Data analysis and processing methods of corn seed stress resistance
Through environmental sensors and high-resolution imaging technology combined with machine learning models, Bayesian networks are used to establish a stress resistance prediction model, and the planting strategy is dynamically adjusted, which solves the problem of long time and high cost of evaluating crop stress resistance in the existing technology. Fast and accurate stress resistance assessment and planting strategy optimization are achieved.
Patent Information
- Application Number
- CN202411111175.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-14
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2044-08-14
AI Technical Summary
The prior art has shortcomings in rapid and accurate assessment and improvement of crop stress resistance. Traditional methods require years of cycles and rely on expensive bioinformatics analysis equipment to adjust planting strategies in real time to cope with environmental changes.
The corn seed stress resistance data analysis and processing method are used to collect data through environmental sensors, combine high-resolution imaging and machine learning models, and use Bayesian network to establish a stress resistance prediction model, and dynamically adjust the planting strategy.
The rapid and accurate assessment of the stress resistance of corn seeds is achieved, which reduces the evaluation time and cost, improves the practicality of the planting strategy and the adaptability of crops, and enhances stress resistance and production stability.
Smart Images

Figure CN118626810B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of agricultural science and technology, and in particular to a method for analyzing and processing corn seed stress resistance data. Background Art
[0002] Corn is one of the most important crops in global food crop production. Its yield and quality are affected by a variety of environmental factors, especially stress resistance, which is the ability of corn to maintain growth and yield when encountering abiotic stress (such as drought, high salt and extreme temperature) and biotic stress (such as disease and insect pests). Traditional methods for improving seed stress resistance include field trials and phenotypic screening, which usually require multi-year cycles and are not always effective under rapidly changing environmental conditions. In addition, these methods often require a lot of manpower and material resources, making them costly in actual operations.
[0003] Existing technologies have obvious shortcomings in quickly and accurately evaluating and improving crop stress resistance. Although genetic markers and genomic selection technologies have made progress in crop improvement, these methods rely on complex bioinformatics analysis and expensive technical equipment, which is not conducive to the rapid application of research results to a wide range of agricultural production practices. In addition, traditional methods find it difficult to take into account immediate changes in environmental factors, thereby adjusting planting strategies in real time to maximize crop yield and quality.
[0004] Therefore, there is an urgent need to develop a new method that can integrate phenotypic characteristics, environmental data and advanced data analysis technology to achieve rapid and accurate assessment of crop stress resistance and dynamic optimization of planting strategies to adapt to rapidly changing agricultural needs and environmental pressures. Summary of the invention
[0005] Based on the above purpose, the present invention provides a method for analyzing and processing corn seed stress resistance data.
[0006] The method for analyzing and processing corn seed stress resistance data comprises the following steps:
[0007] S1: Collect corn seed samples from different regions and use environmental sensors to collect environmental data related to seed growth, including temperature, humidity, light intensity, and soil salt concentration;
[0008] S2: performing non-destructive phenotypic imaging on the sample, and obtaining the growth status and morphological characteristics of the seeds using high-resolution imaging technology;
[0009] S3: inputting the phenotypic imaging data and the environmental data into a preset machine learning model, wherein the machine learning model is composed of a convolutional neural network and a recurrent neural network, and is used to analyze the association between the phenotypic characteristics of the seeds and the environmental data;
[0010] S4: Using the Bayesian network, identify the features related to stress resistance from the correlation data obtained in S3 and establish a stress resistance prediction model;
[0011] S5: using the stress resistance prediction model established in S4, predicting the stress resistance of the newly collected corn seed samples and generating prediction results;
[0012] S6: Based on the prediction results of S5, dynamically adjust the planting strategy of corn seeds.
[0013] Optionally, the S1 specifically includes:
[0014] S11: Determine the target collection area and select representative climate zones through geographic information system analysis, including high temperature zone, low temperature zone, high humidity zone, low humidity zone, high light zone and low light zone;
[0015] S12: Establish multiple sampling points in each target collection area, and collect a predetermined number of corn seed samples at each sampling point by random sampling method;
[0016] S13: Install a variety of environmental sensors at each sampling point, including temperature sensors, humidity sensors, light intensity sensors, and soil salt concentration sensors, to continuously monitor and record environmental data;
[0017] S14: The data collected by the environmental sensor is transmitted to the central database in real time through wireless transmission technology;
[0018] S15: Assign a unique identifier to each sampling point in the central database, associate the identifier with the corresponding corn seed sample and environmental data, and annotate the environmental data of each seed sample through a data matching algorithm.
[0019] Optionally, the S2 specifically includes:
[0020] S21: Pre-treat the collected corn seed samples, including cleaning and drying, to ensure that there are no impurities on the surface of the samples;
[0021] S22: photographing the sample at multiple angles using a high-resolution imaging device to obtain image data at different angles. The high-resolution imaging device includes a high-resolution digital camera and a near-infrared imager;
[0022] S23: pre-processing the image data captured in S22 using an image processing algorithm, including denoising, contrast enhancement, and color correction;
[0023] S24: Using image segmentation technology, the growth status and morphological characteristics of corn seeds are separated from the background;
[0024] S25: Quantitatively analyze the segmented images using computer vision algorithms to extract phenotypic characteristics of corn seeds, including seed size, shape, color, surface texture, and internal structure;
[0025] S26: The extracted phenotypic characteristic data is stored in a central database and associated with the corresponding seed sample unique identifier.
[0026] Optionally, the S25 includes:
[0027] S251: using an edge detection algorithm to extract the outline and edge of the seed to obtain the size characteristics of the seed, wherein the edge detection algorithm adopts Canny edge detection;
[0028] S252: Extracting the shape features of seeds through morphological operations, including erosion and dilation;
[0029] S253: using a color analysis algorithm to obtain color features of the seeds, specifically including converting the image into an HSV color space;
[0030] S254: Calculate the surface texture features of seeds using gray-level co-occurrence matrix;
[0031] S255: Obtain the internal structural characteristics of seeds through image reconstruction technology.
[0032] Optionally, the S3 specifically includes:
[0033] S31: Preprocess the acquired phenotypic imaging data and environmental data, including data normalization and missing value filling;
[0034] S32: Input the preprocessed phenotypic imaging data into a convolutional neural network, and extract high-order phenotypic characteristics of the seeds through multi-layer convolution and pooling operations. The convolutional neural network specifically includes a convolution layer, an activation layer, a pooling layer, and a fully connected layer;
[0035] S33: inputting the preprocessed environmental data into a recursive neural network, and analyzing the time series characteristics of the environmental data through a cyclic calculation of time steps, wherein the recursive neural network specifically includes an input layer, a hidden layer, and an output layer;
[0036] S34: Fusing the high-order phenotypic features extracted by the convolutional neural network and the time series features analyzed by the recurrent neural network to form a comprehensive feature vector;
[0037] S35: The comprehensive feature vector is input into a fully connected neural network, and the association between the phenotypic characteristics of the seeds and the environmental data is analyzed through multi-layer neuron calculations and activation functions. The fully connected neural network specifically includes an input layer, a hidden layer, and an output layer.
[0038] Optionally, the S35 specifically includes:
[0039] S351: Input the comprehensive feature vector into the input layer of the fully connected neural network, each neuron in the input layer receives a feature value and passes it to the hidden layer;
[0040] S352: In the hidden layer, each neuron performs a weighted summation of the received input features and adds a bias term. The calculation formula is: ,in, For the The weighted sum of the hidden layer neurons, For input features To The weights of the neurons in the hidden layer, For the The bias term of the hidden layer neurons;
[0041] S353: Input the calculated weighted sum into an activation function to obtain an activation output, where the activation function is ReLU and the calculation formula is: ,in, For the The activation output of the hidden layer neurons;
[0042] S354: passing the activation output of the hidden layer to the next hidden layer or output layer, and repeating the calculation process of S352 and S353;
[0043] S355: In the output layer, each neuron receives the activation output from the last hidden layer and calculates the final output value through weighted summation and activation function. The activation function of the output layer is a linear activation function or a Softmax function, which is used for classification or regression tasks.
[0044] S356: Calculate the error between the predicted value of the output layer and the actual observed data, and use the mean square error or cross entropy loss function to calculate the error;
[0045] S357: Adjust the model parameters through the back propagation algorithm to minimize the error. The back propagation algorithm includes calculating the gradient of the loss function for each parameter and updating the parameters according to the gradient descent method. The update formula is:
[0046] ,in, are model parameters, is the learning rate, is the loss function.
[0047] Optionally, the S4 specifically includes:
[0048] S41: Input the comprehensive feature vector extracted in S3 into the Bayesian network to initialize the structure and parameters of the Bayesian network. The Bayesian network consists of nodes and edges. The nodes represent variables, and the edges represent the conditional dependency relationship between variables.
[0049] S42: Constructing Bayesian network structures using prior knowledge and data-driven methods. The data-driven methods are structure learning algorithms, including K2 algorithms or greedy search algorithms.
[0050] S43: According to the constructed Bayesian network structure, the conditional probability distribution of each node in the network is calculated, and the parameters of the conditional probability distribution are estimated by maximum likelihood estimation. The specific formula is:
[0051] ,in, Indicates simultaneous occurrence and its parent node, Indicates the number of times the parent node appears;
[0052] S44: Inference is performed in the Bayesian network. The conditional probability of each feature in the comprehensive feature vector for the stress resistance of corn seeds is calculated by the Bayesian reasoning method. The reasoning formula is:
[0053] ,in, For under evidence The posterior probability of For The probability of evidence under the condition, for The prior probability of for the probability of evidence;
[0054] S45: Using the identified features, establish a stress resistance prediction model.
[0055] Optionally, the S45 specifically includes:
[0056] S451: Select the identified features as input nodes in the Bayesian network, and specifically represent the feature variables as , where each Represents a feature;
[0057] S452: Set seed stress resistance as the target variable , and use it as the output node in the Bayesian network;
[0058] S453: Calculate the conditional probability distribution between each input feature and the target variable. The calculation formula of the conditional probability distribution is: ,in, For a given feature Conditional target variable The posterior probability of For the target variable Conditional Features The conditional probability of is the prior probability of the target variable, is the joint probability of the features;
[0059] S454: estimating the conditional probability table of each node in the Bayesian network using the maximum likelihood estimation method;
[0060] S455: Calculate the target variable under the predetermined input feature combination using the estimated conditional probability table through the Bayesian inference method The predicted value of , the formula of Bayesian inference is:
[0061] ,in, is the target variable under the evidence condition The posterior probability of For the target variable Conditional probability of evidence given, is the prior probability of the target variable, for the probability of evidence;
[0062] S456: The final stress resistance prediction model is expressed as:
[0063] ,in, For a given feature Conditional target variable The posterior probability represents the prediction result of stress resistance.
[0064] Optionally, the S5 specifically includes:
[0065] S51: Collect new corn seed samples and obtain their phenotypic characteristic data through preprocessing and imaging steps, and record corresponding environmental data;
[0066] S52: inputting the phenotypic characteristic data and environmental data of the newly collected corn seed samples into the final stress resistance prediction model;
[0067] S53: In the final stress resistance prediction model, conditional probability calculation is performed on the input phenotypic characteristic data and environmental data, and the conditional probability between each characteristic and the stress resistance target variable is calculated using the conditional probability table determined in the model;
[0068] S54: Through Bayesian reasoning, the conditional probabilities of all input features are integrated to generate the stress resistance prediction results of the newly collected samples;
[0069] S55: post-processing the generated stress resistance prediction results, including determining a final stress resistance score or classification label;
[0070] S56: The stress resistance prediction result is stored in a database and associated with the identifier of the corresponding newly collected sample.
[0071] Optionally, the S6 specifically includes:
[0072] S61: analyzing the prediction results of corn seed stress resistance obtained in S5, determining the stress resistance levels of various subsamples, and classifying the results into high stress resistance, medium stress resistance, and low stress resistance;
[0073] S62: Correlate predicted stress resistance results with environmental data of seed samples to identify seed characteristics that exhibit high stress resistance;
[0074] S63: Based on the analysis results of S62, adjust the corn seed planting strategy, including selecting seed varieties suitable for high stress resistance and adjusting the planting density and planting period.
[0075] Beneficial effects of the present invention:
[0076] The present invention provides an innovative corn seed stress resistance data analysis and processing method by integrating phenotypic imaging data, environmental monitoring data and advanced Bayesian network technology. This method can quickly and accurately evaluate the stress resistance of corn seeds under various environmental conditions. Compared with traditional methods that rely on laboratory or field tests, the present invention significantly reduces evaluation time and cost, and improves operational efficiency and practicality. In addition, through real-time data analysis, it can more accurately capture the impact of environmental changes on seed stress resistance, thereby optimizing planting strategies and improving the overall adaptability and yield of crops.
[0077] The present invention, through the dynamic planting strategy adjustment function, enables agricultural producers to adjust planting behavior according to real-time analysis results and optimize resource allocation. This method not only enhances the stress resistance and production stability of crops, but also provides strong technical support for addressing the agricultural production challenges brought about by global climate change. Therefore, the application of the present invention can help agricultural producers reduce crop losses caused by environmental factors and increase economic benefits. It is also of great significance to ensuring global food security. BRIEF DESCRIPTION OF THE DRAWINGS
[0078] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0079] Figure 1 Schematic diagram of a method for analyzing and processing corn seed stress resistance data according to an embodiment of the present invention;
[0080] Figure 2 The figure is a schematic diagram of the process of obtaining the growth status and morphological characteristics of seeds according to an embodiment of the present invention. DETAILED DESCRIPTION
[0081] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.
[0082] It should be noted that the references to "one embodiment", "embodiment", "exemplary embodiments", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).
[0083] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.
[0084] like Figure 1-Figure 2 As shown, the corn seed stress resistance data analysis and processing method comprises the following steps:
[0085] S1: Collect corn seed samples from different regions and use environmental sensors to collect environmental data related to seed growth, including temperature, humidity, light intensity, and soil salt concentration;
[0086] S2: Perform non-destructive phenotypic imaging of the samples, using high-resolution imaging technology to obtain the growth status and morphological characteristics of the seeds;
[0087] S3: The phenotypic imaging data and environmental data are input into a preset machine learning model, which consists of a convolutional neural network (CNN) and a recurrent neural network (RNN) to analyze the association between the phenotypic characteristics of seeds and environmental data;
[0088] S4: Using the Bayesian network, identify the features related to stress resistance from the correlation data obtained in S3 and establish a stress resistance prediction model;
[0089] S5: using the stress resistance prediction model established in S4, predicting the stress resistance of the newly collected corn seed samples and generating prediction results;
[0090] S6: Based on the prediction results of S5, dynamically adjust the planting strategy of corn seeds.
[0091] S1 specifically includes:
[0092] S11: Determine the target collection area and select representative climate zones through geographic information system (GIS) analysis, including high temperature zone, low temperature zone, high humidity zone, low humidity zone, high light zone and low light zone;
[0093] S12: Multiple sampling points are set up in each target collection area, and a predetermined number of corn seed samples are collected at each sampling point by random sampling method to ensure the representativeness of the samples;
[0094] S13: Install a variety of environmental sensors at each sampling point, including temperature sensors, humidity sensors, light intensity sensors, and soil salt concentration sensors, to continuously monitor and record environmental data;
[0095] S14: Through wireless transmission technology, the data collected by environmental sensors are transmitted to the central database in real time to ensure the timeliness and accuracy of the data;
[0096] S15: Assign a unique identifier to each sampling point in the central database, and associate the identifier with the corresponding corn seed sample and environmental data, and annotate the environmental data of each seed sample through a data matching algorithm to ensure data consistency and traceability; the above steps can ensure the breadth and representativeness of the samples by using a geographic information system to select sampling areas and set up multiple sampling points, and can obtain accurate and timely environmental data by installing environmental sensors at the sampling points and transmitting data in real time, making subsequent data analysis more reliable.
[0097] S2 specifically includes:
[0098] S21: Pre-process the collected corn seed samples, including cleaning and drying, to ensure that the sample surface is free of impurities and is in a suitable imaging state;
[0099] S22: photographing the sample at multiple angles using a high-resolution imaging device to obtain image data at different angles. The high-resolution imaging device includes a high-resolution digital camera and a near-infrared imager;
[0100] S23: pre-processing the image data captured in S22 using an image processing algorithm, including denoising, contrast enhancement, and color correction;
[0101] S24: Use image segmentation technology to separate the growth status and morphological characteristics of corn seeds from the background to ensure the clarity and accuracy of image data;
[0102] S25: Quantitatively analyze the segmented images using computer vision algorithms to extract phenotypic characteristics of corn seeds, including seed size, shape, color, surface texture, and internal structure;
[0103] S26: The extracted phenotypic characteristic data are stored in a central database and associated with the corresponding seed sample unique identifier to ensure data consistency and traceability.
[0104] The specific steps of quantitative analysis in S25 include:
[0105] S251: Use edge detection algorithm to extract the outline and edge of the seed to obtain the size characteristics of the seed. The edge detection algorithm uses Canny edge detection. The specific formula is: ,in, is the gradient amplitude at the pixel point, and are the gradients in the horizontal and vertical directions respectively;
[0106] S252: Extract the shape features of the seeds through morphological operations, including corrosion and expansion. The formula for corrosion operation is: ,in, is the input image, is the structural element, Indicates location, express In Location The translation of; the expansion operation formula is: ,in, express Reflection;
[0107] S253: Using a color analysis algorithm to obtain the color characteristics of the seed, specifically including converting the image into the HSV color space and calculating the hue, saturation and brightness by the following formula:
[0108] ; ; ,in, is the chromaticity, is saturation, is brightness, It represents the difference between the maximum and minimum values. is the red channel value, is the green channel value, is the blue channel value;
[0109] S254: Calculate the surface texture features of the seeds using a gray level co-occurrence matrix (GLCM), specifically including the following steps:
[0110] a. Calculate the gray-level co-occurrence matrix:
[0111] ;
[0112] in, Grayscale and The probability of symbiosis, and is the pixel position, is the gray value of the pixel, and is the pixel pitch;
[0113] b. Extract texture features from the gray-level co-occurrence matrix, including energy, contrast, entropy, and homogeneity. The specific formula is as follows:
[0114] energy ;
[0115] Contrast ;
[0116] entropy ;
[0117] Homogeneity ,in, is the gray level of the image, is the gray level in the gray level co-occurrence matrix and The probability of symbiosis;
[0118] S255: The internal structural characteristics of the seeds are obtained through image reconstruction technology, specifically using tomography reconstruction, the formula is: ,in, is the reconstructed image, is the projection data, is the projection angle, is the projection position;
[0119] Through the above detailed quantitative analysis steps, the key phenotypic characteristics of corn seeds can be extracted and analyzed more accurately. The combination of edge detection, morphological operations, color analysis, gray-level co-occurrence matrix and tomography reconstruction algorithms makes the extracted phenotypic characteristic data more comprehensive and accurate, providing a solid foundation for subsequent stress resistance data analysis.
[0120] S3 specifically includes:
[0121] S31: Preprocess the acquired phenotypic imaging data and environmental data, including data normalization and missing value filling, to ensure data consistency and integrity;
[0122] S32: The preprocessed phenotypic imaging data is input into a convolutional neural network (CNN) to extract high-order phenotypic features of the seeds through multi-layer convolution and pooling operations. The convolutional neural network specifically includes a convolution layer, an activation layer, a pooling layer, and a fully connected layer;
[0123] S33: inputting the preprocessed environmental data into a recursive neural network (RNN), and analyzing the time series characteristics of the environmental data through cyclic calculation of time steps, wherein the recursive neural network specifically includes an input layer, a hidden layer, and an output layer;
[0124] S34: Fusing the high-order phenotypic features extracted by the convolutional neural network and the time series features analyzed by the recursive neural network to form a comprehensive feature vector, which is fused by feature connection or feature weighting;
[0125] S35: Input the comprehensive feature vector into a fully connected neural network (FCNN), and analyze the association between the phenotypic characteristics of the seeds and the environmental data through multi-layer neuron calculations and activation functions. The fully connected neural network specifically includes an input layer, a hidden layer, and an output layer. The above steps input the phenotypic imaging data and the environmental data into a machine learning model composed of a convolutional neural network and a recurrent neural network. The present invention can comprehensively analyze and extract high-order phenotypic characteristics and time series characteristics of corn seeds, and achieve high-precision prediction of seed stress resistance through feature fusion and a fully connected neural network. This process not only improves the accuracy and robustness of the model, but also provides reliable data support for optimizing planting strategies.
[0126] S35 specifically includes:
[0127] S351: Input the comprehensive feature vector into the input layer of the Fully Connected Neural Network (FCNN). Each neuron in the input layer receives a feature value and passes it to the hidden layer.
[0128] S352: In the hidden layer, each neuron performs a weighted summation of the received input features and adds a bias term. The calculation formula is: ,in, For the The weighted sum of the hidden layer neurons, For input features To The weights of the neurons in the hidden layer, For the The bias term of the hidden layer neurons;
[0129] S353: Input the calculated weighted sum into the activation function to obtain an activation output. The activation function is ReLU and the calculation formula is: ,in, For the The activation output of the hidden layer neurons;
[0130] S354: passing the activation output of the hidden layer to the next hidden layer or output layer, repeating the calculation process of S352 and S353, the specific calculation formula is the same;
[0131] S355: In the output layer, each neuron receives the activation output from the last hidden layer and calculates the final output value through weighted summation and activation function. The activation function of the output layer is a linear activation function or a Softmax function, which is used for classification or regression tasks.
[0132] S356: Calculate the error between the predicted value of the output layer and the actual observed data, using the mean square error (MSE) or cross-entropy loss function (Cross-EntropyLoss) for error calculation. The specific formula is: ;in, For the The actual value of the samples, For the The predicted value of samples, is the sample size;
[0133] S357: Adjust the model parameters through the back propagation algorithm to minimize the error. The back propagation algorithm includes calculating the gradient of the loss function for each parameter and updating the parameters according to the gradient descent method. The update formula is: ,in, are model parameters, is the learning rate, is the loss function;
[0134] Step By inputting the comprehensive feature vector into a fully connected neural network and using multi-layer neuron calculation and activation function to analyze the association between seed phenotypic characteristics and environmental data, high-precision prediction of corn seed stress resistance can be achieved, providing a scientific basis for optimizing planting strategies.
[0135] S4 specifically includes:
[0136] S41: Input the comprehensive feature vector extracted in S3 into the Bayesian network to initialize the structure and parameters of the Bayesian network. The Bayesian network consists of nodes and edges. The nodes represent variables, and the edges represent the conditional dependency relationship between variables.
[0137] S42: Use prior knowledge and data-driven methods to build a Bayesian network structure. Prior knowledge can be obtained through expert experience. The data-driven method is a structural learning algorithm, including the K2 algorithm or the greedy search algorithm. The formula of the K2 algorithm is: ,in, represents the score of the network structure, For the nodes, For Node The parent node set of
[0138] S43: According to the constructed Bayesian network structure, the conditional probability distribution of each node in the network is calculated, and the parameters of the conditional probability distribution are estimated by maximum likelihood estimation (MLE). The specific formula is:
[0139] ,in, Indicates simultaneous occurrence and its parent node, Indicates the number of times the parent node appears;
[0140] S44: Inference is performed in the Bayesian network. The conditional probability of each feature in the comprehensive feature vector for the stress resistance of corn seeds is calculated by the Bayesian reasoning method. The reasoning formula is:
[0141] ,in, For under evidence The posterior probability of For The probability of evidence under the condition, for The prior probability of for the probability of evidence;
[0142] S45: Using the identified features, a stress resistance prediction model is established, and these features are used as input variables to predict the stress resistance level of seeds through a Bayesian network; by using a Bayesian network to identify features related to stress resistance from a comprehensive feature vector and establish a stress resistance prediction model, the present invention can achieve efficient and accurate stress resistance prediction, and by combining prior knowledge and data-driven methods to construct a Bayesian network, and using conditional probability distribution for reasoning, it can make full use of the information in the data and improve the prediction ability of the model, which provides a scientific basis for optimizing planting strategies.
[0143] The stress resistance prediction model established in S45 specifically includes:
[0144] S451: Select the identified features as input nodes in the Bayesian network, and specifically represent the feature variables as , where each Represents a feature;
[0145] S452: Set seed stress resistance as the target variable , and use it as the output node in the Bayesian network;
[0146] S453: Calculate the conditional probability distribution between each input feature and the target variable. The calculation formula of the conditional probability distribution is: ,in, For a given feature Conditional target variable The posterior probability of For the target variable Conditional Features The conditional probability of is the prior probability of the target variable, is the joint probability of the features;
[0147] S454: Estimate the conditional probability table (CPT) of each node in the Bayesian network using the maximum likelihood estimation (MLE) method;
[0148] S455: Calculate the target variable under the predetermined input feature combination using the estimated conditional probability table through the Bayesian inference method The predicted value of , the formula of Bayesian inference is:
[0149] ,in, is the target variable under the evidence condition The posterior probability of For the target variable Conditional probability of evidence given, is the prior probability of the target variable, for the probability of evidence;
[0150] S456: The final stress resistance prediction model is expressed as:
[0151] ,in, For a given feature Conditional target variable The posterior probability represents the stress resistance prediction result; the above steps establish a stress resistance prediction model by using Bayesian network and maximum likelihood estimation method, which can effectively calculate the conditional probability between each feature and stress resistance, and accurately predict the stress resistance level of seeds, thereby improving the accuracy of stress resistance prediction.
[0152] S5 specifically includes:
[0153] S51: Collect new corn seed samples and obtain their phenotypic characteristic data through preprocessing and imaging steps, and record corresponding environmental data;
[0154] S52: input the phenotypic characteristic data and environmental data of the newly collected corn seed samples into the final stress resistance prediction model, and the model parameters have been determined through previous data training and verification;
[0155] S53: In the final stress resistance prediction model, conditional probability calculation is performed on the input phenotypic characteristic data and environmental data, and the conditional probability between each characteristic and the stress resistance target variable is calculated using the conditional probability table determined in the model;
[0156] S54: Through Bayesian reasoning, the conditional probabilities of all input features are integrated to generate the stress resistance prediction results of the newly collected samples, and the prediction results are presented in the form of probability distribution of stress resistance levels;
[0157] S55: post-processing the generated stress resistance prediction results, including determining a final stress resistance score or classification label, and making a judgment based on the probability distribution output by the model and a preset threshold or standard;
[0158] S56: The stress resistance prediction results are stored in the database and associated with the identifiers of the corresponding newly collected samples to ensure the traceability and integrity of the data; the above steps use the Bayesian network to predict the stress resistance of newly collected corn seed samples, which can quickly and accurately evaluate the stress resistance level of seeds, thereby providing a scientific basis for seed selection and breeding decisions in actual agricultural production, and significantly improving the optimization effect of planting strategies.
[0159] S6 specifically includes:
[0160] S61: analyzing the prediction results of corn seed stress resistance obtained in S5, determining the stress resistance levels of various subsamples, and classifying the results into high stress resistance, medium stress resistance, and low stress resistance;
[0161] S62: Correlate predicted stress resistance results with environmental data of seed samples to identify seed characteristics that exhibit high stress resistance;
[0162] S63: According to the analysis results of S62, the planting strategy of corn seeds is adjusted, including selecting seed varieties suitable for high stress resistance and adjusting the planting density and planting period. The above steps can effectively improve the adaptability and efficiency of agricultural production by dynamically adjusting the planting strategy of corn seeds by utilizing the prediction results of the Bayesian network, so that planting decisions can be flexibly adjusted according to specific environmental conditions and seed characteristics, thereby improving the overall performance and yield of crops.
[0163] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.
[0164] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.
Claims
1. A method for analyzing and processing corn seed stress resistance data, characterized in that: The following steps are involved: S1: Collect corn seed samples from different regions and use environmental sensors to collect environmental data related to seed growth, including temperature, humidity, light intensity, and soil salt concentration; S2: performing non-destructive phenotypic imaging on the sample, and obtaining the growth status and morphological characteristics of the seeds using high-resolution imaging technology; S3: inputting the phenotypic imaging data and the environmental data into a preset machine learning model, wherein the machine learning model is composed of a convolutional neural network and a recurrent neural network, and is used to analyze the association between the phenotypic characteristics of the seeds and the environmental data; S4: using a Bayesian network, identifying features related to stress resistance from the associated data obtained in S3, and establishing a stress resistance prediction model, wherein S4 specifically includes: S41: Input the comprehensive feature vector extracted in S3 into the Bayesian network to initialize the structure and parameters of the Bayesian network. The Bayesian network consists of nodes and edges. The nodes represent variables, and the edges represent the conditional dependency relationship between variables. S42: Constructing Bayesian network structures using prior knowledge and data-driven methods. The data-driven methods are structure learning algorithms, including K2 algorithms or greedy search algorithms. S43: According to the constructed Bayesian network structure, the conditional probability distribution of each node in the network is calculated, and the parameters of the conditional probability distribution are estimated by maximum likelihood estimation. The specific formula is: ; in, Indicates simultaneous occurrence and its parent node, Indicates the number of times the parent node appears; S44: Inference is performed in the Bayesian network. The conditional probability of each feature in the comprehensive feature vector for the stress resistance of corn seeds is calculated by the Bayesian reasoning method. The reasoning formula is: ,in, For under evidence The posterior probability of For The probability of the evidence conditional on for The prior probability of for the probability of evidence; S45: Using the identified features, establish a stress resistance prediction model; S5: using the stress resistance prediction model established in S4, predicting the stress resistance of the newly collected corn seed samples and generating prediction results; S6: Based on the prediction results of S5, dynamically adjust the planting strategy of corn seeds.
2. The method for analyzing and processing corn seed stress resistance data according to claim 1, characterized in that: The S1 specifically includes: S11: Determine the target collection area and select representative climate zones through geographic information system analysis, including high temperature zone, low temperature zone, high humidity zone, low humidity zone, high light zone and low light zone; S12: Establish multiple sampling points in each target collection area, and collect a predetermined number of corn seed samples at each sampling point by random sampling method; S13: Install a variety of environmental sensors at each sampling point, including temperature sensors, humidity sensors, light intensity sensors, and soil salt concentration sensors, to continuously monitor and record environmental data; S14: The data collected by the environmental sensor is transmitted to the central database in real time through wireless transmission technology; S15: Assign a unique identifier to each sampling point in the central database, associate the identifier with the corresponding corn seed sample and environmental data, and annotate the environmental data of each seed sample through a data matching algorithm.
3. The corn seed stress resistance data analysis and processing method according to claim 1, characterized in that: The S2 specifically includes: S21: Pre-treat the collected corn seed samples, including cleaning and drying, to ensure that there are no impurities on the surface of the samples; S22: photographing the sample at multiple angles using a high-resolution imaging device to obtain image data at different angles. The high-resolution imaging device includes a high-resolution digital camera and a near-infrared imager; S23: pre-processing the image data captured in S22 using an image processing algorithm, including denoising, contrast enhancement, and color correction; S24: Using image segmentation technology, the growth status and morphological characteristics of corn seeds are separated from the background; S25: Quantitatively analyze the segmented images using computer vision algorithms to extract phenotypic characteristics of corn seeds, including seed size, shape, color, surface texture, and internal structure; S26: The extracted phenotypic characteristic data is stored in a central database and associated with the corresponding seed sample unique identifier.
4. The method for analyzing and processing corn seed stress resistance data according to claim 3, characterized in that: The S25 includes: S251: extracting the outline and edge of the seed using an edge detection algorithm to obtain the size characteristics of the seed, wherein the edge detection algorithm adopts Canny edge detection; S252: Extracting the shape features of seeds through morphological operations, including erosion and dilation; S253: using a color analysis algorithm to obtain color features of the seeds, specifically including converting the image into an HSV color space; S254: Calculate the surface texture features of seeds using gray-level co-occurrence matrix; S255: Obtain the internal structural characteristics of seeds through image reconstruction technology.
5. The corn seed stress resistance data analysis and processing method according to claim 1, characterized in that: The S3 specifically includes: S31: Preprocess the acquired phenotypic imaging data and environmental data, including data normalization and missing value filling; S32: Input the preprocessed phenotypic imaging data into a convolutional neural network, and extract high-order phenotypic characteristics of the seeds through multi-layer convolution and pooling operations. The convolutional neural network specifically includes a convolution layer, an activation layer, a pooling layer, and a fully connected layer; S33: inputting the preprocessed environmental data into a recursive neural network, and analyzing the time series characteristics of the environmental data through a cyclic calculation of time steps, wherein the recursive neural network specifically includes an input layer, a hidden layer, and an output layer; S34: Fusing the high-order phenotypic features extracted by the convolutional neural network and the time series features analyzed by the recurrent neural network to form a comprehensive feature vector; S35: The comprehensive feature vector is input into a fully connected neural network, and the association between the phenotypic characteristics of the seeds and the environmental data is analyzed through multi-layer neuron calculations and activation functions. The fully connected neural network specifically includes an input layer, a hidden layer, and an output layer.
6. The method for analyzing and processing corn seed stress resistance data according to claim 5, characterized in that: The S35 specifically includes: S351: Input the comprehensive feature vector into the input layer of the fully connected neural network, each neuron in the input layer receives a feature value and passes it to the hidden layer; S352: In the hidden layer, each neuron performs a weighted summation of the received input features and adds a bias term. The calculation formula is: ,in, For the The weighted sum of the hidden layer neurons, For input features To The weights of the neurons in the hidden layer, For the The bias term of the hidden layer neurons; S353: Input the calculated weighted sum into the activation function to obtain an activation output, where the activation function is ReLU and the calculation formula is: ,in, For the The activation output of the hidden layer neurons; S354: passing the activation output of the hidden layer to the next hidden layer or output layer, and repeating the calculation process of S352 and S353; S355: In the output layer, each neuron receives the activation output from the last hidden layer and calculates the final output value through weighted summation and activation function. The activation function of the output layer is a linear activation function or a Softmax function, which is used for classification or regression tasks. S356: Calculate the error between the predicted value of the output layer and the actual observed data, and use the mean square error or cross entropy loss function to calculate the error; S357: Adjust the model parameters through the back propagation algorithm to minimize the error. The back propagation algorithm includes calculating the gradient of the loss function for each parameter and updating the parameters according to the gradient descent method. The update formula is: ,in, are model parameters, is the learning rate, is the loss function.
7. The method for analyzing and processing corn seed stress resistance data according to claim 1, characterized in that: The S45 specifically includes: S451: Select the identified features as input nodes in the Bayesian network, and specifically represent the feature variables as , where each Represents a feature; S452: Set seed stress resistance as the target variable , and use it as the output node in the Bayesian network; S453: Calculate the conditional probability distribution between each input feature and the target variable. The calculation formula of the conditional probability distribution is: ; in, For a given feature Conditional target variable The posterior probability of For the target variable Conditional Features The conditional probability of is the prior probability of the target variable, is the joint probability of the features; S454: estimating the conditional probability table of each node in the Bayesian network using the maximum likelihood estimation method; S455: Calculate the target variable under the predetermined input feature combination using the estimated conditional probability table through the Bayesian inference method The predicted value of , the formula of Bayesian inference is: ,in, is the target variable under the evidence condition The posterior probability of For the target variable Conditional probability of evidence given, is the prior probability of the target variable, for the probability of evidence; S456: The final stress resistance prediction model is expressed as: ; in, For a given feature Conditional target variable The posterior probability represents the prediction result of stress resistance.
8. The method for analyzing and processing corn seed stress resistance data according to claim 1, characterized in that: The S5 specifically includes: S51: Collect new corn seed samples and obtain their phenotypic characteristic data through preprocessing and imaging steps, and record corresponding environmental data; S52: inputting the phenotypic characteristic data and environmental data of the newly collected corn seed samples into the final stress resistance prediction model; S53: In the final stress resistance prediction model, conditional probability calculation is performed on the input phenotypic characteristic data and environmental data, and the conditional probability between each characteristic and the stress resistance target variable is calculated using the conditional probability table determined in the model; S54: Through Bayesian reasoning, the conditional probabilities of all input features are integrated to generate the stress resistance prediction results of the newly collected samples; S55: post-processing the generated stress resistance prediction results, including determining a final stress resistance score or classification label; S56: The stress resistance prediction result is stored in a database and associated with the identifier of the corresponding newly collected sample.
9. The method for analyzing and processing corn seed stress resistance data according to claim 1, characterized in that: The S6 specifically includes: S61: Analyze the prediction results of corn seed stress resistance obtained in S5, determine the stress resistance levels of various subsamples, and classify the results into high stress resistance, medium stress resistance and low stress resistance; S62: Correlate predicted stress resistance results with environmental data of seed samples to identify seed characteristics that exhibit high stress resistance; S63: Based on the analysis results of S62, adjust the corn seed planting strategy, including selecting seed varieties suitable for high stress resistance and adjusting the planting density and planting period.
Citation Information
Patent Citations
Methylation sequencing data processing method and device
CN112599194A
System for training integrated neural network device to assess predictive uncertainty
CN117501280A
Elymus resource stress resistance evaluation system
CN117910875A
Method for identifying stress-resistant germplasm by using image
CN118172661A
Method for detecting plant diseases and physiological properties influenceable by environmental stress in plant material
EP4047351A1
