Artificial intelligence precious fine medicine microscopic attribute detection method and device and electronic equipment

Abstract

The invention provides an artificial intelligence precious fine medicine microscopic attribute detection method and device and electronic equipment, and aims to improve the quality control level of traditional Chinese medicine large honeyed pill precious fine medicine raw powder. The integrity and definition of microscopic images are ensured by establishing a microscopic imaging system, optimizing imaging parameters of a microscope, preferably selecting a precious and fine medicine dispersion medium, a dispersion mode and time and constructing a microscopic image acquisition method. Image processing technologies of self-adaptive threshold segmentation, binarization processing, opening operation and hole filling are adopted to realize identification and objective quantification of microscopic characteristics of precious and fine drugs. Through correlation analysis, grey correlation analysis and principal component analysis, microscopic image key quality attributes of precious and fine drugs are screened out, the normal distribution condition of the microscopic image key quality attributes is verified through a histogram statistics method, and volatility of the key quality attributes is represented. Technical support is provided for optimization and quality control of a traditional Chinese medicine preparation process by monitoring identified microscopic key quality attributes.

Description

Technical Field

[0001] The present invention relates to the technical field of traditional Chinese medicine detection, and in particular to an artificial intelligence method, device and electronic equipment for detecting microscopic properties of precious and fine medicines. Background Art

[0002] Precious and fine medicinal materials play an important role in traditional Chinese medicine preparations, and their quality directly affects the efficacy and safety of traditional Chinese medicine preparations. Precious and fine medicinal materials are generally used as fine powders, which are difficult to distinguish with the naked eye. In addition, due to the scarcity of raw materials and the dispersion of characteristics, their microstructure is difficult to detect. The morphology, structure and powder properties of precious and fine medicinal powders have a significant impact on the preparation, mixing, and molding processes in the production process, making it difficult to ensure the quality of the finished product and clinical efficacy.

[0003] Optical microscopy is an important means of analyzing the microstructure of precious fine drug powders, and can provide detailed information on the fine structure and quality characteristics of powders. However, traditional microscopic detection methods often rely on manual operations, which have problems such as strong subjectivity, low efficiency, and poor repeatability, making it difficult to meet the high standards of quality control in modern Chinese medicine production.

[0004] With the development of artificial intelligence technology, microscopic image analysis methods based on artificial intelligence have gradually received attention. Artificial intelligence technology can realize the automated processing and analysis of microscopic images, improve detection efficiency and accuracy, and thus provide strong support for the quality control of precious and fine medicinal materials. In the process of quality evaluation of traditional Chinese medicine manufacturing, it can effectively reduce the information loss caused by human subjective judgment. Through comprehensive image data preprocessing, feature extraction, pattern recognition and other machine learning algorithms, it is possible to scientifically and objectively characterize the microscopic structural characteristic information of traditional Chinese medicine. Summary of the invention

[0005] In order to overcome the defects and deficiencies in the prior art, the artificial intelligence precious fine medicine microscopic property detection method, device and electronic equipment of the present invention are intended to provide a new way to improve the objectivity, accuracy and reliability of precious fine medicine quality control. Through precise microscopic image analysis, comprehensive detection and analysis of the microscopic properties of precious fine medicine powders can be achieved, thereby improving the quality control level of traditional Chinese medicine preparations and ensuring their efficacy and safety.

[0006] In particular, the detection method of the present invention focuses on summarizing and generalizing the quality control method of carrier musk (artificial musk) by exemplary research, so as to provide a reference for the quality control of precious fine medicines.

[0007] In order to achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0008] In a first aspect, the present invention provides an artificial intelligence method for detecting microscopic properties of precious fine medicines, the detection method comprising the following steps:

[0009] Step 1: Establish a microscopic imaging system and optimize the imaging parameters of the microscope;

[0010] Step 2: Optimize the dispersion medium, dispersion method and dispersion time of the precious fine drug, and construct a microscopic image acquisition method;

[0011] Step 3: Collect the microscopic image of precious medicine and pre-process the microscopic image;

[0012] Step 4: Extract the region of interest (ROI) of the microscopic image, perform binarization on the ROI, and further use image processing methods to remove noise and improve particle contours;

[0013] Step 5: Statistical analysis of the microscopic images is performed to quantify the microscopic characteristics of the precious drug and further identify the key quality attributes of the microscopic images;

[0014] Step 6: Characterize the variability of critical quality attributes of microscopic images.

[0015] As an optional manner, in the above detection method, in step 3, image preprocessing includes grayscale, geometric transformation, image enhancement and other methods.

[0016] As an optional manner, in the above detection method, in step 4, ROI extraction is achieved through adaptive threshold segmentation, and the image processing method includes methods such as opening operation and hole filling, which are used to remove noise, improve particle contour missing, and smooth particle boundaries.

[0017] As an optional manner, in the above-mentioned detection method, in step 5, the statistical analysis of the microscopic image includes quantification of morphological parameters such as aspect ratio, minimum diameter, fractal dimension and roundness of the microscopic particles, and the method for identifying key quality attributes of the microscopic image includes correlation analysis, grey correlation analysis and principal component analysis.

[0018] As an optional manner, in the above detection method, in step 6, the volatility of the key quality attributes of the microscopic image is characterized by a histogram statistical method.

[0019] In a second aspect, the present invention provides an application of the method described in the first aspect in a quality control method based on microscopic images of precious and fine medicines, wherein the precious and fine medicines include precious and fine medicine powders of large honeyed Chinese medicine pills.

[0020] As an optional method, in the above application, the application includes establishing a method for identifying the microscopic characteristics of precious and fine medicines, objectively quantifying them, and screening key evaluation indicators of microscopic images. The application improves the mixing uniformity and formability of the finished products of precious and fine medicines by monitoring the microscopic morphological parameters, thereby realizing quality control of the preparation and mixing process of large honey pills of traditional Chinese medicine.

[0021] In a third aspect, the present invention provides an artificial intelligence device for detecting microscopic properties of precious and fine medicines, wherein the device detects microscopic properties of precious and fine medicines using the method described in the first aspect above, and the device comprises:

[0022] a) Image acquisition module: used to obtain microscopic image information of the precious fine drug to be tested;

[0023] b) Image preprocessing module: used for preprocessing the microscopic image information by grayscale conversion, geometric transformation, image enhancement, etc.;

[0024] c) Critical quality attribute identification module: used to extract the region of interest (ROI) of the microscopic image from the information after image preprocessing, and perform binarization on the ROI; further use image processing methods such as opening operation and hole filling to remove noise and improve particle contours; perform statistical analysis on the processed microscopic images to quantify the microscopic characteristics of precious fine drugs; perform correlation analysis, grey correlation analysis and principal component analysis on the microscopic characteristics of precious fine drugs to identify the critical quality attributes; use histogram statistical methods to characterize the volatility of microscopic critical quality attributes.

[0025] In a fourth aspect, the present invention provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein when the processor executes the program, the method for detecting microscopic properties of precious fine medicines based on artificial intelligence as described in the first aspect above is implemented.

[0026] In a fifth aspect, the present invention provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for detecting microscopic properties of precious fine medicines based on artificial intelligence as described in the first aspect above is implemented.

[0027] The present invention has the following beneficial effects:

[0028] The present invention uses artificial musk in the original powder of Tongren Niuhuang Qingxin Pills as a research carrier, adopts microscopic imaging technology, establishes a method for collecting images of precious and fine medicines, and optimizes the dispersion medium and dispersion mode. Further combined with digital image processing technology, adaptive threshold segmentation, binarization processing, open operation, and hole filling system processing methods are used in sequence to realize the identification of microscopic features of precious and fine medicines. 27 microscopic characteristic indicators of precious and fine medicines are statistically analyzed, and a combined method for screening microscopic image characteristic indicators of precious and fine medicines by correlation analysis, gray correlation analysis, and principal component analysis is established. Seven key evaluation indicators of microscopic images are obtained by screening artificial musk. Finally, it is found that roundness is the key quality attribute of artificial musk, and the identification of key microscopic quality attributes of precious and fine medicines is realized. It provides information technology support for the study of three-dimensional characteristics of traditional Chinese medicine image visualization, and also provides a new approach for the study of quality control of traditional Chinese medicine manufacturing. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 : The original microscopic image of artificial musk and the image processing result. Among them, (a) the original microscopic image of artificial musk, (b) the image processing result of artificial musk microscopic image.

[0030] Figure 2 : The results of key quality attribute identification of artificial musk microscopic images. Among them, (a) two-dimensional correlation matrix heat map of morphological indicators of artificial musk microscopic images, (b) gray correlation degree of morphological indicators of artificial musk microscopic images, (c) principal component score map, and (d) principal component loading map.

[0031] Figure 3 : Frequency distribution histogram of key quality attributes of artificial musk microscopic images.

[0032] Figure 4 : Schematic diagram of the device structure provided by the present invention.

[0033] Figure 5 : Schematic diagram of the structure of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0034] The following is further described in conjunction with the accompanying drawings and specific implementation examples, but the examples do not limit the scope of the method of the present invention.

[0035] Embodiment: An artificial intelligence method, device and electronic device for detecting microscopic properties of precious and fine medicines 1. Optimization of a microscopic image acquisition method based on artificial musk of Tongren Niuhuang Qingxin Pills

[0036] (1) Instruments and materials: Olympus BX51 universal microscope imaging system (Olympus Corporation). The hardware of the system mainly consists of an optical microscope, an optical interface device, a CCD digital camera, a USB bus interface, and a computer processing system. The precious and fine medicine of Tongren Niuhuang Qingxin Pills is artificial musk (30 batches, Beijing Tongrentang Co., Ltd., China). n-Butanol and Tween 80 (Tianjin Damao Chemical Reagent Factory, analytical grade).

[0037] (2) Microscopic image measurement system for precious fine medicine: The microscopic image measurement system for precious fine medicine mainly consists of five parts: light source, optical imaging system, CCD, image acquisition card, and computer. The working principle of the system is as follows: the powder dispersed by the medium is placed on a glass slide, and the glass slide is placed on the microscope workbench. The powder is magnified by the microscope to form an imaging plane, which is converted by the optical system and imaged on the photosensitive surface of the CCD. The CCD converts the light signal into an electrical signal, and then the image acquisition card converts the electrical signal into a digital image. The digital image is read by the computer processing system and loaded into the Olympus BX51 microscopic image detection software for subsequent processing.

[0038] (3) Establish a method for acquiring microscopic images of precious and fine medicines by selecting the dispersion medium, dispersion method, dispersion time, and setting image acquisition parameters. When glycerol acetic acid test solution was used as the dispersion medium, the particles of artificial musk expanded; when glycerol test solution was used as the dispersion medium, the particles were evenly dispersed and there was no swelling phenomenon; when Tween 80 test solution and n-butanol test solution were used, the particles were prone to agglomeration and uneven dispersion. Therefore, glycerol test solution was selected as the dispersion medium for microscopic image detection of artificial musk. When manual oscillation was used for dispersion, the sample was unevenly dispersed, while the oscillator was used for dispersion, so the oscillator was selected to disperse the sample during microscopic detection. The results showed that when the dispersion time exceeded 3 minutes, the sample was more evenly dispersed. At the same time, considering factors such as detection efficiency, the oscillation dispersion time was set to 3 minutes. Appropriate image acquisition parameters were set to finally obtain microscopic images of precious and fine medicines. To ensure that the microscopic images of the four kinds of precious and fine medicines were clear and complete, the acquisition parameters: system magnification, scene mode, analog gain, and exposure compensation were set as shown in Table 1. The final image was good and could be used for subsequent analysis and processing.

[0039] Table 1 Microscopic image acquisition parameters of artificial musk

[0040]

[0041] (4) Microscopic image characteristics of artificial musk: According to literature and experimental research, the microscopic characteristics of artificial musk were analyzed. The microscopic images of the samples are as follows: Figure 1 As shown in (a), artificial musk is mostly composed of granular objects of various shapes, with numerous oil droplets, either scattered individually or connected into clumps, and reddish brown in color; there are translucent square or quasi-circular crystals, some of which have uneven edges, and there are many small granular or irregular crystals.

[0042] 2. Microscopic image processing and identification of artificial musk in Tongren Niuhuang Qingxin Pills

[0043] (1) Establishment of microscopic image processing method for precious fine medicine

[0044] First, the image is preprocessed using grayscale, geometric transformation, and image enhancement methods. After grayscale processing, some areas of the grayscale image have obvious noise, uneven illumination, and white spots appear in the middle of the starch granules due to the focal depth of the microscope. At the same time, the outline of the starch granules is close to the background color, which is not conducive to subsequent image segmentation. Therefore, the power law transformation in the grayscale transformation is further used to process the grayscale image to achieve the purpose of highlighting the starch granules while avoiding the influence of the background on the target granules. After the image is grayscale transformed, the uneven light and dark in the background is eliminated, the background and the middle of the starch granules are close to white, and the outline of the starch granules is deepened.

[0045] The adaptive threshold segmentation method was further used to binarize the microscopic image of artificial musk after grayscale transformation. The background grayscale value was set to 1, and the grayscale value of the starch granule outline was set to 0 to obtain a binary image. After the edge filling algorithm was processed, some starch granules still had holes due to the lack of outline pixel values, and the complete granule image could not be obtained. Therefore, the opening operation method was used and then the hole filling was performed, which not only improved the outline of the starch granules, smoothed the boundaries of the starch granules, but also removed the noise in the image. Figure 1 (b) is the result after the artificial musk starch granules were opened and processed.

[0046] The processed particles were further objectively quantified through image statistical analysis. The morphological parameter indicators obtained by parameter statistics include: angle, area, aspect ratio, major axis, minor axis, equivalent rectangle length, equivalent rectangle width, equivalent rectangle aspect ratio, centroid (X), centroid (Y), maximum diameter, average diameter, minimum diameter, maximum Feret diameter, average Feret diameter, minimum Feret diameter, fractal dimension, perimeter, contour perimeter, equivalent ellipse perimeter, equivalent ellipse perimeter ratio, maximum radius, minimum radius, radius ratio, roundness, major axis diameter, and minor axis diameter, a total of 27 morphological parameter indicators.

[0047] (2) Analysis of morphological characteristics of precious fine medicine microscopic images

[0048] As shown in Table 2, the roundness is between 0.238 and 0.513, indicating that most of the artificial musk microscopic particles are rectangular or polygonal, and the aspect ratio is between 1.429 and 1.636, indicating that the particle edge flatness is poor. Among the 27 index parameters, the radius ratio RSD is as high as 0.36, and the area and roundness RSD are both over 0.2, indicating that the size and shape of artificial musk particles fluctuate greatly. It can also be observed from the image that the artificial musk has different shapes and sizes, and there are circular, rectangular, and polygonal particles. The RSD of other indicators is between 0.018 and 0.148, among which the equivalent ellipse perimeter ratio RSD is the smallest. The angle, equivalent rectangular aspect ratio, centroid (Y), fractal dimension and maximum radius are right-skewed distributions, and the other indicators are left-skewed distributions. The kurtosis of the 27 indicators is mostly negative, indicating a flat distribution.

[0049] Table 2 Statistical results of microscopic morphological parameters of 30 batches of artificial musk

[0050]

[0051]

[0052] 3. Characterization of key quality attributes of artificial musk microscopic images

[0053] (1) Screening of key indicators of artificial musk microscopic images

[0054] 1) Correlation analysis of quality attributes of artificial musk microscopic images

[0055] Correlation analysis was performed on 27 morphological parameters of 30 batches of artificial musk microscopic images, and the two-dimensional correlation matrix heat map is shown in Figure 2 (a). The correlation coefficients among the 27 morphological indices ranged from -0.94 to 0.99. Among them, the correlation coefficients of Feret diameter (maximum Feret diameter, average Feret diameter, minimum Feret diameter) and 11 indices, including area, major axis, minor axis, equivalent rectangle length, equivalent rectangle width, maximum diameter, average diameter, equivalent ellipse circumference, maximum radius, major axis diameter, and minor axis diameter, were greater than 0.95, showing extremely significant correlation (P < 0.05). Therefore, Feret diameter was selected to replace the above indices. The correlation coefficient between the minimum diameter and the minimum radius was 0.99, showing a significant correlation (P < 0.05). Therefore, the minimum diameter was selected to replace the minimum radius. The correlation coefficient between the circumference and the contour circumference ratio was 0.97, showing a significant correlation (P < 0.05). The circumference was selected to replace the contour circumference ratio. Finally, the correlation analysis screened out the following indicators: angle, aspect ratio, equivalent rectangle aspect ratio, centroid (X), centroid (Y), minimum diameter, maximum Feret diameter, average Feret diameter, minimum Feret diameter, fractal dimension, perimeter, equivalent ellipse perimeter ratio, radius ratio, and roundness, a total of 14 indicators.

[0056] 2) Grey correlation analysis of quality attributes of artificial musk microscopic images

[0057] Based on the data screened by correlation analysis, the grey correlation analysis of 14 indicators is carried out. Figure 2 (b). Calculate the grey correlation coefficients between the characteristic parameters of the artificial musk microscopic image. As shown in Table 3, the correlations of the 14 indicators are all greater than 0.5, indicating that the 14 indicators are correlated. Among them, the correlation between the maximum Feret diameter and the average Feret diameter is 0.955, and the correlation between the maximum Feret diameter and the minimum Feret diameter is 0.900, both of which are greater than the threshold of 0.9, indicating that the average Feret diameter and the minimum Feret diameter are strongly correlated with the maximum Feret diameter. Therefore, the three indicators are combined, and the maximum Feret diameter is used as the representative of such indicators. Combined with correlation analysis, the indicators finally screened out are: angle, aspect ratio, equivalent rectangular aspect ratio, center of mass (X), center of mass (Y), minimum diameter, maximum Feret diameter, fractal dimension, perimeter, equivalent ellipse perimeter ratio, radius ratio, and roundness, a total of 12 indicators.

[0058]

[0059] 3) Principal component analysis of quality attributes of artificial musk microscopic images

[0060] Based on the above data, the results of 12 morphological indicators of 30 batches of artificial musk microscopic images were standardized to eliminate the dimension between indicators and then subjected to principal component analysis. With the eigenvalue greater than 1, the cumulative contribution rate of the first four principal components was 83.7%, of which PC1 contributed 47.0% and PC2 contributed 16.0%. The score chart is shown in Figure 2 (a). Load diagram is shown in Figure 2 (b), aspect ratio, radius ratio, and fractal dimension have large loads on PC1; maximum feret diameter and circumference have large loads on PC2. The results show that the first two principal components mainly integrate the information of the size and fractal dimension of the microscopic image particles of artificial musk. The loading factors of the first four principal components are calculated, as shown in Table 4. The absolute value of the loading factor is greater than 0.7 as the standard, and the screening indicators are: aspect ratio, equivalent rectangular aspect ratio, centroid (X), centroid (Y), minimum diameter, maximum feret diameter, fractal dimension, perimeter, equivalent ellipse perimeter ratio, radius ratio, roundness, a total of 11 indicators.

[0061] Table 4 Loading factors of the first four principal components of the morphological indexes of artificial musk microscopic images

[0062]

[0063] In summary, the characteristic indicators of artificial musk microscopic images obtained through correlation analysis, grey correlation analysis, and principal component analysis are: aspect ratio, equivalent rectangle aspect ratio, centroid (X), centroid (Y), minimum diameter, maximum feret diameter, fractal dimension, perimeter, equivalent ellipse perimeter ratio, radius ratio, and roundness, a total of 11 indicators. The histogram of the 8 characteristic indicators of artificial musk microscopic images is drawn, as shown in Figure 5 As shown, they are all normally distributed, with the RSD of roundness being 24%, and the RSD of other indicators being between 2% and 8%, indicating that the roundness of the 30 batches of artificial musks has a large volatility.

[0064] Taking the microscopic image characteristics of 30 batches of artificial musk samples as the research object, the key morphological index screening method of artificial musk microscopic images was established by using correlation analysis, gray key analysis and principal component analysis. The characteristic indicators screened from the microscopic images of artificial musk are: aspect ratio, equivalent rectangular aspect ratio, centroid (X), centroid (Y), minimum diameter, maximum feret diameter, fractal dimension, perimeter, equivalent ellipse perimeter ratio, radius ratio, roundness, a total of 11 indicators. The identification of key quality attributes of artificial musk microscopic images was achieved.

[0065] The histogram statistical method was further used to characterize the key quality attributes of artificial musk, see Figure 3The results showed that the above indicators were normally distributed, the RSD of the roundness of the artificial musk particles was 24%, and the RSD of the other 10 indicators was between 2% and 8%, indicating that the roundness of the 30 batches of artificial musk had a large fluctuation.

[0066] In summary, the screening of key indicators of artificial musk can be achieved by combining the three methods of correlation analysis, grey correlation analysis and principal component analysis. Eleven key evaluation indicators of microscopic images were obtained by screening artificial musk, which provide a basis for the comprehensive evaluation of the overall quality attributes of artificial musk. Particle morphology is an important factor affecting the macroscopic properties of precious fine medicines, especially the transmission of internal forces of particles and the composition characteristics of finished products. Since the particle morphology, structure and powder properties of precious fine medicines have an impact on the subsequent mixing and molding processes in the production process, such as the agglomeration and adsorption of particles during the mixing process, which causes problems such as uneven mixing, it is difficult to ensure the quality and clinical efficacy of the finished product. Therefore, through online monitoring of key microscopic quality attributes such as particle roundness, aspect ratio, and average feret diameter, the mixing uniformity and molding of the finished product can be improved to inherit and develop the traditional dosage forms of precious fine medicines.

[0067] Based on the same inventive concept, an apparatus corresponding to the method for detecting the microscopic properties of precious fine medicines is also provided in the embodiments of the present application. Since the principle of the problem solved by the apparatus in the embodiments of the present application is similar to the above-mentioned microscopic image detection method in the embodiments of the present application, the implementation of the apparatus can refer to the implementation of the method, and the repetitive parts will not be repeated.

[0068] See also Figure 4 , Figure 4 This is a schematic diagram of a microscopic detection device for precious fine drugs provided in the embodiment of the present application. Figure 4 As shown in , the microscopic image detection device 200 provided in the embodiment of the present application includes:

[0069] Image acquisition module 210, used to obtain microscopic image information of precious fine medicine

[0070] The image preprocessing module 220 is used to input the acquired precious fine medicine microscopic image into grayscale, geometric transformation, image enhancement and other methods to obtain a processed microscopic image.

[0071] The key quality attribute identification module 230 is used to perform adaptive threshold segmentation on the acquired processed microscopic image information and extract ROI; after the ROI is binarized, the open operation and hole filling method are used to remove noise, improve the missing particle contour, and smooth the particle boundary, so as to realize the identification of the microscopic characteristics of the precious fine medicine. Then, the identified image is statistically analyzed to realize the objective quantification of the microscopic image of the precious fine medicine. In addition, the RSD value of each indicator is calculated to examine the overall repeatability. Further, through the methods of correlation analysis, gray correlation analysis and principal component analysis, the key quality attributes of the microscopic image of the precious fine medicine are screened to provide a basis for the comprehensive evaluation of the overall quality attributes of the precious fine medicine. Finally, the histogram statistical method is used to characterize the volatility of the key microscopic quality attributes, observe the distribution of each indicator, and calculate the RSD value of each indicator to examine the volatility of each indicator. The large fluctuation indicates that the indicator should be focused on during the production process, providing theoretical and technical support for improving the mixing uniformity and formability of the finished product.

[0072] For descriptions of the processing flow of each module in the device and the interaction flow between each module, reference may be made to the relevant descriptions in the above method embodiment, which will not be described in detail here.

[0073] The present application also provides an electronic device 100, such as Figure 5 FIG. 1 is a schematic diagram of the structure of an electronic device 100 provided in an embodiment of the present application, including:

[0074] Processor 310: used to process data and perform computing tasks;

[0075] Memory 320: used to store data and programs, including internal memory and external memory;

[0076] Communication interface 330: used for communication connection with other devices;

[0077] Communication bus 340: used to connect the processor 310, the memory 320 and the communication interface 330, and provide a data transmission channel.

[0078] In this embodiment, the processor 310 may include one or more central processing units (CPUs), graphics processing units (GPUs), and other processing units. The memory 320 includes a memory for storing temporary data and an external memory for storing long-term data. The communication interface 330 may include wired and wireless communication interfaces, supporting a variety of data transmission protocols. The communication bus 340 may use a bus structure such as PCIe, I2C, SPI, etc.

[0079] During the operation of the system, the processor 310 reads program codes from the memory 320 and executes related tasks, and exchanges data with the memory 320 and the communication interface 330 through the communication bus 340 to implement data processing and transmission functions.

[0080] The electronic device 100 of this embodiment can be applied to various scenarios such as smart terminal devices, embedded systems and computer systems. Through this structural design, it can efficiently process and transmit data, meet diverse application needs, and improve the overall performance and stability of the system.

[0081] In an optional embodiment of the present application, each functional module of the electronic device 100 can be specifically implemented as hardware, software, or a combination of hardware and software. In an optional embodiment, the processor 310 can be specifically embodied as a computer storage medium, and in another optional embodiment, the computer program product is specifically embodied as a software product, such as a software development kit (SDK).

[0082] The memory 320 may further include a memory 321 and an external memory 322 for storing temporary data and long-term data respectively. The memory 321 may be a random access memory (RAM), and the external memory 322 may be a flash memory or a hard disk.

[0083] 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 equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. An artificial intelligence method for detecting microscopic properties of precious fine medicines, characterized in that: The detection method comprises the following steps: Step 1: Establish a microscopic imaging system and optimize the imaging parameters of the microscope; Step 2: Optimize the dispersion medium, dispersion method and dispersion time of the precious fine drug, and construct a microscopic image acquisition method; Step 3: Collect the microscopic image of precious medicine and pre-process the microscopic image; Step 4: Extract the region of interest (ROI) of the microscopic image, perform binarization on the ROI, and further use image processing methods to remove noise and improve particle contours; Step 5: Statistical analysis of the microscopic images is performed to quantify the microscopic characteristics of the precious drug and further identify the key quality attributes of the microscopic images; Step 6: Characterize the variability of critical quality attributes of microscopic images.

2. The method according to claim 1, characterized in that In step 3, image preprocessing includes grayscale, geometric transformation, image enhancement and other methods.

3. The method according to claim 1, characterized in that In the step 4, the ROI is extracted by adaptive threshold segmentation, and the image processing method includes methods such as opening operation and hole filling, which are used to remove noise, improve particle outline loss, and smooth particle boundaries.

4. The method according to claim 1, characterized in that: In step 5, the statistical analysis of the microscopic image includes quantification of morphological parameters such as aspect ratio, minimum diameter, fractal dimension and roundness of the microscopic particles, and the key quality attribute identification method of the microscopic image includes correlation analysis, grey correlation analysis and principal component analysis.

5. The method according to claim 1, characterized in that In step 6, the volatility of the key quality attributes of the microscopic image is characterized by a histogram statistical method.

6. Application of the method according to any one of claims 1 to 5 in a quality control method based on microscopic images of precious fine medicines, characterized in that: The precious fine medicine includes precious fine medicine powder of large honey pills of traditional Chinese medicine.

7. The use according to claim 6, characterized in that: The application includes establishing methods for identifying the microscopic features of precious and fine medicines, objectively quantifying them, and screening key evaluation indicators of microscopic images. The application improves the mixing uniformity and formability of precious and fine medicine finished products by monitoring microscopic morphological parameters, and realizes quality control of the preparation and mixing process of large honey pills of traditional Chinese medicine.

8. An artificial intelligence device for detecting microscopic properties of precious and fine medicines, characterized in that: The detection device uses the method described in any one of claims 1 to 5 to detect the microscopic properties of precious medicines, and the detection device includes: a) Image acquisition module: used to obtain microscopic image information of the precious fine drug to be tested; b) Image preprocessing module: used for preprocessing the microscopic image information by grayscale conversion, geometric transformation, image enhancement, etc.; c) Critical quality attribute identification module: used to extract the region of interest (ROI) of the microscopic image from the information after image preprocessing, and perform binarization on the ROI; further use image processing methods such as opening operation and hole filling to remove noise and improve particle contours; perform statistical analysis on the processed microscopic images to quantify the microscopic characteristics of precious fine drugs; perform correlation analysis, grey correlation analysis and principal component analysis on the microscopic characteristics of precious fine drugs to identify the critical quality attributes; use histogram statistical methods to characterize the volatility of microscopic critical quality attributes.

9. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and running on the processor. When the processor executes the program, the method for detecting microscopic properties of precious fine medicines based on artificial intelligence according to any one of claims 1 to 5 is implemented.

10. A non-transitory computer-readable storage medium, characterized in that: The storage medium stores a computer program, which, when executed by a processor, implements the artificial intelligence method for detecting microscopic properties of precious fine medicines according to any one of claims 1 to 5.

Images