Traditional Chinese medicine decoction piece machine vision classification and recognition device based on multi-feature fusion
By integrating a high-definition industrial camera, an adjustable light source, and a deep learning model, the machine vision classification and recognition device for Chinese herbal medicine slices based on multi-feature fusion solves the problems of accuracy and efficiency in classification and recognition, and achieves adaptive and efficient classification and recognition.
Patent Information
- Application Number
- CN202510926689.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-07
- Publication Date
- 2025-10-28
AI Technical Summary
Traditional manual classification methods are time-consuming, labor-intensive, and highly susceptible to subjective factors. Existing machine vision systems are ill-suited to the diverse types, complex appearances, and sensitivity to lighting conditions of traditional Chinese medicine decoction pieces, resulting in low classification accuracy and an inability to adapt to the needs of new categories.
A machine vision classification and recognition device for Chinese herbal medicine slices based on multi-feature fusion is adopted. It integrates a high-definition industrial camera, an adjustable light source, multiple image processing algorithms and deep learning models to extract multiple feature information and achieve accurate classification through feature fusion and self-learning mechanisms.
It improves the accuracy and efficiency of classifying and identifying Chinese herbal medicine pieces, adapts to the characteristics of different Chinese herbal medicine pieces, has self-learning and updating capabilities, and meets the needs of different users.
Smart Images
Figure CN120852859A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of visual classification and recognition technology, and in particular to a machine vision classification and recognition device for traditional Chinese medicine decoction pieces based on multi-feature fusion. Background Technology
[0002] Accurate classification and identification of Chinese medicinal herbs are crucial in the production, processing, and sales of prepared herbal pieces. Traditional manual classification methods are not only time-consuming and labor-intensive but also heavily influenced by subjective factors, making them unsuitable for large-scale, high-efficiency classification. With the rapid development of computer and machine vision technologies, using machine vision for the classification and identification of prepared herbal pieces has become a trend. However, prepared herbal pieces are characterized by their diverse types, complex appearances, and subtle differences in features. A single image feature cannot comprehensively and accurately describe the characteristics of prepared herbal pieces, thus affecting the accuracy of classification and identification. Furthermore, different prepared herbal pieces have varying requirements for lighting conditions and shooting angles, further complicating the image acquisition and processing process. Furthermore, with the continuous updating and increase of the types of Chinese medicinal herbs, classification and recognition models need to have the ability to learn and update themselves to adapt to new classification requirements. Traditional machine vision classification and recognition methods often only extract single features of Chinese medicinal herbs, such as color or shape, ignoring other important features, resulting in low classification and recognition accuracy. Due to the wide variety of Chinese medicinal herbs, they are sensitive to conditions such as lighting and shooting angles. Existing machine vision systems are difficult to adapt to the characteristics of different Chinese medicinal herbs, affecting image acquisition and processing effects. As the types of Chinese medicinal herbs are constantly updated and increased, existing classification and recognition models cannot automatically adapt to new classification requirements, requiring manual model adjustment and optimization. Summary of the Invention
[0003] To address the aforementioned problems, this invention proposes a machine vision classification and recognition device for traditional Chinese medicine decoction pieces based on multi-feature fusion, which more accurately solves the problems mentioned in the background art.
[0004] This invention is achieved through the following technical solutions.
[0005] This invention proposes a machine vision classification and recognition device for traditional Chinese medicine (TCM) decoction pieces based on multi-feature fusion. The device includes an image acquisition unit equipped with a high-definition industrial camera and an adjustable light source to capture clear and detailed images of TCM decoction pieces, ensuring the image data contains sufficient detail for subsequent processing; a multi-feature extraction unit integrating various image processing algorithms and deep learning models to extract color, texture, shape, edge, and local detail features from the acquired images, comprehensively and accurately reflecting the appearance and intrinsic characteristics of the TCM decoction pieces; a feature fusion unit employing advanced fusion strategies and algorithms to organically fuse the various feature information obtained from the multi-feature extraction unit, forming a more comprehensive and accurate feature description to improve classification accuracy; a classification and recognition unit, based on the fused feature information, using machine learning or deep learning classifiers to quickly and accurately classify and recognize the TCM decoction pieces, outputting the classification results; and a control unit for coordinating and controlling the work of each unit, ensuring smooth and efficient overall operation of the device.
[0006] Preferably, the light source of the image acquisition unit is adjustable, including the adjustment of the light source type, light intensity and light angle, to adapt to the characteristics and classification requirements of different Chinese herbal medicine pieces.
[0007] Preferably, the deep learning model used by the multi-feature extraction unit includes, but is not limited to, convolutional neural networks, deep belief networks, or recurrent neural networks, to automatically learn and extract high-level features from the image.
[0008] Preferably, the feature fusion unit employs fusion strategies including but not limited to weighted averaging, principal component analysis, multi-kernel learning, or attention mechanisms to fully consider the complementarity and correlation between different features.
[0009] Preferably, the machine learning or deep learning classifier used by the classification and recognition unit includes, but is not limited to, support vector machines, random forests, K-nearest neighbors, or deep neural networks.
[0010] Preferably, it also includes a preprocessing unit for performing preprocessing operations such as denoising, enhancement, cropping, and rotation correction on the acquired image data to improve the accuracy of feature extraction and classification.
[0011] Preferably, it also includes a result display unit for intuitively displaying the classification and identification results, such as the category and confidence level of Chinese herbal medicine pieces.
[0012] Preferably, it also includes a data storage unit for storing acquired image data, extracted feature information, classification and recognition results, and device operation logs, etc.
[0013] Preferably, it also includes user interaction functions, allowing users to set classification criteria, adjust model parameters, and view classification results through the interactive interface.
[0014] Preferably, it also includes self-learning and updating capabilities, which can automatically update and optimize the feature extraction model, feature fusion strategy, and classification recognition model based on new Chinese herbal medicine samples and classification requirements, so as to improve the adaptability and generalization ability of the device.
[0015] Compared with existing technologies, the present invention provides a machine vision classification and recognition device for traditional Chinese medicine decoction pieces based on multi-feature fusion, which has the following beneficial effects.
[0016] This machine vision classification and recognition device for traditional Chinese medicine (TCM) decoction pieces, based on multi-feature fusion, integrates various image processing algorithms and deep learning models. The device can automatically extract multiple feature information from TCM decoction pieces, such as color, texture, shape, edges, and local details. These comprehensive and accurate feature descriptions provide a strong foundation for classification and recognition, significantly improving accuracy and efficiency. Furthermore, the application of the feature fusion unit further integrates multiple feature information, forming a more comprehensive feature description and further enhancing classification and recognition performance.
[0017] This machine vision classification and recognition device for traditional Chinese medicine (TCM) decoction pieces, based on multi-feature fusion, features an adjustable light source system, user interaction functions, and a self-learning and update mechanism. This allows the device to adapt to the characteristics and classification requirements of different TCM decoction pieces. Users can adjust the light source type, intensity, and angle to obtain optimal image quality. The user interaction function allows users to set classification criteria and adjust model parameters, meeting the needs of different users. The self-learning and update mechanism enables the device to continuously adapt to new TCM decoction piece samples and classification requirements, improving its adaptability and flexibility.
[0018] This machine vision classification and recognition device for traditional Chinese medicine decoction pieces, based on multi-feature fusion, is equipped with a result display unit and a data storage unit. It intuitively displays the classification and recognition results and stores the collected image data, extracted feature information, classification and recognition results, and the device's operation logs. The result display unit provides easy-to-understand classification results for convenient user viewing and verification. The data storage unit supports data backup, recovery, and query functions, ensuring data security and traceability. These designs enhance the user experience and facilitate subsequent data mining and model optimization. Attached Figure Description
[0019] Figure 1 This is a schematic diagram of the structure of a machine vision classification and recognition device for traditional Chinese medicine decoction pieces based on multi-feature fusion proposed in this invention.
[0020] In order to more clearly and completely illustrate the technical solution of the present invention, the present invention will be further described below with reference to the accompanying drawings. Example
[0021] like Figure 1 As shown in the figure, one embodiment of the present invention proposes a machine vision classification and recognition device for traditional Chinese medicine (TCM) decoction pieces based on multi-feature fusion. The image acquisition unit is equipped with a high-definition industrial camera and an adjustable light source (such as an LED ring light source, fiber optic light source, etc.). By adjusting the light source type (such as white light, red light, ultraviolet light, etc.), light intensity, and light angle, clear and detailed images of TCM decoction pieces can be captured under different environments. The camera is set to a high-resolution mode to acquire image data containing rich detail information. The multi-feature extraction unit integrates various image processing algorithms (such as the Sobel operator, Canny edge detection, Gabor filtering, etc.) and deep learning models (such as convolutional neural networks (CNN), deep belief networks (DBN), recurrent neural networks (RNN, etc.). These algorithms and models can extract color features (such as RGB values, HSV spatial distribution), texture features (such as gray-level co-occurrence matrix, LBP local binary pattern), shape features (such as contour, area, perimeter), edge features (such as edge intensity, direction), and local detail features (such as micro-cracks, spots) from the acquired images. The feature fusion unit employs advanced fusion strategies and algorithms such as weighted averaging, principal component analysis (PCA), multi-kernel learning, or attention mechanisms to organically fuse various feature information obtained from multiple feature extraction units. By considering the complementarity and correlation between different features, a more comprehensive and accurate feature description is formed, providing strong support for classification and recognition. Based on the fused feature information, the classification and recognition unit uses machine learning or deep learning classifiers such as Support Vector Machine (SVM), Random Forest (RF), K-Nearest Neighbors (KNN), or Deep Neural Network (DNN) to quickly and accurately classify and recognize traditional Chinese medicine decoction pieces. After training and optimization, the classifier can output the category label and confidence score of the traditional Chinese medicine decoction pieces.
[0022] The control unit coordinates and controls the various units through programming, ensuring smooth and efficient overall operation of the device. The control unit is responsible for receiving user commands, scheduling resources, and monitoring operational status. The device can automatically and accurately identify the categories of Chinese herbal medicine slices, improving the efficiency and accuracy of herbal medicine slice classification.
[0023] This invention designs an adjustable light source system, including a light source type selector, a light intensity adjuster, and a light angle adjuster. Users can adjust the light source type (e.g., selecting white light for general observation, or ultraviolet light for detecting fluorescence properties), light intensity (e.g., adjusting brightness to suit different reflectivities of medicinal herbs), and light angle (e.g., changing the light source position to eliminate shadows or highlight specific features) via a control panel or software interface, based on the characteristics of the medicinal herbs and classification requirements. This adjustability allows the device to adapt to a wider variety of medicinal herbs, improving image acquisition quality and classification accuracy.
[0024] In this invention, deep learning models such as Convolutional Neural Networks (CNNs), Deep Belief Networks (DBNs), or Recurrent Neural Networks (RNNs) are integrated into the multi-feature extraction unit. By training these models, they can automatically learn and extract high-level features from images. For example, CNNs can learn local features (such as edges and textures) and global features (such as shape and structure); DBNs can handle complex data distributions and dependencies; and RNNs can process sequential data and capture temporal dynamic features. The application of deep learning models improves the accuracy and efficiency of feature extraction, providing richer feature information for classification and recognition.
[0025] In this invention, the feature fusion unit implements fusion strategies such as weighted averaging, principal component analysis (PCA), multi-kernel learning, or attention mechanisms. The weighted averaging method assigns different weights based on the importance of features; the PCA method extracts key feature components through dimensionality reduction; the multi-kernel learning method combines the advantages of multiple kernel functions to improve the flexibility of feature fusion; and the attention mechanism dynamically adjusts the fusion weights based on the saliency of features. The application of these fusion strategies allows for the organic combination of various feature information, forming a more comprehensive and accurate feature description, thereby improving the accuracy of classification and recognition.
[0026] In this invention, machine learning or deep learning classifiers such as Support Vector Machine (SVM), Random Forest (RF), K-Nearest Neighbors (KNN), or Deep Neural Network (DNN) are deployed in the classification and recognition unit. By training these classifiers, they can classify and recognize traditional Chinese medicine decoction pieces based on fused feature information. For example, SVM is suitable for handling high-dimensional feature spaces; RF can handle a large number of features and samples; KNN classifies based on similarity metrics; and DNN can learn complex nonlinear relationships. The application of these classifiers enables rapid and accurate classification of traditional Chinese medicine decoction pieces, improving the practicality and reliability of the device.
[0027] In this invention, a preprocessing unit is added to the device to perform preprocessing operations on the acquired image data, such as denoising (e.g., using Gaussian filtering, median filtering, etc.), enhancement (e.g., histogram equalization, contrast stretching, etc.), cropping (e.g., removing irrelevant regions), and rotation correction (e.g., using Hough transform to detect straight lines for rotation correction). The preprocessing operations improve the quality of the image data and provide a better foundation for subsequent feature extraction and classification recognition.
[0028] In this invention, a result display unit, such as an LCD screen or computer interface, is designed in the device to intuitively display the classification and identification results. The result display unit can display information such as the category label of the Chinese herbal medicine slices, confidence level (such as probability value), and classification time, making it convenient for users to view and verify the classification results. The result display unit provides intuitive and easy-to-understand classification and identification results, improving the user-friendliness of the device.
[0029] In this invention, a data storage unit, such as a hard disk or cloud storage service, is configured in the device to store acquired image data, extracted feature information, classification and recognition results, and device operation logs. The data storage unit supports data backup, recovery, and query functions, ensuring data security and traceability. It also provides the device with data management and analysis capabilities, facilitating subsequent data mining and model optimization.
[0030] In this invention, a user interface, such as a touchscreen or computer software interface, is designed within the device. This interface supports users in setting classification criteria (e.g., selecting categories of Chinese herbal medicine slices of interest), adjusting model parameters (e.g., learning rate, number of iterations), and viewing classification results (e.g., real-time display or exporting reports). The user interface provides a user-friendly graphical interface and a simple operating process, reducing the difficulty of use. The user interaction functions make the device more flexible and customizable, meeting the needs and usage scenarios of different users.
[0031] In this invention, a self-learning and updating mechanism is implemented in the device. Through online learning or incremental learning, the feature extraction model, feature fusion strategy, and classification recognition model are automatically updated and optimized based on new Chinese herbal medicine samples and classification requirements. For example, new sample data can be periodically retrieved from the data storage unit and used to train and optimize the model; or, through a user feedback mechanism, user evaluations and suggestions on the classification results can be collected to improve model performance. This self-learning and updating capability enables the device to adapt to constantly changing Chinese herbal medicine classification requirements, improving its adaptability and generalization ability.
[0032] Finally, it should be noted that the basic concepts have been described above. Obviously, for those skilled in the art, the detailed disclosure above is merely illustrative and does not constitute a limitation of this specification. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this specification. Such modifications, improvements, and corrections are suggested in this specification, and therefore, such modifications, improvements, and corrections still fall within the spirit and scope of the exemplary embodiments of this specification. Furthermore, this specification uses specific terms to describe embodiments of this specification. For example, "an embodiment," "one embodiment," and / or "some embodiments" refer to a feature, structure, or characteristic associated with at least one embodiment of this specification. Therefore, it should be emphasized and noted that "an embodiment," "one embodiment," or "an alternative embodiment" mentioned twice or more in different locations in this specification do not necessarily refer to the same embodiment. In addition, certain features, structures, or characteristics in one or more embodiments of this specification can be appropriately combined. Furthermore, unless expressly stated in the claims, the order of processing elements and sequences, the use of numbers and letters, or other names described in this specification are not intended to limit the order of the processes and methods of this specification.
[0033] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A machine vision classification and recognition device for traditional Chinese medicine decoction pieces based on multi-feature fusion, characterized in that, include The image acquisition unit is equipped with a high-definition industrial camera and an adjustable light source to capture clear and detailed images of Chinese herbal medicine slices, ensuring that the image data contains sufficient detail for subsequent processing. The multi-feature extraction unit integrates various image processing algorithms and deep learning models, and can extract a variety of feature information from the acquired images, such as color features, texture features, shape features, edge features, and local detail features. These feature information comprehensively and accurately reflect the appearance and internal characteristics of Chinese herbal medicine slices. The feature fusion unit employs advanced fusion strategies and algorithms to organically integrate various feature information obtained from multiple feature extraction units, forming a more comprehensive and accurate feature description to improve the accuracy of classification and recognition. The classification and recognition unit, based on the fused feature information, uses machine learning or deep learning classifiers to quickly and accurately classify and recognize Chinese herbal medicine pieces, and outputs the classification results. And a control unit, used to coordinate and control the work of each unit, to ensure that the overall operation of the device is smooth and efficient.
2. The classification and identification device according to claim 1, characterized in that, The light source of the image acquisition unit is adjustable, including the adjustment of the light source type, light intensity and light angle, to adapt to the characteristics and classification requirements of different Chinese herbal medicine pieces.
3. The classification and identification device according to claim 1, characterized in that, The deep learning model used by the multi-feature extraction unit includes, but is not limited to, convolutional neural networks, deep belief networks, or recurrent neural networks, to automatically learn and extract high-level features from images.
4. The classification and identification device according to claim 1, characterized in that, The feature fusion unit employs fusion strategies including, but not limited to, weighted average, principal component analysis, multi-kernel learning, or attention mechanisms, to fully consider the complementarity and correlation between different features.
5. The classification and identification device according to claim 1, characterized in that, The machine learning or deep learning classifiers used by the classification and recognition unit include, but are not limited to, support vector machines, random forests, K-nearest neighbors, or deep neural networks.
6. The classification and identification device according to claim 1, characterized in that, It also includes a preprocessing unit, which performs preprocessing operations such as denoising, enhancement, cropping, and rotation correction on the acquired image data to improve the accuracy of feature extraction and classification.
7. The classification and identification device according to claim 1, characterized in that, It also includes a results display unit, which is used to intuitively display the classification and identification results, such as the category and confidence level of Chinese herbal medicine pieces.
8. The classification and identification device according to claim 1, characterized in that, It also includes a data storage unit for storing acquired image data, extracted feature information, classification and recognition results, and device operation logs.
9. The classification and identification device according to claim 1, characterized in that, It also includes user interaction features, allowing users to set classification criteria, adjust model parameters, and view classification results through the interactive interface.
10. The classification and identification device according to claim 1, characterized in that, It also includes self-learning and updating capabilities, which can automatically update and optimize feature extraction models, feature fusion strategies, and classification recognition models based on new Chinese herbal medicine samples and classification requirements, so as to improve the adaptability and generalization ability of the device.