) of the pixels in the image, color features such as the main hue and color uniformity are extracted, and the color feature vector and feature value vector are determined.

[0053] In this embodiment, for example: when detecting the cut pieces of Lycium barbarum, after converting to the HSV color space, it is analyzed that its main hue is concentrated at about H = 30° (red series), the saturation is relatively high (S value is about 0.8), the value is moderate (V value is about 0.6), and the color distribution of the whole image is relatively uniform. These color features form a color feature vector, and the corresponding numerical values form a color feature value vector, which can be used to judge whether the color of the cut pieces of Lycium barbarum meets the standard.

[0054] In this embodiment, first, the appearance image is grayscale to eliminate the interference of color information, and then algorithms such as gray-level co-occurrence matrix and local binary pattern are used to extract texture features such as roughness, directionality, and periodicity, and then the texture feature vector and feature value vector are determined.

[0055] In this embodiment, for example: for the cut pieces of Poria cocos, after grayscale processing, the contrast, entropy, energy and other parameters of its texture are calculated using the gray-level co-occurrence matrix. It is found that the contrast of the texture of the cut pieces of Poria cocos is relatively high and the entropy value is relatively low, indicating that its texture is relatively regular and clear. These texture features form a texture feature vector, and the corresponding parameter numerical values form a texture feature value vector, which can be used to evaluate the texture characteristics of the cut pieces of Poria cocos.

[0056] In this embodiment, the shape feature vector, color feature vector, and texture feature vector are integrated together, and principal component analysis is performed to achieve dimensionality reduction, generating an appearance feature vector to comprehensively describe the appearance characteristics of the cut Chinese medicine pieces. At the same time, the numerical values in the shape eigenvalue vector, color eigenvalue vector, and texture eigenvalue vector corresponding to all features in the appearance feature vector are integrated to determine the appearance eigenvalue vector, providing a specific numerical basis for subsequent deviation calculation.

[0057] In this embodiment, the shape feature vector may include the contour area, perimeter, aspect ratio, circularity, etc. of the Chinese medicine cut pieces to be detected.

[0058] In this embodiment, the color feature vector may include the HSV hue distribution, main color proportion, etc. of the Chinese medicine cut pieces to be detected.

[0059] In this embodiment, the texture feature vector may include the roughness, contrast, correlation, energy, homogeneity, etc. of the Chinese medicine cut pieces to be detected.

[0060] The beneficial effects of the above technology are as follows: By analyzing the physical data of the Chinese medicine cut pieces to be detected and determining the appearance feature vector and appearance eigenvalue vector of the Chinese medicine cut pieces to be detected, an objective quantitative evaluation of the physical quality of the Chinese medicine cut pieces can be achieved, improving the screening accuracy and consistency. Embodiment 5:

[0061] Based on Embodiment 4, an intelligent screening method for Chinese medicine cut pieces analyzes the physical data of the Chinese medicine cut pieces to be detected, determines the appearance feature vector, appearance eigenvalue vector, and physical feature vector of the Chinese medicine cut pieces to be detected, calculates the feature deviation value of each feature in the physical feature vector of the Chinese medicine cut pieces to be detected, and calculates the physical deviation value of the Chinese medicine cut pieces to be detected. It further includes: Based on all texture features in the texture data of the Chinese medicine cut pieces to be monitored, all weight features in the weight data, and the appearance feature vector, determine the physical feature vector of the Chinese medicine cut pieces to be monitored; Based on the physical feature vector of the Chinese medicine cut pieces to be detected, historical detection data, and cut piece standard data, calculate the feature deviation value of each feature in the physical feature vector of the Chinese medicine cut pieces to be detected; Based on the feature deviation values of all features in the physical feature vector of the Chinese medicine cut pieces to be detected, calculate the physical deviation value of the Chinese medicine cut pieces to be detected.

[0062] In this embodiment, all texture features in the texture data of the Chinese medicine cut pieces to be monitored and all weight features in the weight data are integrated with the appearance feature vector. These physical features in different dimensions are arranged and combined in a certain order to form a vector containing various physical attribute information, that is, the physical feature vector.

[0063] In this embodiment, based on the physical feature vector of the traditional Chinese medicine decoction pieces to be monitored, historical detection data, and decoction piece standard data, the feature deviation value of each feature in the physical feature vector of the traditional Chinese medicine decoction pieces to be detected is calculated. The calculation formula for the feature deviation value can be: ; Where, represents the feature deviation value of the i-th feature in the physical feature vector, represents the first feature sub-deviation value of the i-th feature in the physical feature vector, represents the i-th feature in the physical feature vector, represents the feature value of the i-th feature in the physical feature vector, represents the feature importance factor of the i-th feature in the physical feature vector, represents the lower limit of the decoction piece standard sub-range of the parameter of the sub-standard without an existence mark corresponding to the i-th feature in the physical feature vector based on the standard parameter vector, represents the upper limit of the decoction piece standard sub-range of the parameter of the sub-standard without an existence mark corresponding to the i-th feature in the physical feature vector based on the standard parameter vector, represents the average value of the features of all historical traditional Chinese medicine decoction pieces in the historical monitoring data for the i-th feature in the physical feature vector, represents the standard deviation of the features of all historical traditional Chinese medicine decoction pieces in the historical monitoring data for the i-th feature in the physical feature vector. β represents the second standard data penalty factor, represents the lower limit of the decoction piece standard sub-range of the parameter of the sub-standard with an existence mark corresponding to the i-th feature in the physical feature vector based on the standard parameter vector, represents the upper limit of the decoction piece standard sub-range of the parameter of the sub-standard with an existence mark corresponding to the i-th feature in the physical feature vector based on the standard parameter vector, represents the lower limit of the secondary decoction piece standard range of the parameter of the sub-standard with an existence mark corresponding to the i-th feature in the physical feature vector based on the standard parameter vector, The upper limit of the secondary decoction piece standard range of the parameter of the sub-standard with an existence mark corresponding to the i-th feature in the physical feature vector based on the standard parameter vector. PV represents the standard parameter vector, represents the historical penalty factor.

[0064] In this embodiment, the feature importance factor of the i-th feature in the physical feature vector is determined according to the type of the traditional Chinese medicine decoction pieces to be detected, and the value range is 0-1. For example: for the rhizome type, the texture feature importance factor is relatively large and can take a value of 0.5, the texture importance factor is relatively small and can take a value of 0.3; for the flower and leaf type, the color feature importance factor is relatively large and can take a value of 0.6, and the morphological feature importance factor is relatively small and can take a value of 0.2.

[0065] In this embodiment, the value range of the second standard data penalty factor β is 0.3 - 0.7. For example, it can take the value of 0.6.

[0066] In this embodiment, the historical penalty factor indicates that the i-th feature in the physical feature vector does not have a corresponding cut herbal medicine standard sub-range, and it can take the value of 1.2.

[0067] In this embodiment, based on the feature deviation values of all features in the physical feature vector of the cut herbal medicine to be detected, the physical deviation value of the cut herbal medicine to be detected is calculated. The calculation formula of the physical deviation value can be: ; where DP represents the physical deviation value of the cut herbal medicine to be detected, represents the label penalty factor of the i-th feature in the physical feature vector, represents the reinforcement factor for the absence of the cut herbal medicine standard sub-range, represents the average feature deviation value of all features in the physical feature vector, N1 represents the number of features in the physical feature vector, and N2 represents the number of parameters in the standard parameter vector.

[0068] In this embodiment, the reinforcement factor for the absence of the cut herbal medicine standard sub-range has a value range of 0 - 1. For example, it can take the value of 0.3.

[0069] In this embodiment, the label penalty factor of the i-th feature in the physical feature vector has a value range of 0 - 1. For example, if there is no existence mark, it takes the value of 0.2, if there is an existence mark, it takes the value of 0.5.

[0070] The beneficial effects of the above technology are as follows: By determining the physical feature vector of the cut herbal medicine to be detected, calculating the feature deviation value of each feature in the physical feature vector of the cut herbal medicine to be detected, and calculating the physical deviation value of the cut herbal medicine to be detected, it is possible to achieve accurate and dynamic determination of the physical quality of cut herbal medicine, improve the screening accuracy and reliability, effectively reduce manual subjective errors, and promote the intelligent upgrading of cut herbal medicine screening. Embodiment 6:

[0071] On the basis of Embodiment 5, an intelligent screening method for cut herbal medicine constructs a chemical detection model based on the cut herbal medicine chemical data, historical detection data, and cut herbal medicine standard data of the cut herbal medicine to be detected, and determines the chemical deviation data of the cut herbal medicine to be detected based on the cut herbal medicine chemical data and the chemical detection model, including: Construct a cut herbal medicine chemical detection model based on the cut herbal medicine standard data, the historical chemical detection vectors and historical chemical detection grades of all historical cut herbal medicines in the historical detection data; Input the chemical data of the Chinese herbal pieces to be detected into the constructed chemical detection model of Chinese herbal pieces, and determine the chemical deviation data and chemical detection grade of the Chinese herbal pieces to be detected based on the output result of the chemical detection model of Chinese herbal pieces. Among them, the chemical deviation data includes multiple deviated component features and the deviation values of each deviated component feature, and the chemical detection grade includes excellent, good, medium, and poor.

[0072] In this embodiment, a large amount of historical chemical detection data of Chinese herbal pieces is collected, including the historical chemical detection vectors (including multiple component features and corresponding feature values) and historical chemical detection grades (excellent, good, medium, poor) of each historical Chinese herbal piece.

[0073] In this embodiment, the historical chemical detection vectors and historical chemical detection grades are used as training data, and machine learning algorithms (such as support vector machine, random forest algorithm, or neural network algorithm) are used to establish the mapping relationship between data features and quality grades. During the training process, the model is optimized and adjusted in combination with the standard data of Chinese herbal pieces, so that the model can accurately identify the quality grades corresponding to different chemical detection data, thereby constructing a chemical detection model of Chinese herbal pieces.

[0074] In this embodiment, an excellent chemical detection grade indicates that the content of active ingredients is high and the content of other ingredients is low, meeting the high-quality standards.

[0075] In this embodiment, a good chemical detection grade indicates that the content of active ingredients meets the standards, and the content of other ingredients is slightly higher but does not affect safety.

[0076] In this embodiment, a medium chemical detection grade indicates that the content of active ingredients is close to the lower limit of the standard, and the content of other ingredients is relatively high but still within the allowable range.

[0077] In this embodiment, a poor chemical detection grade indicates that the content of active ingredients is low and the content of other ingredients is high, not meeting the standards.

[0078] The beneficial effects of the above technology are as follows: A chemical detection model is constructed based on the chemical data of Chinese herbal pieces to be detected, historical detection data, and standard data of Chinese herbal pieces. Based on the chemical data of Chinese herbal pieces and the chemical detection model, the chemical deviation data of Chinese herbal pieces to be detected is determined, which can realize the intelligent evaluation of the chemical quality of Chinese herbal pieces, improve the efficiency and accuracy of the chemical quality detection of Chinese herbal pieces, and provide reliable technical support for the quality control of traditional Chinese medicine. Example 7:

[0079] On the basis of Example 6, an intelligent screening method for Chinese herbal pieces determines the grade of Chinese herbal pieces to be detected based on the physical deviation value and chemical deviation data of Chinese herbal pieces to be detected, including: Determine the physical detection level of the Chinese herbal medicine pieces to be detected based on the physical deviation values of the Chinese herbal medicine pieces to be detected; Determine the grade of the Chinese herbal medicine pieces to be detected based on the physical detection level of the Chinese herbal medicine pieces to be detected and the chemical detection level in the chemical deviation data.

[0080] In this embodiment, the corresponding relationship between the physical deviation value and the physical detection level is preset, and the calculated physical deviation value of the Chinese herbal medicine pieces to be detected is compared with these preset ranges to determine its physical detection level.

[0081] In this embodiment, a comprehensive evaluation rule is formulated, and the physical detection level and the chemical detection level are combined and analyzed to determine the final grade of the Chinese herbal medicine pieces. For example, when the physical detection level is "excellent" and the chemical detection level is "excellent", the grade of the Chinese herbal medicine pieces is determined as "special grade"; when the physical detection level is "good" and the chemical detection level is "good", the grade of the Chinese herbal medicine pieces is determined as "first grade"; when one of the physical detection level and the chemical detection level is "poor", the grade of the Chinese herbal medicine pieces is determined as "unqualified", etc. According to the actual physical detection level and chemical detection level of the Chinese herbal medicine pieces to be detected, the final grade of the Chinese herbal medicine pieces is obtained based on the evaluation rule.

[0082] The beneficial effects of the above technology are as follows: Based on the physical deviation values and chemical deviation data of the Chinese herbal medicine pieces to be detected, the grade of the Chinese herbal medicine pieces to be detected is determined, which can break through the limitation of single-dimensional evaluation, construct a multi-dimensional and systematic quality grading system, realize accurate and comprehensive evaluation of the quality of Chinese herbal medicine pieces, effectively improve the scientificity and reliability of screening, and provide a more reasonable and efficient technical means for the quality control of traditional Chinese medicine. Example 8:

[0083] The present invention provides an intelligent screening system for Chinese herbal medicine pieces, which is used to execute the intelligent screening method of any one of the Chinese herbal medicine pieces in Embodiments 1 to 7, refer to Figure 2 , including: Acquisition module: Obtain the standard data of the Chinese herbal medicine pieces to be detected, collect the physical data of the Chinese herbal medicine pieces to be detected, collect the chemical data of the Chinese herbal medicine pieces to be detected, and obtain the historical detection data of the Chinese herbal medicine pieces to be detected; Physical module: Analyze the physical data of the Chinese herbal medicine pieces to be detected, determine the appearance feature vector, appearance eigenvalue vector and physical feature vector of the Chinese herbal medicine pieces to be detected, calculate the feature deviation value of each feature in the physical feature vector of the Chinese herbal medicine pieces to be detected, and calculate the physical deviation value of the Chinese herbal medicine pieces to be detected; Chemical module: Construct a chemical detection model based on the chemical data, historical detection data and standard data of the Chinese herbal medicine pieces to be detected, and determine the chemical deviation data of the Chinese herbal medicine pieces to be detected based on the chemical data and the chemical detection model; Screening module: Based on the physical deviation value and chemical deviation data of the Chinese herbal medicine pieces to be detected, determine the grade of the Chinese herbal medicine pieces to be detected, and perform intelligent screening based on the grade of the Chinese herbal medicine pieces to be detected.

[0084] The beneficial effects of the above technology are as follows: By analyzing the obtained standard data of the Chinese herbal medicine pieces to be detected, historical detection data, as well as the physical data and chemical data of the Chinese herbal medicine pieces to be detected collected, calculate the physical deviation value of the Chinese herbal medicine pieces to be detected, determine the chemical deviation data of the Chinese herbal medicine pieces to be detected, and determine the grade of the Chinese herbal medicine pieces to be detected to achieve intelligent screening of the Chinese herbal medicine pieces to be detected. A comprehensive and accurate standard system for herbal medicine pieces can be constructed to realize multi-dimensional and quantitative evaluation of the quality of herbal medicine pieces, comprehensively and accurately judge the grade of herbal medicine pieces, avoid subjective errors and quality misjudgments, improve the screening efficiency and accuracy, and provide a scientific solution for the quality control of traditional Chinese medicine.

[0085] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention also intends to include these changes and modifications.

Claims

1. An intelligent screening method for traditional Chinese medicine decoction pieces, characterized in that, Including: S1: Obtain the standard data of the Chinese herbal pieces to be detected, collect the physical data of the Chinese herbal pieces to be detected, collect the chemical data of the Chinese herbal pieces to be detected, and obtain the historical detection data of the Chinese herbal pieces to be detected; S2: Analyze the physical data of the Chinese herbal pieces to be detected, determine the appearance feature vector, appearance eigenvalue vector and physical feature vector of the Chinese herbal pieces to be detected, calculate the feature deviation value of each feature in the physical feature vector of the Chinese herbal pieces to be detected, and calculate the physical deviation value of the Chinese herbal pieces to be detected; S3: Construct a chemical detection model based on the chemical data, historical detection data and standard data of the Chinese herbal pieces to be detected, and determine the chemical deviation data of the Chinese herbal pieces to be detected based on the chemical data of the Chinese herbal pieces and the chemical detection model; S4: Determine the grade of the Chinese herbal pieces to be detected based on the physical deviation value and chemical deviation data of the Chinese herbal pieces to be detected, and perform intelligent screening based on the grade of the Chinese herbal pieces to be detected.

2. The intelligent screening method for a traditional Chinese medicine decoction piece according to claim 1, wherein, Obtain the standard data of the Chinese herbal pieces to be detected, including: Extract the first standard data of the Chinese herbal pieces to be detected from the pharmacopoeia, and extract the second standard data of the Chinese herbal pieces to be detected from the industry standard database. Among them, the first standard data includes the first standard sub-data of multiple sub-standards, and the second standard data includes the second standard sub-data of multiple sub-standards; Traverse each sub-standard in the second standard data. If the first standard data includes the first standard sub-data of the sub-standard currently traversed in the second standard data, determine the standard sub-range of each parameter of each sub-standard currently traversed in the second standard data based on the first standard sub-data, and mark the existence of the sub-standard currently traversed in the second standard data; For each sub-standard in the second standard data with an existence mark, determine the standard sub-range of each parameter of each sub-standard with an existence mark based on the second standard sub-data of each sub-standard in the second standard data; For each sub-standard in the second standard data without an existence mark, determine the standard sub-range of each parameter of each sub-standard without an existence mark based on the second standard sub-data of each sub-standard in the second standard data; For each sub-standard in the first standard data without an existence mark, determine the standard sub-range of each parameter of each sub-standard without an existence mark based on the first standard sub-data of each sub-standard in the first standard data; Based on all the parameters of all the sub-standards with existence marks and all the parameters of all the sub-standards without existence marks, determine the standard parameter vector; Based on the standard parameter vector, the standard sub-range of all the parameters of all the sub-standards with existence marks, the standard sub-range and the standard sub-range of all the parameters of all the sub-standards without existence marks, determine the standard data of the Chinese herbal pieces to be detected.

3. The intelligent screening method of a traditional Chinese medicine decoction piece according to claim 1, characterized in that, Collect the physical data of the Chinese herbal pieces to be detected, collect the chemical data of the Chinese herbal pieces to be detected, and obtain the historical detection data of the Chinese herbal pieces to be detected, including: Collect the appearance image data of the Chinese herbal pieces to be detected based on the image acquisition device. Among them, the appearance image data includes the appearance images of multiple acquisition positions of the Chinese herbal pieces to be detected; Collect the texture data of the Chinese herbal pieces to be detected based on a texture detection device, where the texture data includes multiple texture features of the Chinese herbal pieces to be detected and the texture feature values of each texture feature; Collect the weight data of the Chinese herbal pieces to be detected based on a weight detection device, where the weight data includes multiple weight features of the Chinese herbal pieces to be detected and the weight feature values of each weight feature; Determine the physical data of the Chinese herbal pieces to be detected based on the appearance image data, texture data, and weight data of the Chinese herbal pieces to be detected; Collect the chemical data of the Chinese herbal pieces to be detected based on a chemical detection device, where the chemical data of the herbal pieces includes multiple component features of the Chinese herbal pieces to be detected and the component feature values of each component feature; Obtain the historical detection data of the Chinese herbal pieces to be detected, where the historical detection data includes the historical physical detection vectors, historical chemical detection vectors, and historical chemical detection grades of multiple historical Chinese herbal pieces.

4. The intelligent screening method of a traditional Chinese medicine decoction piece according to claim 3, characterized in that, Analyze the physical data of the Chinese herbal pieces to be detected, determine the appearance feature vector, appearance feature value vector, and physical feature vector of the Chinese herbal pieces to be detected, calculate the feature deviation value of each feature in the physical feature vector of the Chinese herbal pieces to be detected, and calculate the physical deviation value of the Chinese herbal pieces to be detected, including: Perform image preprocessing on the appearance images at each acquisition position in the appearance image data of the Chinese herbal pieces to be detected; Perform binarization processing on the appearance images at all acquisition positions in the preprocessed appearance image data of the Chinese herbal pieces to be detected and extract the contour of the Chinese herbal pieces to be detected, extract features from the extracted contour of the Chinese herbal pieces to be detected, and determine the shape feature vector and shape feature value vector of the Chinese herbal pieces to be detected, where the shape feature vector includes multiple shape features, and the shape feature value vector includes the shape feature values of each shape feature in the shape feature vector; Convert the appearance images at all acquisition positions in the preprocessed appearance image data of the Chinese herbal pieces to be detected to the HSV color space, extract color features from the converted appearance images at all acquisition positions, and determine the color feature vector and color feature value vector of the Chinese herbal pieces to be detected, where the color feature vector includes multiple color features, and the color feature value vector includes the color feature values of each color feature in the color feature vector; Perform grayscale processing on the appearance images at all acquisition positions in the preprocessed appearance image data of the Chinese herbal pieces to be detected, extract texture features from the grayscale processed appearance images at all acquisition positions, and determine the texture feature vector and texture feature value vector of the Chinese herbal pieces to be detected, where the texture feature vector includes multiple texture features, and the texture feature value vector includes the texture feature values of each texture feature in the texture feature vector; Generate the appearance feature vector of the Chinese herbal pieces to be detected based on the shape feature vector, color feature vector, and texture feature vector of the Chinese herbal pieces to be detected. At the same time, determine the appearance feature value vector of the Chinese herbal pieces to be detected based on the shape feature value vector, color feature value vector, texture feature value vector, and appearance feature vector of the Chinese herbal pieces to be detected.

5. The intelligent screening method of a traditional Chinese medicine decoction piece according to claim 4, wherein, Analyze the physical data of the traditional Chinese medicine decoction pieces to be detected, determine the appearance feature vector, appearance eigenvalue vector and physical feature vector of the traditional Chinese medicine decoction pieces to be detected, calculate the feature deviation value of each feature in the physical feature vector of the traditional Chinese medicine decoction pieces to be detected, and calculate the physical deviation value of the traditional Chinese medicine decoction pieces to be detected. It also includes: Based on all texture features in the texture data of the traditional Chinese medicine decoction pieces to be monitored, all weight features in the weight data, and the appearance feature vector, determine the physical feature vector of the traditional Chinese medicine decoction pieces to be monitored; Based on the physical feature vector of the traditional Chinese medicine decoction pieces to be monitored, historical detection data, and decoction piece standard data, calculate the feature deviation value of each feature in the physical feature vector of the traditional Chinese medicine decoction pieces to be detected; Based on the feature deviation values of all features in the physical feature vector of the traditional Chinese medicine decoction pieces to be detected, calculate the physical deviation value of the traditional Chinese medicine decoction pieces to be detected.

6. The intelligent screening method for a traditional Chinese medicine decoction piece according to claim 5, characterized in that Construct a chemical detection model based on the chemical data of the traditional Chinese medicine decoction pieces to be detected, historical detection data, and decoction piece standard data. Based on the chemical data of the decoction pieces and the chemical detection model, determine the chemical deviation data of the traditional Chinese medicine decoction pieces to be detected, including: Based on the decoction piece standard data, historical chemical detection vectors and historical chemical detection grades of all historical traditional Chinese medicine decoction pieces in the historical detection data, construct a chemical detection model for decoction pieces; Input the chemical data of the traditional Chinese medicine decoction pieces to be detected into the constructed chemical detection model for decoction pieces. Based on the output results of the chemical detection model for decoction pieces, determine the chemical deviation data and chemical detection grade of the traditional Chinese medicine decoction pieces to be detected. Among them, the chemical deviation data includes multiple component features with deviations and the deviation values of each component feature with deviations. The chemical detection grade includes excellent, good, medium, and poor.

7. The intelligent screening method of a traditional Chinese medicine decoction piece according to claim 6, characterized in that, Based on the physical deviation value and chemical deviation data of the traditional Chinese medicine decoction pieces to be detected, determine the grade of the traditional Chinese medicine decoction pieces to be detected, including: Based on the physical deviation value of the traditional Chinese medicine decoction pieces to be detected, determine the physical detection grade of the traditional Chinese medicine decoction pieces to be detected; Based on the physical detection grade of the traditional Chinese medicine decoction pieces to be detected and the chemical detection grade in the chemical deviation data, determine the grade of the traditional Chinese medicine decoction pieces to be detected.

8. An intelligent screening system for traditional Chinese medicine decoction pieces, characterized in that, For implementing the intelligent screening method of traditional Chinese medicine decoction pieces described in any one of claims 1 to 7, it includes: Collection module: Obtain the decoction piece standard data of the traditional Chinese medicine decoction pieces to be detected, collect the physical data of the traditional Chinese medicine decoction pieces to be detected, collect the chemical data of the traditional Chinese medicine decoction pieces to be detected, and obtain the historical detection data of the traditional Chinese medicine decoction pieces to be detected; Physical module: Analyze the physical data of the traditional Chinese medicine decoction pieces to be detected, determine the appearance feature vector, appearance eigenvalue vector and physical feature vector of the traditional Chinese medicine decoction pieces to be detected, calculate the feature deviation value of each feature in the physical feature vector of the traditional Chinese medicine decoction pieces to be detected, and calculate the physical deviation value of the traditional Chinese medicine decoction pieces to be detected; Chemical module: Construct a chemical detection model based on the chemical data of the traditional Chinese medicine decoction pieces to be detected, historical detection data, and decoction piece standard data. Based on the chemical data of the decoction pieces and the chemical detection model, determine the chemical deviation data of the traditional Chinese medicine decoction pieces to be detected; Screening module: Based on the physical deviation value and chemical deviation data of the Chinese herbal medicine pieces to be detected, determine the grade of the Chinese herbal medicine pieces to be detected, and perform intelligent screening based on the grade of the Chinese herbal medicine pieces to be detected.

Images