An orange-yellow sopheratain raw material microscopic image recognition method
By extracting basic microscopic features and mapping parameters, combined with multivariate regression models and similarity calculations, the problems of unstable feature extraction and inaccurate impurity identification in the microscopic image recognition of cassia seed raw materials were solved, and high-precision microscopic analysis was achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHAANXI TIANXINGJIAN BIOCHEMICAL TECH CO LTD
- Filing Date
- 2026-01-30
- Publication Date
- 2026-06-05
AI Technical Summary
Existing microscopic detection methods for identifying cassia aurantiacus raw materials suffer from unstable feature extraction and inaccurate impurity identification, leading to subjective biases in purity assessment and raw material grading. In particular, under conditions of crystal overlap, particle aggregation, or localized reflective disturbance, it is difficult to accurately identify the differences between abnormally disturbed areas and pure crystal areas.
We employ methods such as microscopic basic feature extraction, parameter mapping, microscopic partition recognition, and texture consistency determination. Through Z-score normalization, multivariate regression model, and parameter registration techniques, we establish a quantitative mapping from microscopic features to structural parameters. Combined with Euclidean distance and cosine similarity calculation, we achieve dynamic texture change recognition of microscopic images.
It significantly improved the sensitivity and stability of identifying abnormal areas in the microscopic images of cassia seed raw materials, improved the accuracy of identification under complex lighting and texture interference conditions, and enhanced the overall accuracy of microscopic analysis.
Smart Images
Figure CN122157250A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent microscopic detection technology, specifically to a method for microscopic image recognition of cassia seed raw materials. Background Technology
[0002] As a natural flavonoid compound with multiple physiological activities such as anti-oxidation, liver protection, and lipid reduction, the purity and microstructure of cassia seed extract have a significant impact on subsequent efficacy and quality control. Existing microscopic detection methods mostly rely on manual microscopic observation or texture recognition based on simple image segmentation algorithms. However, due to the complex crystal morphology and significant differences in light reflection of cassia seed extract at the microscopic scale, traditional methods often suffer from unstable feature extraction and inaccurate impurity identification in microscopic image recognition, leading to subjective biases in purity assessment and raw material grading.
[0003] In existing technologies, the feature extraction process of microscopic images is often based on global texture descriptors or grayscale histogram features. These features cannot fully reflect the differences in microstructure between different regions of the orange-yellow cassia seed microscopic image. Especially in the presence of crystal overlap, particle aggregation, or local reflective disturbances, the system has difficulty in accurately identifying the differences between abnormal disturbance areas and pure crystal areas, resulting in subsequent errors in impurity judgment, distortion of texture analysis, and unstable sample classification. Summary of the Invention
[0004] To address the shortcomings of existing technologies, this invention provides a method for identifying microscopic images of cassia seed raw materials, thereby solving the problems mentioned in the background section.
[0005] To achieve the above objectives, the present invention provides the following technical solution: In a first aspect, embodiments of the present invention provide a method for microscopic image recognition of cassia seed raw materials, comprising the following steps: S1. Extract the basic microscopic characteristics of the raw material of Cassia tora extract and obtain the basic microscopic characteristics; S2. Use the basic microscopic features to perform parameter mapping to obtain the set of microscopic structure parameters; S3. Use the set of microscopic structure parameters to identify microscopic regions and obtain the target region set; S4. Based on the target region set, perform impurity structure comparison to obtain a pure microscopic region dataset; S5. Use the pure microscopic region dataset to determine the texture consistency and obtain the microscopic identification results of the orange-yellow cassia seed raw material.
[0006] To further optimize this technical solution, step S1 involves microscopic image standardization processing, feature region identification and boundary extraction, microscopic feature analysis and quantification, and feature data standardization and output, ultimately forming a basic microscopic feature set. Each It represents a standardized microstructural characteristic parameter.
[0007] To further optimize this technical solution, step S2 is based on the microscopic basic feature set output in step S1. The descriptive features in the microscopic images of the orange-yellow cassia seed raw material are transformed into quantifiable structural parameters, thus providing a unified quantitative basis for subsequent partition identification. Step S2 specifically includes the following steps: Feature normalization and scale unification; Mapping function fitting and parameter solving; Parameter validity verification and output transformation.
[0008] To further optimize this technical solution, in step S2, when performing feature normalization and scale unification: For input features Normalization is performed using the Z-score standardization method, and the calculation formula is as follows: ; in and These are the first and second images in the entire training sample set. The mean and standard deviation of each feature; standardization ensures the comparability of different features in terms of units and numerical range, providing a consistent input for fitting the mapping function.
[0009] To further optimize this technical solution, step S2 involves fitting the mapping function and solving for the parameters: Mapping function This is achieved through a multivariate regression model, and its expression is: ; in: The characteristic contribution coefficient represents the first... The microscopic feature is related to the first The influence weight of each structural parameter; This is a bias term used to correct for baseline differences; By fitting the standard sample dataset provided in step S1, determine and The value is set such that the fitting residuals satisfy the mean square error being lower than a preset threshold, and the physical rationality of the parameters is guaranteed.
[0010] Further optimization of this technical solution includes parameter validity verification and output conversion: After fitting, the obtained parameters Perform a consistency check: Outliers were removed using residual analysis and correlation analysis; Perform range verification on each parameter to ensure that the value is within a reasonable physical range; The qualified parameters are structured and encoded to form a set of microstructure parameters: ; In the formula, ; Indicates the first Each of the following basic microscopic features is described numerically: particle morphology and outline, optical density distribution, cell cavity structure, and tissue arrangement orientation. Represents the first in the set of microstructure parameters Several parameters, including edge integrity, layer density difference, cavity ratio, and arrangement angle; This represents a mapping function, established through fitting a multivariate linear or nonlinear regression model, which enables a quantitative correspondence between input features and structural parameters.
[0011] To further optimize this technical solution, step S3 is based on the limit structure parameter set obtained in step S2. By logically determining the stability, continuity, and structural hierarchy differences of different parameter combinations, the microscopic images of the orange-yellow cassia seed raw material are divided into multiple functional recognition regions. Step S3 specifically includes the following stages: Data mapping phase; Similarity calculation stage; Partition identification stage.
[0012] To further optimize this technical solution, step S3 is performed during the data mapping stage: The parameter set P generated in step S2 is mapped according to the microscopic image coordinate index corresponding to the parameter extraction. The mapping is performed using a parameter registration technique based on image coordinate index mapping, so that the parameter matrix is consistent with the microscopic image in spatial dimension, forming a microscopic parameter matrix M(x,y). Through this mapping, it is ensured that each parameter value corresponds one-to-one with a specific pixel or pixel block in the image space.
[0013] To further optimize this technical solution, step S3 in the similarity calculation stage: The similarity of the parameter vectors of each pixel in matrix M(x,y) is measured. Mature Euclidean distance or cosine similarity calculation models can be used to calculate the feature distance of local structural parameters, thereby obtaining the structural similarity matrix S(x,y).
[0014] To further optimize this technical solution, step S3 is performed during the partition identification stage: Based on similarity matrix The microscopic image space is clustered or segmented; a region segmentation algorithm based on feature similarity clustering is used to aggregate pixel regions with high parameter similarity to obtain microscopic partitioning results; this process generates a set of region indexes. ; Each of them It represents a target region with similar microstructural features.
[0015] In a second aspect, embodiments of the present invention provide a computer device, including a memory and a processor, wherein the memory stores a computer program, and the computer program instructions, when executed by the processor, implement the steps of a method for microscopic image recognition of cassia seed raw materials as described in the first aspect of the present invention.
[0016] Thirdly, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, they implement the steps of a method for microscopic image recognition of cassia seed raw materials as described in the first aspect of the present invention.
[0017] Compared with the prior art, the present invention provides a method for microscopic image recognition of cassia seed raw materials, which has the following beneficial effects: This method for identifying microscopic images of cassia seed raw materials utilizes an anomaly perturbation recognition mechanism based on an electromagnetic reference standard. This enables the system to identify potential impurity regions from dynamic texture changes in microscopic images, transforming the process from static image recognition to dynamic microscopic perturbation feature recognition. This method not only significantly improves the sensitivity of identifying anomalous regions in microscopic images of cassia seed raw materials but also enhances recognition stability under complex lighting and texture interference conditions, thereby improving the overall accuracy of microscopic analysis. Attached Figure Description
[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0019] Figure 1 This is a flowchart illustrating a method for identifying the microscopic image of cassia seed raw material proposed in this invention. Figure 2 This is a schematic diagram of the parameter mapping process for a microscopic image recognition method for cassia seed raw materials proposed in this invention; Figure 3 This is a schematic diagram of the microscopic partitioning identification process of a microscopic image recognition method for cassia seed raw materials proposed in this invention. Figure 4 This invention relates to a method for determining the texture consistency of a microscopic image recognition of cassia seed raw materials. Detailed Implementation
[0020] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
[0021] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.
[0022] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single embodiment or an embodiment selectively excluded from other embodiments.
[0023] Example 1: Reference Figures 1-4 This is the first embodiment of the present invention, which provides a method for microscopic image recognition of cassia seed raw material, including the following steps: S1. Extract the basic microscopic characteristics of the raw material of Cassia tora extract and obtain the basic microscopic characteristics; Step S1 provides quantifiable microscopic feature data input for subsequent parametric mapping and recognition, and specifically includes the following steps: Normalization of microscopic images: We used existing mature microscopic image correction and standardization techniques (such as multi-scale correction methods based on resolution resampling and image registration) to process the microscopic images of the cassia seed raw material; ensuring that the pixel size of different image sources is consistent with the microscopic field of view scale, so that each pixel corresponds to a fixed physical length unit, providing a unified benchmark for subsequent feature quantization.
[0024] Feature region identification and boundary extraction: By using mature edge detection and region growing algorithms (such as Canny edge detection and threshold-based region growing techniques), the boundaries of microstructures in the image are identified, including the outer edge of the cassia seed raw material particles and the contours of their internal cell cavities. This process uses morphological filtering techniques to remove noise interference, so that the extraction results retain only the true boundaries related to the raw material structure.
[0025] Microscopic Feature Analysis and Quantification: Using established image texture and optical density analysis methods (such as the Gray-Level Co-occurrence Matrix (GLCM) and Optical Density Gradient Statistical Method), the following basic microscopic features were extracted: Morphological characteristics: aspect ratio of particles, edge curvature, and ratio of perimeter to area; Optical density characteristics: light intensity distribution and optical density gradient in different regions; Tissue arrangement characteristics: directional distribution and arrangement angle of cell cavities; These features are all output in numerical form, forming a set of microscopic basic feature vectors.
[0026] Feature data standardization and output: To ensure the accuracy of subsequent parameter mapping, mature data normalization and feature standardization algorithms (such as Z-score standardization) are used to unify the range of feature values.
[0027] Ultimately, a set of fundamental microscopic features is formed. Each It represents a standardized microstructural characteristic parameter.
[0028] S2. Use the basic microscopic features to perform parameter mapping to obtain the set of microscopic structure parameters; Step S2 is based on the microscopic fundamental feature set output in step S1. The descriptive features in the microscopic images of the cassia seed raw material are transformed into quantifiable structural parameters, thereby providing a unified quantitative basis for subsequent partition identification. Step S2 specifically includes the following steps: Feature normalization and scale unification: For input features Normalization is performed using the Z-score standardization method, and the calculation formula is as follows: ; in and These are the first and second images in the entire training sample set. The mean and standard deviation of each feature. Standardization ensures the comparability of different features in terms of units and numerical range, providing consistent input for fitting the mapping function.
[0029] Mapping function fitting and parameter solving: Mapping function This is achieved through a multivariate regression model, and its expression is: ; in: The characteristic contribution coefficient represents the first... The microscopic feature is related to the first The influence weight of each structural parameter; This is a bias term used to correct for baseline differences; By fitting the standard sample dataset provided in step S1, determine and The value is set such that the fitting residuals satisfy the mean square error (MSE) being lower than a preset threshold, and the physical rationality of the parameters is guaranteed.
[0030] Parameter validity verification and output transformation: After fitting, the obtained parameters Perform a consistency check: Outliers were removed using residual analysis and correlation analysis; Perform range verification on each parameter to ensure that the value is within a reasonable physical range; The qualified parameters are structured and encoded to form a set of microstructure parameters: ; In the formula, ; Indicates the first Each of the following basic microscopic features is described numerically: particle morphology and outline, optical density distribution, cell cavity structure, tissue arrangement orientation, etc. Represents the first in the set of microstructure parameters Several parameters, such as edge integrity, layer density difference, cavity ratio, and arrangement angle; The mapping function can be established by fitting a multivariate linear or nonlinear regression model, so that the input features and structural parameters form a quantitative correspondence.
[0031] Existing technologies typically remain at the stage of statistical description of microscopic features or texture analysis, lacking a direct quantitative mapping between microscopic features and structural parameters. Step S2 establishes a mapping model from multidimensional microscopic features to microscopic structural parameters, realizing the structured expression of features and providing a quantifiable and verifiable parameter basis for subsequent partition identification and raw material consistency judgment. This logical processing approach differs from traditional microscopic image analysis methods that rely solely on statistical description.
[0032] S3. Use the set of microscopic structure parameters to identify microscopic regions and obtain the target region set; Step S3 is based on the limit structure parameter set obtained in step S2. By logically determining the stability, continuity, and structural hierarchy differences of different parameter combinations, the microscopic image of the orange-yellow cassia seed raw material is divided into multiple functional recognition regions; step S3 specifically includes the following steps: Data mapping phase: The parameter set generated in step S2 The parameters are mapped according to the coordinate indices of the microscopic image during parameter extraction. The mapping uses a mature "parameter registration technique based on image coordinate index mapping" to ensure that the parameter matrix is spatially consistent with the microscopic image, thus forming a microscopic parameter matrix. This mapping ensures that each parameter value corresponds one-to-one with a specific pixel or pixel block in the image space.
[0033] Similarity calculation stage: For matrix The similarity is measured using the parameter vectors of each pixel (or pixel block); mature Euclidean distance or cosine similarity calculation models can be used to calculate the feature distance of local structural parameters, thereby obtaining the structural similarity matrix. This matrix reflects the consistency of microscopic features between local regions.
[0034] Partition identification stage: Based on similarity matrix Clustering or segmentation of the microscopic image space is performed. Mature "feature similarity-based clustering region segmentation algorithms" (such as improved K-means or region growing methods) can be used to aggregate pixel regions with high parameter similarity, obtaining microscopic partitioning results. This process generates a set of region indices: ; Each of them It represents a target region with similar microstructural features.
[0035] target area set The final output of this step is represented as a set of region indexes or a matrix of region labels. Each region contains its corresponding spatial location and characteristic average parameters, which are used for regional feature comparison in the subsequent impurity analysis stage.
[0036] Existing microscopic image partitioning methods are mostly based on gray-level gradients or texture thresholds, which cannot effectively distinguish differences driven by structural parameters. Step S3 establishes a spatial clustering logic based on structural parameters by using spatial mapping of structural parameter sets and similarity-driven partitioning, thus differentiating itself from traditional image thresholding or edge detection-based partitioning methods in terms of logical path.
[0037] S4. Based on the target region set, perform impurity structure comparison to obtain a pure microscopic region dataset; Step S4 involves using the target region set identified in step S3. The set of microstructure parameters output in step S2 By combining these methods, structural consistency analysis was performed. Through parameter mapping, impurity identification, and region screening, a dataset of pure microscopic regions that conform to the microscopic characteristics of cassia aurantium was obtained. Step S4 specifically includes the following steps: Region and parameter mapping synchronization: Based on each region in step S3 A one-to-one correspondence is established between the spatial index and the parameter index in step S2. Using a mature "microscopic image parameter registration algorithm," synchronization between regions and parameters is achieved at the pixel level, forming a region parameter matrix. ; This matrix ensures that each region is associated with the mean and variance of the corresponding structural parameters, giving the comparison results a quantifiable basis.
[0038] Impurity characteristic comparison and analysis: For the mapping matrix Multidimensional comparison of parameter characteristics in each region was performed; using the mature "parameter threshold interval comparison method", the parameter vector of each region was compared with the standard microscopic feature parameter interval of cassia aurantium; When the mean value of parameters, texture density, or optical reflectance within a region deviates from the standard range by more than a set threshold, the region is marked as an impurity region. In specific calculations, the assessment of deviation can be achieved through the "statistical difference test method", which involves calculating the mean squared deviation between the regional parameter and the standard parameter, and using the standard deviation multiple as the screening criterion.
[0039] Region removal and recoding: Regions identified as impurities are removed, and the remaining pure regions are recoded.
[0040] In implementation, data consistency correction is achieved through "region renumbering and spatial index update technology," ensuring that the removed regions maintain continuity in spatial index and parameter references, thus forming an updated clean region set. .
[0041] Clean dataset generation and standardization: regional set Structural parameters are normalized to form a clean microscopic region dataset. This dataset represents the feature information of each region with a unified parameter dimension structure, including fields such as region index, spatial coordinate range, mean microscopic parameters, and texture features, providing standardized input for the next step of microscopic identification and analysis.
[0042] S5. Use the pure microscopic region dataset to determine the texture consistency and obtain the microscopic identification results of the orange-yellow cassia seed raw material; Step S5: Target region set obtained in step S3 Compared with the clean microscopic region dataset obtained in step S4 Based on this, a matching relationship is established through microscopic coordinate indexing and region labels to achieve a one-to-one or many-to-one mapping structure, providing a comparison benchmark for texture consistency analysis and yielding a set of microscopic identification results for the orange-yellow cassia seed raw material. Step S5 specifically includes the following steps: Texture feature extraction: For target region set With pure microscopic region dataset Texture features such as grayscale, directional gradient, and particle density were extracted from the corresponding microscopic regions. Grayscale features were obtained using the Gray-Level Co-occurrence Matrix (GLCM) method, directional gradient features were calculated using the Histogram of Oriented Gradients (HOG) algorithm, and particle density was statistically analyzed using the connected component labeling method. All obtained features were normalized and mapped to the [0,1] interval.
[0043] Consistency calculation: Each pair of corresponding areas The grayscale feature similarity is The directional gradient similarity is Particle density similarity is Then the overall texture consistency function is defined as: ; in, , respectively, represent the importance weights of different texture features; (Gray-level feature weight) is determined by the stability analysis of the gray-level distribution of pure samples. Its value reflects the contribution of the gray-level co-occurrence matrix to sample differentiation. It is usually determined by the gray-level variance stability assessment results of pure regions. (Oriented gradient feature weights) are determined based on the significance evaluation results of sample edges and structural directions, and mainly depend on the magnitude of information entropy change in the distribution of oriented gradients among sample differences; (Particle density feature weight) is determined based on the uniformity of particle distribution and the stability of density variance in the pure sample, reflecting the degree of influence of particle density within the material on the identification process; Similarity of each item , , The mature Structural Similarity metric (SSIM) is used to calculate the degree of feature similarity, which is represented by a continuous value from 0 to 1.
[0044] Consistency determination and identification output: Based on the statistical results of the characteristics of clean samples, the consistency threshold is determined. ,in: This represents the average texture consistency score among clean microscopic samples. This represents the standard deviation of the consistency score; when Time, region If it is determined to be consistent with the pure sample, it is otherwise marked as a non-consistent region. The determination results are output in the form of a binary image mask or a region label matrix, forming the final microscopic recognition result set. ,in Indicates the first Does the region match the texture of the clean sample?
[0045] Existing technologies often employ single texture features (such as gray-level co-occurrence matrix or Gabor filtering) for local region comparison, lacking cross-scale and cross-feature fusion mechanisms, which easily leads to misjudgments in complex grains or particle transition regions. Step S5 introduces a multi-parameter weighted texture consistency function within the microscopic region level, achieving joint comparison of three structural elements: gray level, orientation, and density. Furthermore, it adaptively determines the threshold based on the statistical characteristics of pure samples, thereby achieving high-precision consistency determination of the microscopic identification results of cassia seed raw materials.
[0046] Example 2: A practical application of the microscopic image recognition method for cassia seed raw materials described in Example 1: At a certain medicinal herb quality testing center, when conducting microscopic purity testing on raw materials containing cassia seed extract, the microscopic image recognition method provided in this invention was used to automate the analysis of batches of microscopic image samples. The specific process is as follows: Sample preparation and image acquisition: After extraction with a standard solvent and drying, the raw material of cassia seed extract was sampled and placed on a microscope slide. Imaging was then performed using a high-resolution polarizing microscope under constant illumination. The system automatically acquired multiple images at different depths of focus and illumination intensities, and established a raw image dataset.
[0047] Establishment of microscopic reference features: The system first extracts the optical and texture feature parameters of a subset of manually verified pure orange-yellow cassia seed crystal regions and establishes electromagnetic and optical reference standards. This process forms a benchmark for subsequent anomaly identification, used to measure the texture stability and optical consistency of various regions in the image.
[0048] Anomaly identification and region segmentation: In the microscopic image dataset, the system performs perturbation detection on each region based on the aforementioned reference benchmark. When a sudden change in local brightness or a discontinuity in texture direction is detected, the system marks the region as a "suspicious signal area." Simultaneously, regions with stable texture features are identified as candidate pure crystal regions. Through this distinction, the automatic separation of impurities and crystal regions in microscopic images can be initially achieved.
[0049] Optimization of clean areas in microtexture: For the identified suspicious and clean areas, the system further filters the microscopic textures, removing false anomalies caused by focusing errors or local reflections. Ultimately, a clean microscopic region dataset is obtained, ensuring that the regions used for analysis are representative and free of optical artifacts.
[0050] Microscopic recognition results output: The system uses the obtained target region set and the pure microscopic region dataset to determine texture consistency. If the texture distribution, optical features, and microstructural orientation of the pure region in the sample are consistent with the standard reference sample, the system outputs the identification conclusion "microscopic features meet the standard"; if there are significant differences, it outputs the result "contains impurities or has an uneven structure". Finally, a microscopic identification report is generated, including a schematic diagram of the microscopic image, region annotation results, and purity determination level, which is used for medicinal material quality assessment and production process traceability.
[0051] Example 3; This embodiment also provides a computer device applicable to a method for identifying microscopic images of cassia seed raw materials, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the method for identifying microscopic images of cassia seed raw materials as proposed in the above embodiment.
[0052] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements a method for identifying microscopic images of cassia seed raw materials as proposed in the above embodiments.
[0053] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.
[0054] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0055] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.
[0056] More specific examples (a non-exhaustive list) of computer-readable media include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.
[0057] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.
[0058] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.
Claims
1. A method for microscopic image recognition of cassia seed raw material, characterized in that, Includes the following steps: S1. Extract the basic microscopic characteristics of the raw material of Cassia tora extract and obtain the basic microscopic characteristics; S2. Use the basic microscopic features to perform parameter mapping to obtain the set of microscopic structure parameters; S3. Use the set of microscopic structure parameters to identify microscopic regions and obtain the target region set; S4. Based on the target region set, perform impurity structure comparison to obtain a pure microscopic region dataset; S5. Use the pure microscopic region dataset to determine the texture consistency and obtain the microscopic identification results of the orange-yellow cassia seed raw material.
2. The method for microscopic image recognition of cassia seed raw material according to claim 1, characterized in that, Step S1, through the processes of microscopic image standardization, feature region identification and boundary extraction, microscopic feature analysis and quantification, and feature data standardization and output, ultimately forms a basic set of microscopic features. Each It represents a standardized microstructural characteristic parameter.
3. The method for microscopic image recognition of cassia seed raw material according to claim 1, characterized in that, Step S2 is based on the microscopic fundamental feature set output in step S1. The descriptive features in the microscopic images of the orange-yellow cassia seed raw material are transformed into quantifiable structural parameters, thus providing a unified quantitative basis for subsequent partition identification. Step S2 includes the following steps: Feature normalization and scale unification; Mapping function fitting and parameter solving; Parameter validity verification and output transformation.
4. The method for microscopic image recognition of cassia seed raw material according to claim 3, characterized in that, In step S2, when performing feature normalization and scale unification: For input features Normalization is performed using the Z-score standardization method, and the calculation formula is as follows: ； in and These are the first and second images in the entire training sample set. The mean and standard deviation of each feature; standardization ensures the comparability of different features in terms of units and numerical range, providing a consistent input for fitting the mapping function.
5. The method for microscopic image recognition of cassia seed raw material according to claim 3, characterized in that, Step S2 involves fitting the mapping function and solving for the parameters: Mapping function This is achieved through a multivariate regression model, and its expression is: ； in: The characteristic contribution coefficient represents the first... The microscopic feature is related to the first The influence weight of each structural parameter; This is a bias term used to correct for baseline differences; By fitting the standard sample dataset provided in step S1, determine and The value is set such that the fitting residuals satisfy the mean square error being lower than a preset threshold, and the physical rationality of the parameters is guaranteed.
6. The method for microscopic image recognition of cassia seed raw material according to claim 3, characterized in that, The validity verification and output transformation of the parameters are as follows: After fitting, the obtained parameters Perform a consistency check: Outliers were removed using residual analysis and correlation analysis; Perform range verification on each parameter to ensure that the value is within a reasonable physical range; The qualified parameters are structured and encoded to form a set of microstructure parameters: ； In the formula, ； Indicates the first Each of the following basic microscopic features is described numerically: particle morphology and outline, optical density distribution, cell cavity structure, and tissue arrangement orientation. Represents the first in the set of microstructure parameters Several parameters, including edge integrity, layer density difference, cavity ratio, and arrangement angle; This represents a mapping function, established through fitting a multivariate linear or nonlinear regression model, which enables a quantitative correspondence between input features and structural parameters.
7. The method for microscopic image recognition of cassia seed raw material according to claim 1, characterized in that, Step S3 is based on the limit structure parameter set obtained in step S2. By logically determining the stability, continuity, and structural hierarchy differences of different parameter combinations, the microscopic images of the orange-yellow cassia seed raw material are divided into multiple functional recognition regions. Step S3 details Includes the following stages: Data mapping phase; Similarity calculation stage; Partition identification stage.
8. The method for microscopic image recognition of cassia seed raw material according to claim 7, characterized in that, Step S3 is in the data mapping stage: The parameter set generated in step S2 The parameters are mapped according to the coordinate index of the microscopic image at the time of parameter extraction. The mapping is carried out using the parameter registration technique based on image coordinate index mapping, so that the parameter matrix is consistent with the microscopic image in spatial dimension, forming a microscopic parameter matrix M(x,y). Through this mapping, it is ensured that each parameter value corresponds one-to-one with a specific pixel or pixel block in the image space.
9. The method for microscopic image recognition of cassia seed raw material according to claim 7, characterized in that, Step S3 is in the similarity calculation stage: The similarity of the parameter vectors of each pixel in matrix M(x,y) is measured; the feature distance of the local structural parameters is calculated by using the Euclidean distance or cosine similarity calculation model, so as to obtain the structural similarity matrix S(x,y).
10. The method for microscopic image recognition of cassia seed raw material according to claim 7, characterized in that, Step S3 is in the partition identification stage: Based on similarity matrix The microscopic image space is clustered or segmented; a region segmentation algorithm based on feature similarity clustering is used to aggregate pixel regions with high parameter similarity to obtain microscopic partitioning results; this process generates a set of region indexes. ； Each of them It represents a target region with similar microstructural features.