Cordyceps sinensis larva infection phenotypic characteristic intelligent detection method and system
The phenotype detection system for Cordyceps sinensis larva infection is established through deep learning and machine learning algorithms, which solves the problems of large errors and low efficiency of traditional detection methods, and achieves efficient and accurate detection and harvesting judgment of larva infection.
Patent Information
- Application Number
- CN202510502564.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-21
- Publication Date
- 2025-08-05
AI Technical Summary
The traditional Cordyceps sinensis detection method relies on manual observation, has large errors and low efficiency, and cannot meet the needs of efficient quality control.
A deep learning algorithm is used to establish an electronic image object detection model of Cordyceps sinensis bacteria, combined with a machine learning algorithm to extract and classify mycelium features, obtain growth rate, infection area and density, and establish appropriate harvest judgments.
It realizes objective and efficient detection of Cordyceps larva infection phenotype, supports grade classification and timely harvesting judgment, and improves detection accuracy and efficiency.
Smart Images
Figure CN120431369A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of Cordyceps sinensis breeding and cultivation, and in particular to an intelligent detection method and system for the infection phenotypic characteristics of Cordyceps sinensis larvae. Background Art
[0002] Cordyceps sinensis is a precious traditional Chinese medicinal ingredient, considered one of the "three treasures of Chinese medicine" along with ginseng and deer antler. Its quality and growth condition have a significant impact on its medicinal value and market competitiveness. Cordyceps sinensis is native to the eastern Qinghai-Tibet Plateau, where wild production is relatively low. In recent decades, the resource has been declining, and the plant faces the risk of resource depletion. Artificially cultivated Cordyceps sinensis offers significant market potential. Cordyceps sinensis cultivation and processing require comprehensive quality control, including mycelial testing and analysis. Traditional Cordyceps testing relies on manual observation and measurement to estimate the time of infection of larvae and assess the infection efficiency of mycelium within the host. The quality of the host is determined by characteristics such as appearance, color, texture, and size. This method requires considerable skill and experience, and the results are subject to operator influence, making them prone to errors and inefficient, making them inefficient and unable to meet the needs of Cordyceps sinensis production. Advances in machine vision technology have provided new methods for mycelial phenotyping, enabling automated and intelligent mycelial phenotyping. Therefore, there is an urgent need to develop a technical method that can objectively and efficiently detect the phenotype of Cordyceps sinensis mycelium.
[0003] To this end, the present invention proposes an intelligent detection method and system for the phenotypic characteristics of Cordyceps sinensis larvae infection. Summary of the Invention
[0004] The purpose of the present invention is to solve at least one of the technical problems raised in the background technology.
[0005] In one aspect, the present invention provides an intelligent method for detecting phenotypic characteristics of Cordyceps sinensis larvae infection, comprising the following steps:
[0006] S1: Collect electronic images of Cordyceps sinensis fungi;
[0007] S2: Processing the collected electronic images of Cordyceps sinensis to generate an electronic image positioning dataset of Cordyceps sinensis in the host;
[0008] S3: Establish a Cordyceps sinensis electronic image target detection model, input the Cordyceps sinensis electronic image positioning dataset, train the Cordyceps sinensis electronic image target detection model, and obtain a trained Cordyceps sinensis electronic image target detection model;
[0009] S4: Measure the morphological parameters of the Cordyceps sinensis mycelium image and obtain the growth rate based on the growth time;
[0010] S5: extracting characteristic blocks from the Cordyceps sinensis mycelium image to obtain a characteristic block image of the Cordyceps sinensis mycelium, and obtaining the infection area and infection density of the mycelium inside the insect body;
[0011] S6: Based on the Cordyceps sinensis mycelium feature block image obtained above, a feature dataset is generated based on different image features;
[0012] S7: Establish a classification model, input the obtained mycelium feature block image into the Cordyceps sinensis mycelium classification model, perform mycelium discrimination, and perform grade classification;
[0013] S8: Integrate information and make decisions on appropriate harvesting.
[0014] Furthermore, in said S1, the electronic image of the Cordyceps sinensis fungi collected includes the morphology, color, texture and distribution characteristics of the hyphae in the host.
[0015] Furthermore, in said S2, the collected electronic images are processed including image denoising, enhancement, segmentation and annotation operations to produce a high-precision electronic image positioning data set.
[0016] Furthermore, in S3, the Cordyceps sinensis electronic image target detection model is constructed based on a deep learning algorithm, including a convolutional neural network or a target detection algorithm.
[0017] Furthermore, in S4, the morphological parameters measured include hypha length, width, number of branches and distribution density, and the growth rate of the hyphae is calculated in combination with the growth time. The calculation formula of the growth rate ν is:
[0018]
[0019] Among them, ΔL is the change in hypha length, and Δt is the growth time.
[0020] Furthermore, in S5, the extraction of the feature blocks is achieved by an image segmentation algorithm, including threshold-based segmentation, edge detection, or region growing algorithm, to obtain the infection area and infection density of the hyphae inside the parasite. The infection density D is calculated as follows:
[0021]
[0022] Among them, A infected is the mycelial infection area, A total is the total cross-sectional area of the insect body.
[0023] Furthermore, in S6, the feature data set includes morphological features, color features, texture features and distribution features of mycelium, which are used to train the classification model.
[0024] Furthermore, in S7, the classification model is constructed based on a machine learning algorithm, including a support vector machine, a random forest or a deep neural network, for discriminating and classifying the mycelium. The discriminant function of the support vector machine is:
[0025]
[0026] Among them, a i is the Lagrange multiplier, y i is the category label, k(x i ,x) is the kernel function and b is the bias term.
[0027] Furthermore, in S8, the information integration includes mycelial growth rate, infection area, infection density and grade classification results, combined with the preset harvesting criteria, to output a judgment result of whether it is suitable for harvesting. The harvesting judgment function is:
[0028]
[0029] Among them, ν threshold is the growth rate threshold, D threshold is the infection density threshold.
[0030] In another aspect, an embodiment of the present application provides an intelligent detection system for phenotypic characteristics of Cordyceps sinensis larvae infection, comprising:
[0031] An image acquisition module, used for acquiring electronic images of Cordyceps sinensis fungi;
[0032] Image processing module, used to process the collected images and produce positioning data sets;
[0033] The target detection module is used to build and train the target detection model for Cordyceps sinensis electronic images;
[0034] Morphological parameter measurement module, used to measure mycelial morphological parameters and calculate growth rate;
[0035] Feature extraction module, used to extract hyphae feature blocks and obtain infection area and infection density;
[0036] Classification module, used to establish classification models and perform discrimination and grade classification of hyphae;
[0037] The decision-making module is used to integrate information and output the judgment results of suitable harvesting.
[0038] Compared with the prior art, the technical solution of this application has the following beneficial effects:
[0039] 1. The present invention can determine the mycelial infection density and infection rate in Cordyceps sinensis hosts by establishing a classification model and extracting feature modules;
[0040] 2. The present invention combines the growth time to obtain the growth rate, and further determines whether the Cordyceps sinensis is suitable for harvesting.
[0041] 3. The present invention objectively and efficiently detects the phenotype of Cordyceps sinensis mycelium to classify the grade of Cordyceps sinensis products. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments with reference to the following drawings:
[0043] Figure 1 This is a flow chart of an intelligent detection method for phenotypic characteristics of Cordyceps sinensis larvae infection proposed by the present invention.
[0044] Figure 2 This is a system block diagram of an intelligent detection system for phenotypic characteristics of Cordyceps sinensis larvae infection proposed by the present invention. DETAILED DESCRIPTION
[0045] The present invention will be further described below with reference to the accompanying drawings and examples, but they are not intended to limit the present invention.
[0046] like Figure 1 As shown, the present invention provides an intelligent detection method for the phenotypic characteristics of Cordyceps sinensis larvae infection, comprising the following steps:
[0047] S1: Collect electronic images of Cordyceps sinensis fungi;
[0048] S2: Processing the collected electronic images of Cordyceps sinensis to generate an electronic image positioning dataset of Cordyceps sinensis in the host;
[0049] S3: Establish a Cordyceps sinensis electronic image target detection model, input the Cordyceps sinensis electronic image positioning dataset, train the Cordyceps sinensis electronic image target detection model, and obtain a trained Cordyceps sinensis electronic image target detection model;
[0050] S4: Measure the morphological parameters of the Cordyceps sinensis mycelium image and obtain the growth rate based on the growth time;
[0051] S5: extracting characteristic blocks from the Cordyceps sinensis mycelium image to obtain a characteristic block image of the Cordyceps sinensis mycelium, and obtaining the infection area and infection density of the mycelium inside the insect body;
[0052] S6: Based on the Cordyceps sinensis mycelium feature block image obtained above, a feature dataset is generated based on different image features;
[0053] S7: Establish a classification model, input the obtained mycelium feature block image into the Cordyceps sinensis mycelium classification model, perform mycelium discrimination, and perform grade classification;
[0054] S8: Integrate information and make decisions about appropriate harvesting;
[0055] Information integration:
[0056] Integrate the mycelial growth rate, infection area, infection density and grade classification results.
[0057] Harvest judgment:
[0058] According to the preset harvesting standards (such as the growth rate threshold V_th and the infection density threshold D_th), whether it is suitable for harvesting is determined.
[0059] Output harvest judgment results and provide visual reports.
[0060] In a specific embodiment of the present application, in said S1, the electronic image of Cordyceps sinensis collected includes the morphology, color, texture and distribution characteristics of the hyphae in the host;
[0061] Image acquisition equipment:
[0062] Use high-resolution electron or optical microscopes equipped with digital imaging systems to ensure image clarity and detail capture.
[0063] The collected images include the morphology, color, texture and distribution characteristics of mycelium in the host.
[0064] Collection conditions:
[0065] Shoot under standard lighting conditions to avoid light interference.
[0066] Collect images of the same insect at multiple angles and levels to ensure the comprehensiveness of the data.
[0067] In a specific embodiment of the present application, in said S2, processing the collected electronic image includes image denoising, enhancement, segmentation and annotation operations to produce a high-precision electronic image positioning dataset;
[0068] Image preprocessing:
[0069] Denoising: Use Gaussian filtering or median filtering to remove image noise.
[0070] Enhancement: Enhance image details through histogram equalization or contrast stretching.
[0071] Segmentation: Use threshold segmentation, edge detection or region growing algorithm to separate hyphae from the background.
[0072] Annotation: Manual or semi-automatic annotation of the segmented mycelium area to generate a high-precision positioning dataset.
[0073] Dataset format:
[0074] The dataset includes image files and their corresponding annotation files (such as XML or JSON format). The annotation information includes the location, morphology and category labels of mycelium.
[0075] In a specific embodiment of the present application, in S3, the Cordyceps sinensis electronic image target detection model is constructed based on a deep learning algorithm, including a convolutional neural network or a target detection algorithm.
[0076] Model selection:
[0077] Build object detection models based on deep learning algorithms, such as Faster R-CNN, YOLO, or SSD.
[0078] Use convolutional neural network (CNN) to extract image features.
[0079] Training process:
[0080] Input the electronic image positioning dataset and divide it into training set and validation set.
[0081] Model training is performed using Stochastic Gradient Descent (SGD) or Adam optimizer.
[0082] Adjust hyperparameters through cross-validation to improve model accuracy.
[0083] Model output:
[0084] The trained model can automatically detect and locate the distribution of Cordyceps sinensis mycelium in the host.
[0085] In a specific embodiment of the present application, in S4, the morphological parameter measurements include hypha length, width, number of branches and distribution density, and the growth rate of the hyphae is calculated in combination with the growth time. The calculation formula of the growth rate ν is:
[0086]
[0087] Among them, ΔL is the change in hypha length, and Δt is the growth time.
[0088] In a specific embodiment of the present application, in S5, the extraction of the feature block is achieved by an image segmentation algorithm, including threshold-based segmentation, edge detection or region growing algorithm, to obtain the infection area and infection density of the hyphae inside the insect body. The infection density D is calculated as follows:
[0089]
[0090] Among them, A infected is the mycelial infection area, A total is the total cross-sectional area of the insect body.
[0091] In a specific embodiment of the present application, in S6, the feature data set includes morphological features, color features, texture features and distribution features of mycelium, which are used to train the classification model.
[0092] In a specific embodiment of the present application, in S7, the classification model is constructed based on a machine learning algorithm, including a support vector machine, a random forest or a deep neural network, for discriminating and classifying the mycelium. The discriminant function of the support vector machine is:
[0093]
[0094] Among them, a i is the Lagrange multiplier, y i is the category label, k(x i ,x) is the kernel function and b is the bias term.
[0095] In a specific embodiment of the present application, in S8, the information integration includes mycelial growth rate, infection area, infection density and grade classification results, combined with the preset harvesting criteria, to output a judgment result of suitability for harvesting. The harvesting judgment function is:
[0096]
[0097] Among them, ν threshold is the growth rate threshold, D threshold is the infection density threshold.
[0098] like Figure 2 As shown, the present invention provides an intelligent detection system for the phenotypic characteristics of Cordyceps sinensis larvae infection, comprising:
[0099] An image acquisition module, used for acquiring electronic images of Cordyceps sinensis fungi;
[0100] Image processing module, used to process the collected images and produce positioning data sets;
[0101] The target detection module is used to build and train the target detection model for Cordyceps sinensis electronic images;
[0102] Morphological parameter measurement module, used to measure mycelial morphological parameters and calculate growth rate;
[0103] Feature extraction module, used to extract hyphae feature blocks and obtain infection area and infection density;
[0104] Classification module, used to establish classification models and perform discrimination and grade classification of hyphae;
[0105] The decision-making module is used to integrate information and output the judgment results of suitable harvesting.
[0106] The present invention proposes an intelligent detection method for phenotypic characteristics of Cordyceps sinensis mycelium, which relates to the technical field of Cordyceps sinensis breeding and cultivation. The method comprises collecting electronic images of Cordyceps sinensis mycelium, processing the collected images, producing a positioning data set, establishing a target detection model, and combining an image processing algorithm to extract the mycelium, obtain growth rate results for the extracted mycelium, and mark them, perform antagonism discrimination, extract feature blocks, produce a data set based on different image features, establish a classification model, input the feature block images of Cordyceps sinensis mycelium into a trained classification model, perform mycelial bundle thickness discrimination, colony morphology discrimination, mycelial density discrimination, integrate information, and output statistical results.
[0107] The above description is merely a preferred embodiment of the present application and is not intended to limit the present application. Various modifications and variations are possible for those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present application shall be included within the scope of protection of the present application.
Claims
1. An intelligent detection method for the phenotypic characteristics of Cordyceps sinensis larvae infection, characterized in that: The following steps are involved: S1: Collect electronic images of Cordyceps sinensis; S2: Processing the collected electronic images of Cordyceps sinensis to generate an electronic image positioning dataset of Cordyceps sinensis in the host; S3: Establish a Cordyceps sinensis electronic image target detection model, input the Cordyceps sinensis electronic image positioning dataset, train the Cordyceps sinensis electronic image target detection model, and obtain a trained Cordyceps sinensis electronic image target detection model; S4: Measure the morphological parameters of the Cordyceps sinensis mycelium image and obtain the growth rate based on the growth time; S5: extracting characteristic blocks from the Cordyceps sinensis mycelium image to obtain a characteristic block image of the Cordyceps sinensis mycelium, and obtaining the infection area and infection density of the mycelium inside the insect body; S6: Based on the Cordyceps sinensis mycelium feature block image obtained above, a feature dataset is generated based on different image features; S7: Establish a classification model, input the obtained mycelium feature block image into the Cordyceps sinensis mycelium classification model, perform mycelium discrimination, and perform grade classification; S8: Integrate information and make decisions on appropriate harvesting.
2. The intelligent detection method for phenotypic characteristics of Cordyceps sinensis larvae infection according to claim 1, characterized in that: In said S1, the collected electronic image of Cordyceps sinensis includes the morphology, color, texture and distribution characteristics of the mycelium in the host.
3. The intelligent detection method for phenotypic characteristics of Cordyceps sinensis larvae infection according to claim 1, characterized in that: In S2, the collected electronic image is processed including image denoising, enhancement, segmentation and annotation operations to produce a high-precision electronic image positioning data set.
4. The intelligent detection method for phenotypic characteristics of Cordyceps sinensis larvae infection according to claim 1, characterized in that: In S3, the Cordyceps sinensis electronic image target detection model is constructed based on a deep learning algorithm, including a convolutional neural network or a target detection algorithm.
5. The intelligent detection method for phenotypic characteristics of Cordyceps sinensis larvae infection according to claim 1, characterized in that: In S4, the morphological parameters measured include hypha length, width, number of branches and distribution density, and the growth rate of the hyphae is calculated in combination with the growth time. The calculation formula of the growth rate ν is: Among them, ΔL is the change in hypha length, and Δt is the growth time.
6. The intelligent detection method for phenotypic characteristics of Cordyceps sinensis larvae infection according to claim 1, characterized in that: In S5, the extraction of the feature blocks is achieved by an image segmentation algorithm, including threshold-based segmentation, edge detection, or region growing algorithm, to obtain the infection area and infection density of the hyphae inside the parasite. The infection density D is calculated as follows: Among them, A infected is the mycelial infection area, A total is the total cross-sectional area of the insect body.
7. The intelligent detection method for phenotypic characteristics of Cordyceps sinensis larvae infection according to claim 1, characterized in that: In S6, the feature data set includes the morphological features, color features, texture features and distribution features of the mycelium, which are used to train the classification model.
8. The intelligent detection method for phenotypic characteristics of Cordyceps sinensis larvae infection according to claim 1, characterized in that: In S7, the classification model is constructed based on a machine learning algorithm, including a support vector machine, a random forest or a deep neural network, for discriminating and classifying the mycelium. The discriminant function of the support vector machine is: Among them, a i is the Lagrange multiplier, y i is the category label, k(x i ,x) is the kernel function and b is the bias term.
9. The intelligent detection method for phenotypic characteristics of Cordyceps sinensis larvae infection according to claim 1, characterized in that: In S8, the information integration includes mycelial growth rate, infection area, infection density and grade classification results, combined with the preset harvesting standards, to output the judgment result of whether it is suitable for harvesting. The harvesting judgment function is: Among them, ν threshold is the growth rate threshold, D threshold is the infection density threshold.
10. An intelligent detection system for the phenotypic characteristics of Cordyceps sinensis larvae infection, characterized in that: include: An image acquisition module, used for acquiring electronic images of Cordyceps sinensis fungi; Image processing module, used to process the collected images and produce positioning data sets; The target detection module is used to build and train the target detection model for Cordyceps sinensis electronic images; Morphological parameter measurement module, used to measure mycelial morphological parameters and calculate growth rate; Feature extraction module, used to extract hyphae feature blocks and obtain infection area and infection density; Classification module, used to establish classification models and perform discrimination and grade classification of hyphae; The decision-making module is used to integrate information and output the judgment results of suitable harvesting.