Traditional Chinese medicine analysis and identification system based on large model

Abstract

The invention discloses a traditional Chinese medicine analysis and identification system based on a large model, and relates to the field of traditional Chinese medicine identification. Comprising a sensor module used for acquiring multi-dimensional data of traditional Chinese medicine; the large model database is used for storing the information of the traditional Chinese medicinal materials by adopting a block type storage structure; the multi-dimensional data comparison module performs data similarity comparison of each dimension according to the multi-dimensional data acquired by the sensor module and the large model database to generate a comparison result; and the result output module is used for outputting a fusion result of the multi-dimensional data comparison module, and the sensor module comprises a spectral data acquisition unit, a microscopic imaging unit and a smell sensing unit. The traditional Chinese medicine information is collected through large model data, and the traditional Chinese medicine is identified by adopting multi-dimensional data fusion of the sensor module and comparison of subsequent fused data, so that the identification accuracy is improved on the basis of improving the identification efficiency.

Description

Technical Field

[0001] The present invention relates to the field of traditional Chinese medicine, and particularly to a system for analyzing and identifying traditional Chinese medicine drugs based on a large model. Background Art

[0002] There are many types of traditional Chinese medicine, and it is difficult and inaccurate to identify them only by appearance. Therefore, in the prior art, there are the following identification methods:

[0003] I. Chemical analysis method

[0004] To overcome the limitations of the sensory identification method, the chemical analysis method has gradually been applied to the identification of traditional Chinese medicine. This method extracts the chemical components in the medicinal materials and conducts qualitative or quantitative analysis to judge the authenticity and quality of the medicinal materials. However, the chemical analysis method usually requires destroying the integrity of the medicinal materials, and the detection process is cumbersome and time-consuming, which is not conducive to the subsequent use of the medicinal materials. In addition, the chemical components of medicinal materials from different origins and different growth environments may vary, which also increases the difficulty of chemical analysis.

[0005] II. Simple instrument detection method

[0006] With the development of instrument analysis technology, some simple instrument detection methods have also been applied to the identification of traditional Chinese medicine. For example, a spectrometer is used to detect the spectral characteristics of the medicinal materials, or a microscope is used to observe the microscopic structure of the medicinal materials. These methods have improved the accuracy and objectivity of identification to a certain extent.

[0007] However, due to the complexity of the components of traditional Chinese medicine, single-dimensional detection often fails to comprehensively reflect the characteristics of the medicinal materials, so the accuracy of the identification results still needs to be improved.

[0008] Therefore, the present invention proposes a system for analyzing and identifying traditional Chinese medicine drugs based on a large model. Summary of the Invention

[0009] The purpose of the present invention is to solve the deficiencies existing in the prior art, and to propose a system for analyzing and identifying traditional Chinese medicine drugs based on a large model.

[0010] To achieve the above purpose, the present invention adopts the following technical solutions:

[0011] A system for analyzing and identifying traditional Chinese medicine drugs based on a large model, comprising:

[0012] A sensor module for acquiring multi-dimensional data of traditional Chinese medicine;

[0013] A large model database storing information on traditional Chinese medicine herbs using a block storage structure;

[0014] A multi-dimensional data comparison module that compares the multi-dimensional data obtained by the sensor module with the large model database for data similarity in each dimension to generate a comparison result;

[0015] A result output module that outputs the fusion result of the multi-dimensional data comparison module.

[0016] Preferably, the sensor module includes a spectral data acquisition unit, a microscopic imaging unit, and an odor sensing unit.

[0017] Preferably, the spectral data acquisition unit includes a near-infrared spectrometer and a Raman spectrometer, and the spectral data acquisition unit processes the spectral waveform using an improved 1D-CNN network;

[0018] The microscopic imaging unit is a 40-1000 times digital microscope, and the microscopic imaging unit is based on microscopic image analysis of Vision Transformer;

[0019] The odor sensing unit is composed of multiple arrayed electronic noses.

[0020] Preferably, the method for establishing the large model database includes the following steps:

[0021] S21: Obtain the spectral information, microscopic image information, and odor information of each known traditional Chinese medicine from existing literature through a big data model;

[0022] S22: Establish a comparison model and a verification model based on the obtained information.

[0023] Preferably, in the S21 step, the existing literature includes "Chinese Pharmacopoeia", a formula database, a clinical research literature database, and "Compendium of Chinese Materia Medica".

[0024] Preferably, in the S22 step, the establishment of the comparison model includes the following steps:

[0025] S221A: Establish m + n + p comparison sets according to the types of information obtained, where m, n, and p are the number of spectral features of spectral information, the number of image information of image information, and the number of odor features of odor information respectively;

[0026] S222A: For each comparison set, match all the traditional Chinese medicine names containing the features of this set.

[0027] Preferably, in the S22 step, the establishment of the verification model includes the following steps:

[0028] S221B: Establish M traditional Chinese medicine feature sets according to the number of types of traditional Chinese medicine obtained;

[0029] S222B: For each traditional Chinese medicine feature set, name it with the name of the traditional Chinese medicine and match all the features of this traditional Chinese medicine.

[0030] Preferably, the comparison method of the multi-dimensional data comparison module includes the following steps:

[0031] S31A: Obtain the data of one dimension collected by the sensor module, and this dimension of data is a feature.

[0032] S32A: Search and match this feature with the m + n + p sets of the comparison model in the large model database. After successful matching, mark this set as the selected set.

[0033] S33A: Repeat steps S31A and S32A until after all dimensions of data are processed, multiple selected sets are obtained.

[0034] S34A: Then take the intersection of all selected sets. If the traditional Chinese medicine name in the intersection is unique, there is no need for a verification link and it is directly output. If the traditional Chinese medicine names in the intersection are not unique, verification is performed.

[0035] Preferably, in step S34A, if the intersection is zero, it means that the verified traditional Chinese medicine is not included, and the large model database is updated.

[0036] Preferably, the verification method of the multi-dimensional data comparison module includes the following steps:

[0037] S31B: Select one of the traditional Chinese medicine names in the intersection and match the selected traditional Chinese medicine name in the M traditional Chinese medicine feature sets in the verification model.

[0038] S32B: Obtain the total number of feature data MT in the set and the number of feature data Mt that matches the data collection module, and calculate the matching degree.

[0039] S33B: Output the result with the highest matching degree.

[0040] The beneficial effects of the present invention are as follows:

[0041] 1. The present invention collects traditional Chinese medicine information through the large model data, and uses the multi-dimensional data fusion of the sensor module and the comparison of the subsequent fusion data to identify traditional Chinese medicine, thereby increasing the identification efficiency and also increasing the identification accuracy.

[0042] 2. The present invention, based on the multi-dimensional data comparison, adopts the dual identification of the comparison model and the verification model, thereby further reducing the possibility of misjudgment and increasing the accuracy of the final result output. Description of the Drawings

[0043] Figure 1The architecture and logic diagram of a traditional Chinese medicine drug analysis and identification system based on a large model proposed by the present invention. Specific implementation manners

[0044] The technical solution of the present invention will be further described in detail below in conjunction with specific implementation manners.

[0045] In the description of the present invention, it should be noted that unless otherwise clearly specified and limited, the terms "installation", "connection", "connection", and "setting" should be understood in a broad sense. For example, it can be fixedly connected and set, or detachably connected and set, or integrally connected and set. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0046] Example 1:

[0047] A traditional Chinese medicine drug analysis and identification system based on a large model, which includes:

[0048] A sensor module for obtaining multi-dimensional data of traditional Chinese medicine;

[0049] A large model database that stores information on traditional Chinese medicine materials using a block storage structure;

[0050] A multi-dimensional data comparison module that compares the multi-dimensional data obtained by the sensor module with the large model database for data similarity in each dimension and generates a comparison result;

[0051] A result output module that outputs the fusion result of the multi-dimensional data comparison module.

[0052] The sensor module includes a spectral data acquisition unit, a microscopic imaging unit, and an odor sensing unit;

[0053] The spectral data acquisition unit includes a near-infrared spectrometer and a Raman spectrometer, and the spectral data acquisition unit processes the spectral waveform using an improved 1D-CNN network;

[0054] The microscopic imaging unit is a 40-1000-fold digital microscope, and the microscopic imaging unit is based on microscopic image analysis of Vision Transformer;

[0055] The odor sensing unit consists of multiple electronically noses arranged in an array.

[0056] Example 2:

[0057] A traditional Chinese medicine drug analysis and identification system based on a large model, which includes:

[0058] A sensor module for obtaining multi-dimensional data of traditional Chinese medicine;

[0059] A large model database that stores information on traditional Chinese medicine herbs using a block storage structure;

[0060] A multi-dimensional data comparison module that compares the multi-dimensional data obtained by the sensor module with the large model database for data similarity in each dimension to generate a comparison result;

[0061] A result output module that outputs the fusion result of the multi-dimensional data comparison module.

[0062] The sensor module includes a spectral data acquisition unit, a microscopic imaging unit, and an odor sensing unit;

[0063] The spectral data acquisition unit includes a near-infrared spectrometer and a Raman spectrometer, and the spectral data acquisition unit processes the spectral waveform using an improved 1D-CNN network;

[0064] The microscopic imaging unit is a 40-1000 times digital microscope, and the microscopic imaging unit is based on microscopic image analysis of Vision Transformer;

[0065] The odor sensing unit is composed of multiple electronically noses arranged in an array.

[0066] The method for establishing the large model database includes the following steps:

[0067] S21: Obtain the spectral information, microscopic image information, and odor information of each known traditional Chinese medicine from existing literature through a big data model;

[0068] S22: Establish a comparison model and a verification model based on the obtained information.

[0069] In the S21 step, the existing literature includes "Chinese Pharmacopoeia", prescription database, clinical research literature database, and "Compendium of Chinese Materia Medica".

[0070] In the S22 step, the establishment of the comparison model includes the following steps:

[0071] S221A: Establish m + n + p comparison sets according to the types of information obtained, where m, n, and p are the number of spectral features of spectral information, the number of image information of image information, and the number of odor features of odor information respectively;

[0072] S222A: For each comparison set, match all the Chinese medicine names containing the features of that set.

[0073] In the S22 step, the establishment of the verification model includes the following steps:

[0074] S221B: Establish M Chinese medicine feature sets according to the number of types of Chinese medicine obtained;

[0075] S222B: For each set of traditional Chinese medicine features, name it with the name of the medicine and match all the features of this traditional Chinese medicine.

[0076] The comparison method of the multi-dimensional data comparison module includes the following steps:

[0077] S31A: Obtain the data of one dimension collected by the sensor module, and this dimension of data is a feature.

[0078] S32A: Search and match this feature with the m + n + p sets of the comparison model in the large model database. After successful matching, mark this set as the selected set.

[0079] S33A: Repeat steps S31A and S32A until after all dimensions of data are processed, multiple selected sets are obtained.

[0080] S34A: Then take the intersection of all selected sets. If the traditional Chinese medicine name in the intersection is unique, there is no need for a verification link and it is directly output. If the traditional Chinese medicine names in the intersection are not unique, verification is performed.

[0081] In step S34A, if the intersection is zero, it means that the verified traditional Chinese medicine is not included, and the large model database is updated.

[0082] The verification method of the multi-dimensional data comparison module includes the following steps:

[0083] S31B: Select one of the traditional Chinese medicine names in the intersection and match the selected traditional Chinese medicine name in the M sets of traditional Chinese medicine features in the verification model.

[0084] S32B: Obtain the total number of feature data MT in the set and the number of feature data Mt that matches the data collection module, and calculate the matching degree

[0085] S33B: Output the result with the highest matching degree.

[0086] In the present invention, traditional Chinese medicine information is collected through the large model data, and multi-dimensional data fusion of the sensor module and subsequent comparison of the fused data are used to identify traditional Chinese medicine, thereby increasing the identification efficiency and also increasing the accuracy of identification.

[0087] In the present invention, based on multi-dimensional data comparison, double identification of the comparison model and the verification model is adopted, thereby further reducing the possibility of misjudgment and increasing the accuracy of the final result output.

[0088] As described above, it is only the preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution of the present invention and its inventive concept, makes equivalent substitutions or changes, and should be covered within the protection scope of the present invention.

Claims

1. A Chinese medicine drug analysis and identification system based on a large model, characterized in that: include: A sensor module for acquiring multi-dimensional data of traditional Chinese medicine; A large model database that uses a block storage structure to store information on Chinese medicinal materials; A multi-dimensional data comparison module compares the data similarity of each dimension with the large model database based on the multi-dimensional data acquired by the sensor module, and generates a comparison result; The result output module outputs the fusion results of the multi-dimensional data comparison module.

2. A Chinese medicine drug analysis and identification system based on a large model according to claim 1, characterized in that: The sensor module includes a spectral data acquisition unit, a microscopic imaging unit and an odor sensing unit.

3. A Chinese medicine drug analysis and identification system based on a large model according to claim 2, characterized in that: The spectral data acquisition unit includes a near-infrared spectrometer and a Raman spectrometer, and the spectral data acquisition unit uses an improved 1D-CNN network to process the spectral waveform; The microscopic imaging unit is a 40-1000 times digital microscope, and the microscopic imaging unit is based on the microscopic image analysis of VisionTransformer; The odor sensing unit is composed of a plurality of electronic noses arranged in an array.

4. A Chinese medicine drug analysis and identification system based on a large model according to claim 1, characterized in that: The method for establishing the large model database comprises the following steps: S21: Obtain the spectral information, microscopic image information, and odor information of each known Chinese medicine in the existing literature through the big data model; S22: Establish a comparison model and a verification model based on the acquired information.

5. A Chinese medicine drug analysis and identification system based on a large model according to claim 4, characterized in that: In the step S21, the existing literature includes the Chinese Pharmacopoeia, the prescription database, the clinical research literature library and the Chinese Materia Medica.

6. A Chinese medicine drug analysis and identification system based on a large model according to claim 4, characterized in that: The step S22, establishing the comparison model, comprises the following steps: S221A: establishing m+n+p comparison sets according to the types of information obtained, where m, n, and p are the number of spectral features of the spectral information, the number of image information of the image information, and the number of odor features of the odor information, respectively; S222A: For each comparison set, all TCM names containing the features of the set are matched.

7. A large model-based Chinese medicine analysis and identification system according to claim 6, characterized in that: In the step S22, the establishment of the verification model includes the following steps: S221B: Establish M traditional Chinese medicine feature sets according to the number of traditional Chinese medicine types obtained; S222B: For each TCM feature set, name it using the medicine name and match all the features of the TCM.

8. A large model-based Chinese medicine analysis and identification system according to claim 7, characterized in that: The comparison method of the multidimensional data comparison module comprises the following steps: S31A: Acquire data of one dimension collected by the sensor module, where the data of the dimension is a feature; S32A: searching and matching the feature with the m+n+p sets of comparison models in the large model database, and after a successful match, marking the set as a selected set; S33A: repeating steps S31A and S32A until multiple selected sets are obtained after data of all dimensions are processed; S34A: Then take the intersection of all selected sets. If the name of the Chinese medicine in the intersection is unique, there is no need for verification and it is directly output. If the name of the Chinese medicine in the intersection is not unique, verification is performed.

9. A large model-based Chinese medicine analysis and identification system according to claim 8, characterized in that: In the step S34A, if the intersection is zero, it means that the verified Chinese medicine is not included and the large model database is updated.

10. A large model-based Chinese medicine analysis and identification system according to claim 8, characterized in that: The verification method of the multidimensional data comparison module comprises the following steps: S31B: Select one of the Chinese medicine names in the intersection, and match the selected Chinese medicine name in the M Chinese medicine feature sets in the validation model; S32B: Obtain the number of all feature data in the set MT and the number of feature data matching the data acquisition module Mt, and calculate the matching degree S33B: Output the result with the highest matching degree.

Images