An intelligent management system for agricultural product seed cultivation

Abstract

The invention relates to the technical field of agricultural product cultivation, and specifically to an intelligent management system for agricultural product seed cultivation, comprising: a cultivation collection module, which is used for collecting cultivation index data of seed cultivation, and classifying and processing the cultivation index data; a cultivation management module, which is used for detecting the cultivation index data, and setting a cultivation management model corresponding to the cultivation index data, and generating a cultivation plan corresponding to the current cultivation index data according to the cultivation management model; a cultivation analysis module, which is used for analyzing the cultivation plan, and determining the influence index of the current cultivation index data on the cultivation of seeds; an environmental difference module, which is used for obtaining environmental data during seed cultivation, combining the environmental data with the influence index, and determining the differential factor corresponding to the influence index; and a cultivation adjustment module, which is used for cyclically traversing the process of seed cultivation according to the differential factor of the influence index, and adjusting the cultivation plan according to the traversed data, so as to improve the efficiency and quality of seed cultivation.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural product cultivation, and in particular to an intelligent management system for agricultural product seed cultivation. Background Art

[0002] In the traditional agricultural seed breeding process, managers usually rely on manual experience and regular field observations to evaluate seed growth and adjust breeding plans. However, this approach has many shortcomings. For example, manual observations are often affected by subjective factors, resulting in inaccurate evaluation results; the frequency of regular observations is limited, making it difficult to promptly detect and deal with problems that arise during seed growth; at the same time, traditional breeding plans often lack scientific basis and dynamic adjustment mechanisms, making it difficult to adapt to the needs of different growth environments and seed types.

[0003] For example, Chinese patent publication number CN117829414A discloses a method, system and electronic equipment for managing the cultivation of Citrus aurantium seedlings, which relates to the field of Citrus aurantium seedling cultivation, including: obtaining monitoring information of Citrus aurantium seedlings in different regions through a monitoring module of a management system; obtaining seedling phenotypic characteristics, light characteristics, soil pH value, ambient temperature, stratification humidity, fertilization history, management history and storage information; inputting the monitoring information, the seedling phenotypic characteristics, the light characteristics, the soil pH value, the ambient temperature, the stratification humidity, the fertilization history, the management history and the storage information into the management system, so as to regulate the cultivation process of the Citrus aurantium seedlings through the management system.

[0004] For example, Chinese patent publication number CN117933580A discloses a breeding material optimization evaluation method for wheat breeding management system; pre-construct a decision tree model based on wheat breeding index data to obtain different sub-nodes; obtain the same type of change characteristic value and the first classification effect factor based on the distance characteristics between the breeding index data in the scatter plot. Obtain the interval interpolation amount of the breeding index data and the breeding data curve based on the quantitative difference characteristics between the breeding index data in the sub-nodes; obtain the second classification effect factor based on the trend correlation characteristics between the breeding data curve and the overall breeding data curve.

[0005] The prior art describes the processing of breeding data according to the technical means of environment and feature classification respectively, but does not take into account the multiple indicators measured for seeds during seed development during the processing. As a result, it is difficult to adjust the set environment and feature processing methods according to the performance of seeds under different indicators and environments, so as to improve the effect of dynamic adjustment during seed cultivation. Summary of the invention

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is: an intelligent management system for agricultural product seed cultivation, including: a cultivation collection module, which is used to collect cultivation index data of seed cultivation and classify and process the cultivation index data.

[0007] The cultivation management module is used to detect the cultivation indicator data, set the cultivation management model corresponding to the cultivation indicator data, and generate the cultivation plan corresponding to the current cultivation indicator data according to the cultivation management model.

[0008] The cultivation analysis module is used to analyze the cultivation plan and determine the impact of the current cultivation indicator data on the cultivation seeds.

[0009] The environmental differential module is used to obtain environmental data during seed cultivation, combine the environmental data with the influencing indicators, and determine the differential factors corresponding to the influencing indicators.

[0010] The cultivation adjustment module is used to loop through the seed cultivation process according to the differential factors that affect the indicators, and adjust the cultivation plan according to the traversed data.

[0011] The beneficial effects of the present invention are as follows: 1. Through the cultivation collection module, the system can collect various index data in the seed cultivation process in real time, and classify and process the data, which improves the efficiency and accuracy of data collection while considering the periodicity and mutation of the seed corresponding data. At the same time, after clustering the data, the clustered data is superimposed according to the second classification factor related to periodicity and the first classification factor related to mutation, which can improve the efficiency and classification accuracy of the data during processing, and can further explore the change trends and characteristics in the data, providing a basis for subsequent management and decision-making.

[0012] Second, the present invention sets a cultivation management module based on the collected data. The cultivation management model adopts a decision tree method, and adjusts the decision tree according to the container seedling preservation rate and information gain value, so that the set model can comprehensively consider the growth environment of the seeds, matrix components and other factors to generate the optimal cultivation plan. At the same time, the system can also dynamically adjust and optimize the model according to the update frequency of the data to ensure the timeliness and accuracy of the cultivation plan.

[0013] 3. Through the cultivation analysis module and the environmental difference module, the system of the present invention can accurately evaluate the cultivation plan and determine the key factors affecting seed cultivation. On this basis, the cultivation adjustment module can traverse the seed cultivation process in a loop and adjust and optimize the cultivation plan according to the traversal results. This dynamic adjustment and optimization mechanism can ensure that the seeds grow in the best growth environment and improve the survival rate and quality of the seeds. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0015] Figure 1 It is a system framework diagram of an intelligent management system for agricultural product seed cultivation.

[0016] Figure 2 The invention is a flow chart of the cultivation and collection module of the intelligent management system for agricultural product seed cultivation.

[0017] Figure 3 The invention is a flow chart of a cultivation management model of an intelligent management system for agricultural product seed cultivation. DETAILED DESCRIPTION

[0018] The embodiments of the present invention are described in detail below. The embodiments described below are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. If no specific techniques or conditions are specified in the embodiments, the techniques or conditions described in the literature in the art or the product specifications are used.

[0019] See also Figure 1 , an intelligent management system for agricultural product seed cultivation, comprising: a cultivation collection module, a cultivation management module, a cultivation analysis module, an environmental difference module and a cultivation adjustment module; the output end of the cultivation collection module is connected to the cultivation management module, the output end of the cultivation management module is connected to the output end of the cultivation analysis module, the output end of the cultivation analysis module is connected to the output end of the environmental difference module, and the output end of the environmental difference module is connected to the cultivation adjustment module.

[0020] The cultivation collection module is used to collect the cultivation index data of seed cultivation and classify and process the cultivation index data.

[0021] The cultivation management module is used to detect the cultivation indicator data, set the cultivation management model corresponding to the cultivation indicator data, and generate the cultivation plan corresponding to the current cultivation indicator data according to the cultivation management model.

[0022] The cultivation analysis module is used to analyze the cultivation plan and determine the impact of the current cultivation indicator data on the cultivation seeds.

[0023] The environmental differential module is used to obtain environmental data during seed cultivation, combine the environmental data with the influencing indicators, and determine the differential factors corresponding to the influencing indicators.

[0024] The cultivation adjustment module is used to loop through the seed cultivation process according to the differential factors that affect the indicators, and adjust the cultivation plan according to the traversed data.

[0025] In one embodiment of the present invention, the cultivation index data will represent the data measured under normal seed germination, such as germination rate, germination potential, seedling growth height, biomass, root index, water content, chlorophyll content, germination daily growth, and other measurable indicators, such as protein content, antioxidant level, mineral element content (such as nitrogen, phosphorus, potassium, etc.), enzyme activity (such as peroxidase, superoxide dismutase, etc.) and other parameters. These parameters will indicate the effect of seed cultivation under a specific environment, as well as the success rate of seed cultivation, to express whether the current seed cultivation is effective.

[0026] Germination rate: refers to the ratio of the number of seeds that successfully germinate after a certain period of time under specific conditions to the total number of tested seeds. It is a basic indicator for measuring seed vitality.

[0027] Germination vigor: reflects the ability of seeds to germinate quickly and uniformly under suitable conditions, usually assessed by counting the number of seeds that germinate within the first few days.

[0028] These two data can be measured using culture dishes, germination paper or sand, incubators, etc. to record seed-related conditions.

[0029] Seedling growth height: Measuring the height of seedlings from germination to a certain stage (such as two weeks or one month) is used to evaluate the growth rate and health of seeds after germination. The height can be measured using a ruler or an electronic height meter.

[0030] Biomass: includes the dry weight or fresh weight of the above-ground and underground parts, which is used to measure the growth and health of the plant. At this time, the above-ground and underground parts (roots) are collected separately, and after cleaning, they are weighed in two ways: fresh weight and dry weight. Before measuring the dry weight, the sample needs to be placed in an oven to dry to a constant weight.

[0031] Root system indicators include but are not limited to the following: Root length: refers to the total length of the entire root system from top to base.

[0032] Root volume: The amount of space occupied by the roots.

[0033] Root surface area: The total surface area of ​​all roots, which affects the efficiency of water and nutrient absorption.

[0034] Number of root tips: Root tips are the most active parts in absorbing water and nutrients, and their number is directly related to the plant's absorption capacity.

[0035] The root indicators can be calculated by using a camera to obtain a picture of the root system, and then using image analysis to calculate its volume, area and other data. At this time, transparent culture dishes will be used to cultivate seeds. The culture dishes can directly observe the root system, germination and other conditions of the seeds, so as to timely evaluate the corresponding conditions of the seed cultivation.

[0036] Moisture content: Determining the moisture content of seeds or seedlings is important for evaluating seed vitality, storage stability and stress resistance. The moisture content is measured by weighing a certain amount of fresh samples and placing them in an oven at 105°C for 30 minutes, then drying them at 80°C to constant weight, and calculating the percentage of water loss.

[0037] Chlorophyll content: The concentration of photosynthetic pigments indirectly reflects the photosynthetic efficiency of plants, which can be measured using specialized instruments. Chlorophyll content can be calculated by using solvents to extract chlorophyll from leaves and then measuring the absorbance with a spectrophotometer; portable chlorophyll meters directly read values ​​based on optical principles.

[0038] Germination day growth: record the growth changes every day or every few days to observe and analyze the speed of seed germination and seedling growth. Germination day growth records the growth of seed germination.

[0039] Protein is the main bearer of life activities and is involved in almost all physiological processes in plants, including enzyme catalysis, structural support, transport of substances, signal transduction, etc. Determining the protein content in seeds or seedlings can help evaluate their nutritional status and metabolic activity, which is important for understanding seed vitality and seedling health. Protein content can be measured using a spectrophotometer to identify the proteins present in seeds.

[0040] Antioxidant levels are expressed as antioxidants (such as vitamin C, vitamin E, carotenoids, polyphenols, etc.) that neutralize reactive oxygen species (ROS) produced in plants and prevent cell damage. Under adverse conditions (such as drought, salinity, low temperature, etc.), the generation of ROS in plants increases, and the role of antioxidants is particularly important at this time. Monitoring antioxidant levels helps to understand the adaptability and resistance of plants to environmental stress. Antioxidant levels are measured using high performance liquid chromatography (HPLC).

[0041] Mineral element content (such as nitrogen, phosphorus, potassium, etc.): Nitrogen (N), phosphorus (P) and potassium (K) are the three main nutrients necessary for plant growth and development.

[0042] Nitrogen: Promotes green leaves and lush branches and foliage, and is an important component of amino acids, proteins, and nucleic acids.

[0043] Phosphorus: Participates in key biochemical processes such as energy conversion (ATP formation), photosynthesis, and respiration, and is especially important for root development.

[0044] Potassium: Enhances plant disease resistance and lodging resistance, regulates stomatal opening and closing to control water evaporation, and affects sugar transport.

[0045] Monitoring the levels of these mineral elements can help optimize fertilization strategies and ensure that plants receive an adequate supply of nutrients. Mineral element levels can be measured using atomic absorption spectrometry or inductively coupled plasma mass spectrometry.

[0046] Enzyme activity (e.g. peroxidase, superoxide dismutase, etc.): Enzymes are catalysts for various chemical reactions in organisms, and different enzymes are responsible for specific types of reactions. For example, peroxidase (POD): participates in the removal of hydrogen peroxide and other harmful substances in plants, protecting cells from damage. Superoxide dismutase (SOD): can convert superoxide free radicals into hydrogen peroxide and oxygen, reducing ROS damage to cells.

[0047] Measuring the activity of specific enzymes can reflect the physiological state of plants and their response mechanisms to environmental changes, helping researchers understand how plants cope with biotic and abiotic stresses. Enzyme activity is measured using a spectrophotometer.

[0048] By measuring these cultivation indicator data, we can know the specific changes of seeds during the seedling process. Based on these data, we can comprehensively evaluate whether the seeds germinate normally and whether these measured data are the same as expected.

[0049] When classifying seeds, they will be classified according to the acquisition time and corresponding values ​​of the cultivation indicator data. For example, time series analysis is used to complete the initial classification of seeds. After that, the cultivation indicator data that have completed the initial classification will be tested to complete the model of the cultivation indicator data. After that, the data in the model can be processed in various ways, and the corresponding plan of the cultivation indicator data when it was collected can be checked to complete the cultivation management of the seeds.

[0050] Initial classification of seed usage time series can distinguish relatively obvious indicator values ​​of seeds, and according to these indicator values, the efficiency and survival rate of seeds in the germination process can be analyzed. The trend of seed growth and germination, as well as the periodic characteristics of seed existence can be identified, which is crucial for understanding the optimal planting time and environmental conditions of seeds.

[0051] like Figure 2 As shown, the processing method of the cultivation acquisition module also includes: arranging the cultivation indicator data in the form of a time series to obtain a cultivation time series, processing each indicator existing in the cultivation time series, and obtaining a cultivation change trend corresponding to the cultivation time series. For each indicator in the cultivation time series, each indicator set in the cultivation indicator data is represented, and then these indicators are converted into corresponding cultivation change trends in the form of a time series to identify the trend of the corresponding indicator.

[0052] Based on the acquired cultivation change trends, the turning point characteristics and periodic characteristics in the cultivation change trends are identified. The turning point characteristics in the cultivation change trends indicate the points where the cultivation change trends are reversed and accelerated. These points will reflect the time periods of large-scale germination and rapid growth of seeds, or the points where seeds mutate during normal cultivation. These points will have a significant impact on the normal germination and growth of seeds and need to be identified.

[0053] Periodic characteristics indicate the existence of stable characteristics in the cultivation change trend, that is, the positions where there are stable peaks in the cultivation change trend and the data at the same slope in the cultivation change trend are regarded as periodic characteristics; these periodic characteristics can show the stability of the current seeds during cultivation, and these conditions can reflect whether the seeds grow as expected.

[0054] Calculate the average value of the data corresponding to the inflection point feature, and set the ratio of the average values ​​between adjacent inflection point features as the first classification factor. This first classification factor can represent the relative situation between the data corresponding to these inflection point features when there are multiple inflection point features in the cultivation change trend. At the same time, the average value of the inflection point feature is calculated according to the points adjacent to the corresponding points of the inflection point feature. Then, the ratio of the average values ​​of the multiple inflection point features is calculated to obtain whether the seeds show rapid growth or decline and stable growth after the trend changes.

[0055] The average value of the cosine similarities between all periodic features is calculated as the second classification factor. At this time, when the data represented by the periodic features is in a stable state, the calculated cosine similarities will present positively correlated values. Then, the average value of the positive correlation is taken to obtain the positive correlation between the periodic features. Finally, according to this value, the subsequent clustering is adjusted to obtain the seeds after the classification is completed.

[0056] According to the first classification factor and the second classification factor, the data in the cultivation time series are clustered and analyzed to complete and classify the cultivation indicator data.

[0057] At this time, according to the first classification factor and the second classification factor, the implementation method of clustering the data in the cultivation time series is: regard the data in the cultivation time series as data points, calculate the Euclidean distance between the data points, generate multiple cluster centers according to the Euclidean distance between the data points, and use the first classification factor and the second classification factor to weight each cluster center in turn, superimpose the weighted cluster centers, and output the largest cluster center after superposition to complete the classification of the cultivation index data. At this time, the first classification factor and the second classification factor are selected for weighting in order to enhance the periodic characteristics and the existing mutation conditions. After that, the superimposed cluster centers will comprehensively consider these two conditions to complete the clustering. According to the final output cluster center, the main situation of the current cultivation index data can be known, and according to this situation, a relatively comprehensive classification can be obtained, which is convenient for the subsequent setting of the model for the cultivation index data and improves the processing efficiency of the model.

[0058] In one embodiment of the present invention, the cultivation management module will further process the cultivation index data according to the classified parts and set the corresponding cultivation management model. The cultivation management module set at this time will be set in the form of a decision tree to obtain the decision of the current cultivation index data in the output part of the decision tree, and obtain the corresponding cultivation plan. This cultivation plan will serve as a further reference form of the current cultivation index data. This plan will include multiple parameters set for seed cultivation, and use corresponding indicators to evaluate these multiple parameters to determine subsequent influencing indicators.

[0059] Therefore, the implementation method of the cultivation management model includes: setting a decision tree for cultivation index data, extracting the container seedling preservation rate corresponding to the seeds in the cultivation index data, and the container seedling preservation rate indicates how many seeds have completed normal germination within the normal germination time of the seeds in the current cultivation index data. The ratio of germinated seeds to ungerminated seeds is the container seedling preservation rate.

[0060] The data when the container seedling preservation rate takes the maximum value is set as the root node of the decision tree. This root node will represent the distribution of relevant content in the corresponding cultivation indicator data when all the seeds germinate as much as possible. According to the distribution of these data, the main content when the seeds germinate can be identified, and the cultivation plan can be selected based on this part of the content.

[0061] Calculate the information gain of each intermediate node and leaf node under the root node, and divide the intermediate nodes according to the value of information gain to ensure that the information gain value of each intermediate node is the maximum value. In order to obtain the best split point, the information gain of the intermediate node needs to reach the maximum value to reduce the fitting between the data.

[0062] According to the information gain of the leaf nodes, the leaf nodes with the maximum information gain rate under each intermediate node are deduced, and the decision tree is adjusted according to the leaf nodes with the maximum information gain rate under each intermediate node, and the adjusted decision tree is set as the cultivation management model. The maximum information gain rate is the ratio of the information gain calculated on the leaf node to all information gains. At this time, the leaf node with the maximum information gain rate is selected to find relatively important data in the nodes divided by the decision tree. These data will show high correlation and similar values. At the same time, adjusting the decision tree based on these data will improve the splitting criteria of the decision tree, thereby optimizing the performance of the model; focusing on the ratio of the size of each subset after the split, to punish those splits that lead to too many child nodes. This can reduce the bias towards these features and make the model pay more attention to those features that really help improve purity.

[0063] The method for adjusting the decision tree is as follows: calculate the distance between the leaf nodes with the maximum information gain rate under each intermediate node as the first adjustment factor; calculate the standard deviation of the distance between the leaf node with the maximum information gain rate under each intermediate node and the leaf nodes with the maximum information gain rate under all intermediate nodes as the second adjustment factor, and iterate the first adjustment factor and the second adjustment factor in a loop. When the first adjustment factor and the second adjustment factor take the minimum value, output the decision tree to obtain a cultivation management model.

[0064] The first adjustment factor attempts to quantify the difference between leaf nodes produced by the best split at each intermediate node. The distance here can be understood as a measure of the difference between the data sets represented by different leaf nodes, such as Euclidean distance or other appropriate metrics. If the distance between leaf nodes under some intermediate nodes is too large, it means that these nodes may be too specialized to a specific pattern in the training data, increasing the risk of overfitting. By minimizing this distance, the decision tree can be made smoother and less sensitive to noise in the training data.

[0065] The second adjustment factor focuses on the distribution of leaf nodes from a global perspective. It measures the average difference between the best leaf node under each intermediate node and all other best leaf nodes, and evaluates the consistency or dispersion of this difference through the standard deviation. It can help identify nodes that may lead to unbalanced tree structures, that is, the leaf nodes under some nodes are particularly different from other nodes. This helps guide the algorithm to make more balanced choices and avoid extremely uneven subtree structures. By reducing the standard deviation, the entire decision tree structure can be made more uniform, making the model more adaptable and robust to changes in input data.

[0066] The calculation method of the information gain described above has been described in the prior art and will not be explained here.

[0067] After obtaining the cultivation management model, the cultivation plan will select the cultivation plan to be used according to the output of the decision tree corresponding to the cultivation management model; the cultivation plan at this time is set relatively according to the output of the decision tree, and is the initial plan. After that, it is necessary to conduct further analysis according to the possible influencing indicators in the cultivation plan to find the parts of the cultivation plan that are easy to have an impact, and whether these indicators will be affected by the environment, and finally obtain a cultivation plan with small environmental impact and high germination rate.

[0068] Generating a cultivation plan corresponding to the current cultivation index data according to the cultivation management model is essentially to find the cultivation plan corresponding to the data in the current cultivation index from the preset plans after taking the average value according to the data output by the cultivation management model.

[0069] The cultivation plan will include some cultivation index data output by the cultivation management model and the plan set up for the cultivation index data, such as the matrix for cultivating seeds, environmental data, and data on osmotic regulators used.

[0070] The matrix for seed cultivation refers to the physical medium used to support seed germination and initial growth of seedlings. An ideal matrix should have good air permeability, water retention capacity and appropriate nutrient supply. Commonly used matrix components include the following.

[0071] Sphagnum moss: has excellent moisture retention and breathability, but may lack adequate nutrition.

[0072] Perlite: Increases the air permeability and drainage of the substrate and contains no nutrients itself.

[0073] Vermiculite: Improves water retention and cation exchange capacity of the substrate, aiding in nutrient retention.

[0074] Coconut coir: As a substitute for peat moss, it has better air permeability and water retention.

[0075] Leaf mold: provides rich organic matter and trace nutrients, and improves soil structure.

[0076] Sand: Used to improve substrate texture and enhance drainage.

[0077] Fertilizer: Add slow-release fertilizer or liquid fertilizer as needed to meet the nutritional needs of plant growth.

[0078] Choosing the right substrate formula requires consideration of the specific crop type, growth stage, and local environmental conditions. By adjusting the ratio of various ingredients experimentally, the substrate performance can be optimized to promote seed germination and healthy growth of seedlings.

[0079] The osmotic regulators used can be natural organic substances, such as betaine and proline; they can also be synthetic chemicals, such as polyethylene glycol, sorbitol and mannitol. These osmotic regulators will be recorded in the preset plan and selected based on the cultivation index data output by the cultivation management model. At this time, the basis for selecting the plan is to conduct multiple tests and adjust the current content according to the data after the test to achieve the effect of improving the seed germination rate.

[0080] Betaine: such as glycine betaine, is a quaternary ammonium compound widely found in plants, which helps improve plant tolerance to drought and salt stress.

[0081] Proline: An amino acid whose content increases significantly when plants are subjected to stress such as drought, salinity or low temperature, protecting enzyme activity and stabilizing protein structure.

[0082] Polyethylene glycol: Although not strictly an osmotic regulator, it is often used in experiments to simulate drought conditions because PEG can affect the ability of plant roots to absorb water by reducing water potential.

[0083] Sorbitol and mannitol: These sugar alcohols also act as osmotic regulators, helping plants adapt to high salinity or other stressful environments that cause water loss.

[0084] The use of osmotic regulators during seed production can help improve seed germination and seedling survival, especially in less-than-ideal environments. The choice of osmotic regulator depends on the specific crop, expected environmental challenges, and the purpose of the application. For example, when growing crops in arid regions, the use of natural organic substances such as betaine or proline as osmotic regulators may be preferred.

[0085] At this time, when generating a cultivation plan corresponding to the current cultivation indicator data, it is also necessary to consider the update frequency of the output data of the cultivation management model, and identify the type of each leaf node in the output data according to the update frequency of the output data. This is mainly based on the specific attributes existing on the leaf nodes to identify the update frequency of the leaf nodes on the cultivation management model, so as to know how many updates the cultivation management model has completed to obtain the current output data.

[0086] Therefore, the cultivation management module also includes: for the data on the leaf nodes, the variation of the leaf nodes at continuous time points is calculated. This variation represents the absolute difference of the data of the corresponding leaf nodes at continuous time points when the decision tree is adjusted; the probability of occurrence of the variation of the leaf nodes at continuous time points is set as the update probability of the cultivation management model, and the decision tree is adjusted based on the update probability of the cultivation management model. At this time, it is mainly checked whether the currently set decision tree can complete the processing of the current cultivation index data. When the update frequency is greater than the first preset threshold, the current decision tree is adjusted, and the node with the largest value of the current decision tree is set as the root node of the adjusted decision tree, and the decision tree is reset; when the update frequency is less than the first preset threshold, the current data is recorded and the decision tree is not adjusted. The first preset threshold represents the frequency of data updates of the decision tree leaf nodes. When the data update frequency is high, the traditional static decision tree model may no longer be applicable. At this time, you can consider using an incremental learning algorithm or regularly retraining the model to maintain its accuracy; in this way, the decision tree is adjusted to achieve overall processing; at the same time, the node with the largest value in the decision tree represents the node with the largest calculated information gain value and the largest data mean in the decision tree. This node will represent the point with the best initial splitting effect, so that the re-generated decision tree can quickly complete the classification and improve the efficiency of decision tree processing.

[0087] In one embodiment of the present invention, the cultivation analysis module is mainly used to extract corresponding data in the cultivation plan and determine the degree of influence of the data on the cultivation indicator data to obtain the influence indicator.

[0088] In the cultivation plan, the matrix components and environment used in seed cultivation will be extracted, and the most easily affected indicators will be found from these two parts. At this time, the corresponding data in the cultivation plan will be compared with the historical data, such as obtaining the environment corresponding to the cultivation indicator data, and then determining the matrix components corresponding to the cultivation indicator data, and calculating the correlation coefficient between these corresponding data and the historical data, and taking the largest indicator in the correlation coefficient calculated for the environment and the matrix components as the influencing indicator at this time; in this way, the indicators that are easily affected by the environment can be found, and after obtaining these indicators, it is determined whether these affected indicators will affect the germination and growth of seeds. The matrix components here are the above-mentioned matrix for cultivating seeds.

[0089] like Figure 3 As shown, the implementation method of the cultivation analysis module includes: extracting the environmental data and matrix components corresponding to the cultivation plan, calculating the cultivation index data corresponding to the environmental data and matrix components with the historical data, obtaining the Pearson correlation coefficient, and considering the data with the largest Pearson correlation coefficient value among the cultivation index data corresponding to the environmental data and matrix components as the influence indicator of the current cultivation index data on the cultivated seeds.

[0090] Based on this influencing indicator, it is convenient to make subsequent adjustments to the cultivation plan, such as adjusting the proportion of matrix components, improving environmental control and other measures. This method can help identify the most critical influencing factors under the corresponding environment and matrix components, and provide specific optimization directions to improve the success rate and efficiency of seed cultivation.

[0091] In one embodiment of the present invention, the environmental difference module further processes the environmental data. At this time, the difference factor emphasizes the difference between adjacent time points, highlighting the short-term change pattern in the data rather than the long-term trend; for monitoring the rapidly changing environmental conditions (such as moisture content and soil temperature) in the early stage of seed germination, the differenced data can better reflect these subtle and important changes.

[0092] At this time, the way to combine environmental data and impact indicators is to perform differential calculations on the impact indicators according to the environmental data, that is, to take the impact indicators as the dependent variables and the environmental data as the independent variables, and use the autoregressive moving average model for calculation, and use the autoregressive coefficient and moving average coefficient of the autoregressive moving average model after calculation as the differential factors corresponding to the impact indicators.

[0093] The autoregressive moving average model is a statistical model used to analyze and predict stationary time series. This model can capture the dynamic patterns of environmental data and influencing indicators when using time series analysis, and can further emphasize the parts with significant influence between environmental data and influencing indicators, as well as how the parts with significant influence change over time and affect seed germination. At this time, the differential factor will represent these significant situations to enhance the staff's judgment on seed germination.

[0094] In one embodiment of the present invention, the training adjustment module performs set constraints on the data of each loop traversal based on the obtained differential factors, and merges or unmerges the data according to the different data situations in each loop, so as to find the merge points and unmerge points in the loop processing process. These points will represent the modifiable and unmodifiable parts of the training plan, and then the modifiable parts are adjusted to finally complete the adjustment of the training plan.

[0095] When the seed cultivation process is looped, the main traversed data will include the originally set cultivation plan, cultivation index data and environmental data. When these data are looped, they will produce a variety of different results due to different set cultivation plans and different environmental data, resulting in different values ​​of the differential factors at this time. Afterwards, there will be different places to judge whether there is a relationship and intersection relationship, to merge or not merge the corresponding data. The final merged plan can represent a more comprehensive cultivation setting, thereby improving the germination rate of seeds.

[0096] At this time, the corresponding data will be found according to the differential factors, and the data that needs to be processed in a loop will be determined. That is, the implementation method of looping through the seed cultivation process for the differential factors that affect the indicators is as follows: based on the differential factors that affect the indicators, the cultivation management model and cultivation plan corresponding to the differential factors are obtained, the corresponding nodes and cultivation plan data in the cultivation management model are extracted, and the seed cultivation process is looped through. Afterwards, the data that has been traversed and completed will be further set up and processed in the form of a data set.

[0097] Therefore, the implementation method of the cultivation adjustment module also includes: identifying the cultivation plan and cultivation indicator data contained in the traversed data as the traversal data set, comparing the traversal data sets at each traversal, and merging the corresponding traversal data sets when there is an inclusion relationship and a node intersection relationship among the traversal data sets; if there is no inclusion relationship and a node intersection relationship among the traversal data sets, the corresponding traversal data sets are not merged; if there is any one of the inclusion relationship and the node intersection relationship in the traversal data sets, the intersection of the corresponding traversal data sets is merged, and the merged intersection is merged with the original traversal data set as the merged traversal data set; calculating the Pearson correlation coefficient of the merged traversal data set, and taking the cultivation plan in the traversal data set corresponding to the maximum value of the Pearson correlation coefficient as the adjusted cultivation plan.

[0098] At this time, the inclusion relationship mainly refers to whether there are contents that can be included in the setting plan at this time. The node intersection relationship means that when the cultivation index data is processed and the cultivation management model is set, whether there are intersecting nodes in the cultivation management model of these traversed data. These nodes represent the same situation when these cultivation index data are processed. After that, these data are merged to obtain a corresponding intersection, which can better highlight the comprehensiveness of the seed cultivation plan. If both the inclusion relationship and the node intersection relationship are not satisfied, it means that the data during the two traversals do not have common points and cannot be used as an extension and supplement to the same plan. If one of these two relationships is satisfied, the corresponding intersection is obtained to find the existing common points, and the data represented by these intersections are used as supplements to improve the stability of the current data, and finally a more stable cultivation plan is obtained.

[0099] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations on the present invention. A person skilled in the art may make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention and they are still covered by the protection scope of the present invention.

Claims

1. An intelligent management system for agricultural product seed cultivation, characterized in that: include: The cultivation collection module is used to collect the cultivation index data of seed cultivation and classify and process the cultivation index data; The cultivation management module is used to detect the cultivation indicator data, set the cultivation management model corresponding to the cultivation indicator data, and generate the cultivation plan corresponding to the current cultivation indicator data according to the cultivation management model; The cultivation analysis module is used to analyze the cultivation plan and determine the impact of the current cultivation index data on the cultivation seeds; The environmental difference module is used to obtain environmental data during seed cultivation, combine the environmental data with the influencing indicators, and determine the differential factors corresponding to the influencing indicators; The cultivation adjustment module is used to loop through the seed cultivation process according to the differential factors affecting the indicators, and adjust the cultivation plan according to the traversed data; The processing methods of the cultivation and collection module also include: Arrange the cultivation index data in the form of time series to obtain a cultivation time series, process each indicator in the cultivation time series, and obtain the cultivation change trend corresponding to the cultivation time series; Based on the acquired cultivation change trend, identify the turning point characteristics and periodic characteristics in the cultivation change trend; Calculate the average value of the data corresponding to the inflection point features, and set the ratio of the average values ​​between adjacent inflection point features as the first classification factor; Calculate the average of the cosine similarities between all periodic features as the second classification factor; According to the first classification factor and the second classification factor, cluster analysis is performed on the data in the cultivation time series to complete the classification processing of the cultivation indicator data; The implementation of the cultivation management model includes: Setting a decision tree for cultivation index data, extracting the container seedling preservation rate corresponding to the seeds in the cultivation index data; The data when the container seedling preservation rate takes the maximum value is set as the root node of the decision tree; Calculate the information gain of each intermediate node and leaf node under the root node, and divide the intermediate nodes according to the value of information gain to ensure that the information gain value of each intermediate node is the maximum value; The implementation methods of the training adjustment module also include: Identify the cultivation plan and cultivation indicator data contained in the traversed data as the traversal data set, compare the traversal data sets during each traversal, and when the traversal data sets have an inclusion relationship and a node intersection relationship, merge the corresponding traversal data sets; if the traversal data sets do not have an inclusion relationship and a node intersection relationship, the corresponding traversal data sets will not be merged; if the traversal data sets have any one of an inclusion relationship and a node intersection relationship, merge the intersection of the corresponding traversal data sets, and merge the merged intersection with the original traversal data set as the merged traversal data set.

2. The intelligent management system for seed cultivation of agricultural products according to claim 1, characterized in that: According to the first classification factor and the second classification factor, the implementation method of clustering the data in the cultivation time series is: The data existing in the cultivation time series are regarded as data points, the Euclidean distance between the data points is calculated, and multiple cluster centers are generated according to the Euclidean distance between the data points.

3. The intelligent management system for seed cultivation of agricultural products according to claim 2, characterized in that: Each cluster center is weighted using the first classification factor and the second classification factor in turn, the weighted cluster centers are superimposed, and the largest cluster center after superposition is output to complete the classification of the cultivation indicator data.

4. The intelligent management system for seed cultivation of agricultural products according to claim 1, characterized in that: According to the information gain of the leaf nodes, the leaf node with the maximum information gain rate under each intermediate node is calculated, and the decision tree is adjusted according to the leaf node with the maximum information gain rate under each intermediate node, and the adjusted decision tree is set as the cultivation management model.

5. The intelligent management system for seed cultivation of agricultural products according to claim 4, characterized in that: The decision tree is adjusted as follows: Calculate the distance between the leaf nodes with the maximum information gain rate under each intermediate node as the first adjustment factor; calculate the standard deviation of the distance between the leaf node with the maximum information gain rate under each intermediate node and the leaf nodes with the maximum information gain rate under all intermediate nodes as the second adjustment factor, iterate the first adjustment factor and the second adjustment factor in a loop, and when the first adjustment factor and the second adjustment factor take the minimum value, output the decision tree to obtain the cultivation management model.

6. The intelligent management system for seed cultivation of agricultural products according to claim 4, characterized in that: The cultivation management module also includes: For the data on the leaf nodes, the variation of the leaf nodes at consecutive time points is calculated, the occurrence probability of the variation of the leaf nodes at consecutive time points is set as the update probability of the cultivation management model, and the decision tree is adjusted based on the update probability of the cultivation management model; When the update frequency is greater than the first preset threshold, the current decision tree is adjusted, the node with the largest value in the current decision tree is set as the root node of the adjusted decision tree, and the decision tree is reset; when the update frequency is less than the first preset threshold, the current data is recorded and the decision tree is not adjusted.

7. The intelligent management system for seed cultivation of agricultural products according to claim 1, characterized in that: The implementation methods of the cultivation analysis module include: The environmental data and matrix components corresponding to the cultivation plan are extracted, the cultivation index data corresponding to the environmental data and matrix components are calculated with the historical data, the Pearson correlation coefficient is obtained, and the data with the largest Pearson correlation coefficient among the cultivation index data corresponding to the environmental data and matrix components is regarded as the influence indicator of the current cultivation index data on the cultivation of seeds.

8. The intelligent management system for seed cultivation of agricultural products according to claim 1, characterized in that: Environmental data and impact indicators are combined in the following ways: The impact indicators are calculated differentially according to the environmental data, that is, the impact indicators are taken as dependent variables and the environmental data as independent variables, and the autoregressive moving average model is used for calculation. The autoregressive coefficient and moving average coefficient of the autoregressive moving average model after calculation are used as the differential factors corresponding to the impact indicators.

9. The intelligent management system for seed cultivation of agricultural products according to claim 1, characterized in that: The Pearson correlation coefficient of the merged traversal data set is calculated, and the cultivation scheme in the traversal data set corresponding to the maximum value of the Pearson correlation coefficient is used as the adjusted cultivation scheme.

10. The intelligent management system for seed cultivation of agricultural products according to claim 1, characterized in that: According to the differential factors affecting the index, the implementation method of cyclic traversal of the seed cultivation process is as follows: Based on the differential factors affecting the indicators, the cultivation management model and cultivation plan corresponding to the differential factors are obtained, the corresponding nodes and cultivation plan data in the cultivation management model are extracted, and the seed cultivation process is traversed in a loop.

Images