Method and system for identifying maturity of leafy crop plants based on small sample images

Abstract

The application discloses a method and system for identifying leaf crop maturity based on small sample images, mainly relates to the technical field of identifying harvesting maturity, and aims to solve the problems that the existing identification scheme needs more training samples and consumes more manpower and material resources. The method comprises the following steps: extracting the color step mean values of the Red, Green and Blue channels and the gray scale images of the third preset number of leaves of each preset maturity in the sample group, and forming a table corresponding to each preset maturity; based on the color step mean values, performing a set operation on the table corresponding to each preset maturity by using a pixel set method to obtain a set data set of the Red, Green and Blue channels and the gray scale images corresponding to each preset maturity; extracting a feature interval of the set data set; and taking the end point values of the feature interval as training data to obtain a trained Bayesian model.

Description

Technical Field

[0001] This application relates to the field of agricultural product maturity identification technology, and in particular to a method and system for identifying the maturity of leaf crops based on small sample images. Background Technology

[0002] Harvest maturity refers to the degree of maturity at which agricultural products have completed growth and nutrient accumulation and are ready for harvest. It is a crucial basis for determining the timing of harvest. Harvest maturity is closely related to the yield, quality, shelf life, and profit of agricultural products. In field production, harvest maturity identification mainly relies on the harvester's experience, which is highly subjective and significantly affected by factors such as variety, part of the plant, and sunlight (Chen et al., 2019). Therefore, how to objectively and accurately determine the harvest maturity of agricultural products has become a key focus for agricultural product purchasing and processing personnel. Currently, methods for identifying the harvest maturity of agricultural products mainly include chemical detection (Assefa et al., 2019), spectral analysis (Lee et al., 2020), and image recognition (Yuan et al., 2024). However, due to the long processing time of chemical detection and the high cost and inconvenience of spectral analysis equipment, the application scope of both is greatly limited.

[0003] The ripening process of agricultural products is a process of change from the inside out. The ripening state is manifested not only in changes in shape and size but also in the appearance and / or changes in characteristic colors and textures. Color information is a crucial feature parameter for identifying the ripeness of agricultural products, and it can be directly acquired through digital image acquisition. With the development of digital imaging, artificial intelligence, and big data technologies, the integrated use of digital imagery and machine learning techniques has become the mainstream method for intelligent identification of agricultural product ripeness. Furthermore, with the introduction of the big data concept, the requirement for a large number of input samples is increasing, with the belief that the more samples collected, the closer the dataset will be to the real population, and the better the accuracy of the model constructed. However, this brings new research problems: the more samples collected, the more manpower and resources are required, leading to a rapid increase in the cost of building agricultural product harvest ripeness models, thus restricting the further promotion and application of identification patterns. How to achieve the accuracy of a Bayesian model for agricultural product harvest ripeness with a small sample size has become a highly challenging and valuable research topic. Summary of the Invention

[0004] To address the aforementioned shortcomings of existing technologies, this application provides a method and system for recognizing the maturity of leafy crops based on small sample images, thereby solving the problem that existing recognition schemes require a large number of training samples and consume a lot of manpower and resources.

[0005] Firstly, this application provides a method for recognizing the maturity of leafy crops based on small sample images, the method comprising:

[0006] A first preset number of crop plants were collected, and a second preset number of leaves were collected from bottom to top. The leaves were divided into a sample group and a test group. Both the sample group and the test group were labeled with a preset maturity level. The mean values ​​of the Red, Green, and Blue channels and the grayscale image of the leaves of each preset maturity level in the sample group were extracted to form a table corresponding to each preset maturity level. Based on the mean values ​​of the color levels, the pixel union method was used to perform a union operation on the tables corresponding to each preset maturity level to obtain the union dataset of the Red, Green, and Blue channels and the grayscale image at each preset maturity level. The feature intervals of the union dataset were extracted. The endpoint values ​​of the feature intervals were used as training data to train a Bayesian model. The trained Bayesian model was verified using the test group.

[0007] In one implementation of this application, based on the color level mean, a pixel union method is used to perform a union operation on the tables corresponding to each preset maturity level to obtain the union dataset of Red, Green, Blue channels and grayscale images corresponding to each preset maturity level, specifically including:

[0008] Using the [ ] operator in MATLAB software, the mean values ​​of the color levels of the Red, Green, Blue channels and grayscale images corresponding to each preset maturity level are combined to obtain the combined dataset of the Red, Green, Blue channels and grayscale images at each preset maturity level.

[0009] In one implementation of this application, extracting the feature intervals of the union dataset specifically includes:

[0010] The percentiles Q1 and Q3 of the union dataset corresponding to the Red, Green, and Blue channels and grayscale images at various preset maturity levels were extracted using the prctile function in MATLAB software. Based on this, the surface color feature intervals corresponding to the Red, Green, and Blue channels and grayscale images at various preset maturity levels were constructed.

[0011] In one implementation of this application, extracting the feature intervals of the union dataset specifically includes:

[0012] Using the quartile method, obtain the values ​​located at the 25th and 75th percentiles in the union dataset, denoted as Q1 and Q3 respectively;

[0013] Extract the interval [Q1, Q3] of the union dataset as the feature interval of the union dataset.

[0014] In one implementation of this application, the endpoint values ​​of the feature interval are used as training data to train the Bayesian model, specifically including:

[0015] The feature intervals corresponding to the Red, Green, and Blue channels and grayscale images at various preset maturity levels are obtained. The endpoint values ​​of the feature intervals are used as input data, and then a Bayesian decision model is constructed using the Bayesian toolbox.

[0016] In one implementation of this application, the average values ​​of the Red, Green, and Blue channels and the grayscale images of each preset maturity level and a third preset number of leaves in the sample group are extracted and formed into a table, specifically including:

[0017] Extract the average color levels of the Red, Green, and Blue channels and grayscale images of each preset maturity third preset number of leaves in the sample group;

[0018] Extract the average values ​​of the Red, Green, and Blue channels and grayscale images of leaves of each preset maturity level and fourth preset number in the sample group; where the third preset number is less than the fourth preset number.

[0019] Calculate the value difference rate between the average color levels corresponding to the third preset quantity and the fourth preset quantity;

[0020] When the value difference rate is greater than the preset threshold, the average value of the Red, Green, and Blue channels and the grayscale image of each preset maturity third preset number of leaves in the sample group is re-extracted until the value difference rate between the average value of the color levels corresponding to the third preset number and the fourth preset number is less than or equal to the preset threshold.

[0021] Generate a table of the average color levels of the Red, Green, and Blue channels and grayscale images where the difference rate is less than or equal to a preset threshold.

[0022] In one implementation of this application, the trained Bayesian model is verified according to a test group, specifically including:

[0023] The accuracy of the Bayesian model is tested using a test group. When the accuracy is less than the preset accuracy threshold, the third preset number is updated based on the preset growth rate, and the Bayesian model is obtained again until the accuracy is greater than or equal to the preset accuracy threshold, thus obtaining a trained Bayesian model.

[0024] Secondly, this application provides a system for recognizing the maturity of leafy crops based on small sample images. The system includes:

[0025] The data collection module is used to collect a first preset number of crop leaves from bottom to top, and then divide the leaves into a sample group and a test group. Both the sample group and the test group are labeled with a preset maturity level.

[0026] The union acquisition module is used to extract the mean values ​​of the Red, Green, and Blue channels and grayscale images of the third preset number of leaves at each preset maturity level in the sample group, forming a table corresponding to each preset maturity level; based on the mean values ​​of the color levels, the pixel union method is used to perform a union operation on the tables corresponding to each preset maturity level to obtain the union dataset of the Red, Green, and Blue channels and grayscale images at each preset maturity level.

[0027] The model building module is used to extract feature intervals from the union dataset; the endpoints of the feature intervals are used as training data to train the Bayesian model; and the trained Bayesian model is validated using the test group.

[0028] In one implementation of this application, the union acquisition module includes a union unit.

[0029] This method uses the [ ] operator in MATLAB software to perform a union operation on the mean values ​​of the Red, Green, Blue channels and grayscale images corresponding to each preset maturity level, and obtains the union dataset of the Red, Green, Blue channels and grayscale images at each preset maturity level.

[0030] Those skilled in the art will understand that this application has at least the following beneficial effects:

[0031] This application provides a method and system for recognizing the maturity of leafy crops based on small-sample images. Traditional methods typically require a large number of samples for training, leading to high human and material costs. This application significantly reduces the required sample size by collecting only a first preset number of crop leaves and dividing them into sample and test groups, thereby lowering the collection cost. Furthermore, this application involves extracting the RGB channels and grayscale average values ​​of the leaves to form a table corresponding to a preset maturity level, and merging the tables using a pixel union method to obtain feature intervals. This process simplifies the data preprocessing steps and improves the efficiency of model construction. Additionally, by training a Bayesian model using an interval-based decision method based on feature intervals, this application can better capture the key differences between leaves of different maturity levels, thereby enhancing the model's generalization ability. This allows the model to achieve high recognition accuracy even with a small sample size.

[0032] That is, this application uses RGB digital image pixels as the analysis unit and adopts the union operation to reconstruct the group information data set, so as to obtain the true and accurate surface color information of agricultural product groups to the greatest extent while reducing the number of samples (cost), thereby constructing a high-precision agricultural product maturity determination model.

[0033] Furthermore, to avoid insufficient sample size affecting model accuracy, this application introduces a dynamic optimization mechanism. When the model accuracy tested by the test group falls below a preset accuracy threshold, the first preset sample size is updated according to a preset growth rate, and the model is retrained. This mechanism ensures that the model can gradually improve its accuracy through continuous iteration until it meets application requirements. Attached Figure Description

[0034] The following description refers to some embodiments of this disclosure, in which:

[0035] Figure 1 This is a flowchart of a method for recognizing the maturity of leafy crops based on small sample images, provided in an embodiment of this application.

[0036] Figure 2 This is a schematic diagram of the internal structure of a system for recognizing the maturity of leafy crops based on small sample images, provided in an embodiment of this application. Detailed Implementation

[0037] Those skilled in the art should understand that the embodiments described below are merely preferred embodiments of this disclosure and do not imply that this disclosure can only be implemented through these preferred embodiments. These preferred embodiments are merely used to explain the technical principles of this disclosure and are not intended to limit the scope of protection of this disclosure. Based on the preferred embodiments provided by this disclosure, all other embodiments obtained by those skilled in the art without creative effort should still fall within the scope of protection of this disclosure.

[0038] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0039] Achieving accuracy in Bayesian models of agricultural product harvest maturity under small sample sizes has become a highly challenging and valuable research topic. The primary requirement for solving this problem is for researchers to extract as much population characteristic information as possible from small samples. Since digital images are composed of pixels, even physically separate agricultural products exhibit underlying relationships. Therefore, we consider whether we can reconstruct the population information dataset from the pixel-level of digital images of agricultural products. This would allow us to obtain the most accurate and realistic surface color information of the agricultural product population while reducing the number of samples (and thus the cost), thereby constructing a more accurate agricultural product maturity determination model.

[0040] Tobacco is a widely used model plant in genetics and ecology research. Due to its large leaf area, relatively flat surface, and the distinct color change from green to yellow during ripening, it is an ideal material for acquiring surface color information and analyzing maturity. This can provide valuable reference for constructing Bayesian models of harvest maturity for other leafy agricultural products. The Bayesian classifier, based on the Naive Bayes theorem, is a probabilistic classifier that outperforms multilayer perceptrons, fuzzy logic, and principal component analysis methods when used to classify the color features of digital images of agricultural products. Therefore, this application takes fresh tobacco leaves as the research object, and obtains digital information on the surface color of fresh tobacco leaves at three maturity levels: underripe, mature, and overripe. The pixel union method is used to obtain datasets of color information of fresh tobacco leaves at different maturity levels, and the statistical characteristics of these datasets are analyzed. Then, three methods are used to construct tobacco maturity classification models with different sample sizes: data fitting method, single leaf set interval determination method, and pixel union interval determination method. By comparing the accuracy of the model determination, the advantages of the pixel union interval determination method are clarified, the construction mechanism of the accurate model of harvest maturity of small sample agricultural products is explained, and a research case is provided.

[0041] The technical solutions proposed in the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0042] The embodiment provides a method for identifying the maturity of leafy crops based on small sample images, such as Figure 1 As shown in the embodiments of this application, the method mainly includes the following steps:

[0043] Step 110: Collect a first preset number of crop leaves from bottom to top, and then divide the leaves into a sample group and a test group.

[0044] It should be noted that both the sample group and the test group were labeled with preset maturity levels. The number of leaves was the same for each preset maturity level in the sample group.

[0045] In some embodiments, this step may involve selecting 50 tobacco plants that are growing normally and developing uniformly, collecting leaves from the 6th to 11th leaf positions from the bottom (6 leaves per plant), removing damaged or missing single leaves, totaling 272 leaves, and randomly dividing the tobacco leaves of different maturity levels into two groups. The first group serves as the modeling data source, including 50 samples each of immature, mature, and overripe (a total of 150 samples), and the second group serves as the test data source, including 28 immature samples, 50 mature samples, and 44 overripe samples (a total of 122 samples).

[0046] Step 120: Extract the mean values ​​of the Red, Green, and Blue channels and grayscale images of the third preset number of leaves at each preset maturity level in the sample group to form a table corresponding to each preset maturity level; based on the mean values ​​of the color levels, perform a pixel union operation on the table corresponding to each preset maturity level to obtain the union dataset of the Red, Green, and Blue channels and grayscale images at each preset maturity level.

[0047] It should be noted that this step can specifically involve using the `mean` function in MATLAB to extract the mean values ​​of the Red, Green, and Blue channels and grayscale images of all collected leaves. The mean values ​​of the Red, Green, and Blue channels and grayscale images of each maturity level are randomly selected from the sample group for a third preset number (e.g., 5 single leaves) to form a table corresponding to each preset maturity level. Then, using the "[ ]" operator in MATLAB, a union operation is performed on a second preset number of leaves for each maturity level to obtain a group pixel union dataset of the three color channels and grayscale images under the three maturity levels, forming a union dataset.

[0048] In addition, to reduce the contamination of faulty data during the extraction process, this application can perform data verification by extracting the average values ​​of the Red, Green, and Blue channels and grayscale images of the third preset number of leaves at each preset maturity level in the sample group and forming a table. The specific process can be as follows:

[0049] Extract the average color levels of the Red, Green, and Blue channels and grayscale images of each preset maturity third preset number of leaves in the sample group;

[0050] Extract the average values ​​of the Red, Green, and Blue channels and grayscale images of leaves of each preset maturity level and fourth preset number in the sample group; where the third preset number is less than the fourth preset number.

[0051] Calculate the value difference rate between the average color levels corresponding to the third preset quantity and the fourth preset quantity;

[0052] When the value difference rate is greater than the preset threshold, the average value of the Red, Green, and Blue channels and the grayscale image of each preset maturity third preset number of leaves in the sample group is re-extracted until the value difference rate between the average value of the color levels corresponding to the third preset number and the fourth preset number is less than or equal to the preset threshold.

[0053] Generate a table of the average color levels of the Red, Green, and Blue channels and grayscale images where the difference rate is less than or equal to a preset threshold.

[0054] Those skilled in the art will understand that, regardless of whether the number of samples is 5, 10, or 30, the mean values ​​of the Red channel, Green channel, Blue channel, and Gray image obtained after pixel union are all within 5% of the corresponding feature values ​​after sample union, which is a small difference. Therefore, this small difference can be used to verify whether the current data is mixed with abnormal data.

[0055] Step 130: Extract the feature intervals of the union dataset; use the endpoints of the feature intervals as training data to train the Bayesian model, and verify the trained Bayesian model based on the test group.

[0056] It should be noted that extracting the feature intervals from the union dataset can specifically be done as follows:

[0057] The percentiles Q1 and Q3 of the union dataset corresponding to the Red, Green, and Blue channels and grayscale images at various preset maturity levels were extracted using the prctile function in MATLAB software. Based on this, the surface color feature intervals corresponding to the Red, Green, and Blue channels and grayscale images at various preset maturity levels were constructed.

[0058] In addition, Q1 and Q3 can be determined by those skilled in the art based on the actual situation.

[0059] As an example, the methods for determining Q1 and Q3 can be:

[0060] Using the quartile method, the 25th and 75th percentile values ​​of the union dataset are obtained and denoted as Q1 and Q3, respectively; the interval [Q1, Q3] of the union dataset is extracted as the feature interval of the union dataset.

[0061] Specifically, the Bayesian model is trained using the endpoint values ​​of the feature interval as training data, which can be done as follows:

[0062] The feature intervals corresponding to the Red, Green, and Blue channels and grayscale images at various preset maturity levels are obtained. The endpoint values ​​of the feature intervals are used as input data, and then a Bayesian decision model is constructed using the Bayesian toolbox.

[0063] The verification of the trained Bayesian model, based on the test group, specifically includes:

[0064] The accuracy of the Bayesian model is tested using a test group. When the accuracy is less than the preset accuracy threshold, the third preset number is updated based on the preset growth rate, and the Bayesian model is obtained again until the accuracy is greater than or equal to the preset accuracy threshold, thus obtaining a trained Bayesian model.

[0065] Furthermore, this application can update the Bayesian model based on its accuracy. Specifically, the accuracy of the Bayesian model is tested using a test group. When the accuracy is less than a preset accuracy threshold, a third preset number of updates is made based on a preset growth rate, and the trained Bayesian model is obtained again until the accuracy is greater than or equal to the preset accuracy threshold.

[0066] It should be noted that the preset growth rate can be 10.

[0067] Those skilled in the art will understand that, to avoid insufficient samples affecting model accuracy, this application introduces a dynamic optimization mechanism. Specifically, when the model accuracy tested by the test group is less than a preset accuracy threshold, the first preset number of samples is updated according to a preset growth rate, and the model is retrained. This mechanism ensures that the model can gradually improve its accuracy through continuous iteration until it meets application requirements.

[0068] In addition, this application Figure 2 This application provides a system for recognizing the maturity of leafy crops based on small sample images. For example... Figure 2 As shown in the embodiments of this application, the system mainly includes:

[0069] The acquisition module 210 is used to acquire a first preset number of crop leaves from bottom to top, and to divide the leaves into a sample group and a test group. The sample group and the test group are labeled with a preset maturity level.

[0070] The union acquisition module 220 is used to extract the mean values ​​of the Red, Green, and Blue channels and grayscale images of each preset maturity third preset number of leaves in the sample group, forming a table corresponding to each preset maturity; based on the mean values ​​of the color levels, the pixel union method is used to perform a union operation on the tables corresponding to each preset maturity, to obtain the union dataset of the Red, Green, and Blue channels and grayscale images corresponding to each preset maturity.

[0071] Union acquisition module 220 includes union unit,

[0072] This method uses the [ ] operator in MATLAB software to perform a union operation on the mean values ​​of the Red, Green, Blue channels and grayscale images corresponding to each preset maturity level, and obtains the union dataset of the Red, Green, Blue channels and grayscale images at each preset maturity level.

[0073] The model building module 230 is used to extract the feature intervals of the union dataset; the endpoint values ​​of the feature intervals are used as training data to train the Bayesian model; and the trained Bayesian model is verified according to the test group.

[0074] The technical solutions of this disclosure have been described in conjunction with the preceding embodiments. However, it will be readily understood by those skilled in the art that the scope of protection of this disclosure is not limited to these specific embodiments. Without departing from the technical principles of this disclosure, those skilled in the art can disassemble and combine the technical solutions in the above embodiments, and can also make equivalent changes or substitutions to the relevant technical features. Any changes, equivalent substitutions, improvements, etc., made within the technical concept and / or technical principles of this disclosure will fall within the scope of protection of this disclosure.

Claims

1. A method for recognizing the maturity of leafy crops based on small sample images, characterized in that, The method includes: Collect a first preset number of crop plants and a second preset number of leaves from bottom to top; divide the leaves into a sample group and a test group; both the sample group and the test group are labeled with a preset maturity level; Extract the mean values ​​of the Red, Green, and Blue channels and grayscale images of the leaves at each preset maturity level in the sample group, forming a table corresponding to each preset maturity level. Based on the mean values ​​of the color levels, perform a pixel union operation on the tables corresponding to each preset maturity level to obtain the union dataset of the Red, Green, and Blue channels and grayscale images at each preset maturity level. Specifically, this includes: Using the [ ] operator in MATLAB software, the mean values ​​of the color levels of the Red, Green, Blue channels and grayscale images corresponding to each preset maturity level are combined to obtain the combined dataset of the Red, Green, Blue channels and grayscale images at each preset maturity level. Extracting feature intervals from the union dataset; specifically, using the prctile function in MATLAB software to extract the percentiles Q1 and Q3 of the union dataset corresponding to the Red, Green, Blue channels and grayscale images at each preset maturity level, and using this to construct the surface color feature intervals corresponding to the Red, Green, Blue channels and grayscale images at each preset maturity level; Specifically, extracting feature intervals from the union dataset includes: Using the quartile method, obtain the values ​​at the 25th and 75th percentiles in the union dataset, denoted as Q1 and Q3 respectively; extract the interval [Q1, Q3] of the union dataset as the feature interval of the union dataset; The endpoint values ​​of the feature interval are used as training data to train the Bayesian model. The trained Bayesian model is then validated using the test group.

2. The method for recognizing the maturity of leafy crops based on small sample images according to claim 1, characterized in that, Using the endpoints of the feature interval as training data, a Bayesian model is trained, specifically including: The feature intervals corresponding to the Red, Green, and Blue channels and grayscale images at various preset maturity levels are obtained. The endpoint values ​​of the feature intervals are used as input data, and then a Bayesian decision model is constructed using the Bayesian toolbox.

3. The method for recognizing the maturity of leafy crops based on small sample images according to claim 1, characterized in that, Extract the mean values ​​of the Red, Green, and Blue channels and grayscale images of leaves at each preset maturity level and third preset number in the sample group, and form a table, specifically including: Extract the average color levels of the Red, Green, and Blue channels and grayscale images of each preset maturity third preset number of leaves in the sample group; Extract the average values ​​of the Red, Green, and Blue channels and grayscale images of leaves of each preset maturity level and fourth preset number in the sample group; where the third preset number is less than the fourth preset number. Calculate the value difference rate between the average color levels corresponding to the third preset quantity and the fourth preset quantity; When the value difference rate is greater than the preset threshold, the average value of the Red, Green, and Blue channels and the grayscale image of each preset maturity third preset number of leaves in the sample group is re-extracted until the value difference rate between the average value of the color levels corresponding to the third preset number and the fourth preset number is less than or equal to the preset threshold. Generate a table of the average color levels of the Red, Green, and Blue channels and grayscale images where the difference rate is less than or equal to a preset threshold.

4. The method for recognizing the maturity of leafy crops based on small sample images according to claim 1, characterized in that, Based on the test group, the trained Bayesian model was validated, specifically including: The accuracy of the Bayesian model is tested using a test group. When the accuracy is less than the preset accuracy threshold, the third preset number is updated based on the preset growth rate, and the Bayesian model is obtained again until the accuracy is greater than or equal to the preset accuracy threshold, thus obtaining a trained Bayesian model.

5. A system for recognizing the maturity of leafy crops based on small sample images, characterized in that, The system includes: The data collection module is used to collect a first preset number of crop leaves from bottom to top, and then divide the leaves into a sample group and a test group. Both the sample group and the test group are labeled with a preset maturity level. The union acquisition module is used to extract the mean values ​​of the Red, Green, and Blue channels and grayscale images of the third preset number of leaves at each preset maturity level in the sample group, forming a table corresponding to each preset maturity level; based on the mean values ​​of the color levels, the pixel union method is used to perform a union operation on the tables corresponding to each preset maturity level to obtain the union dataset of the Red, Green, and Blue channels and grayscale images at each preset maturity level. The union acquisition module includes union units. This is used to perform a union operation on the mean values ​​of the Red, Green, Blue channels and grayscale images corresponding to each preset maturity level using the [ ] operator in MATLAB software, to obtain the union dataset of the Red, Green, Blue channels and grayscale images at each preset maturity level; The model building module is used to extract feature intervals from the union dataset. Specifically, it includes: using the prctile function in MATLAB software to extract the percentiles Q1 and Q3 of the union dataset corresponding to the Red, Green, Blue channels and grayscale images at various preset maturity levels, and using this to construct the surface color feature intervals corresponding to the Red, Green, Blue channels and grayscale images at various preset maturity levels; the extraction of feature intervals from the union dataset specifically includes: using the quartile method to obtain the values ​​located at the 25th and 75th percentiles in the union dataset, denoted as Q1 and Q3 respectively; extracting the interval [Q1, Q3] of the union dataset as the feature interval of the union dataset; using the endpoint values ​​of the feature interval as training data to train the Bayesian model, and verifying the obtained trained Bayesian model based on the test group.

Images