Water and fertilizer management method and system for the whole life cycle of blueberry planting

Abstract

The application provides a water and fertilizer management method and system for the whole life cycle of blueberry planting, relates to the field of crop fertilization, and comprises the following steps: acquiring water and fertilizer management experimental data of multiple blueberry growth stages; based on the water and fertilizer management experimental data of the multiple blueberry growth stages, a growth-related weighted undirected graph of the multiple blueberry growth stages is established; based on the water and fertilizer management experimental data of the multiple blueberry growth stages and the growth-related weighted undirected graph of the multiple blueberry growth stages, key growth factors and key water and fertilizer management factors of each blueberry growth stage are determined; based on the water and fertilizer management experimental data of the blueberry growth stages, the key growth factors and the key water and fertilizer management factors, candidate water and fertilizer management schemes of the blueberry growth stages are determined; and based on the candidate water and fertilizer management schemes of each blueberry growth stage, a water and fertilizer management scheme for the whole life cycle of blueberry planting is generated, which has the advantages of realizing precise supply and intelligent regulation and control of water and fertilizer and improving water and fertilizer utilization rate.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of crop fertilization, in particular to a water and fertilizer management method and system for the whole life cycle of blueberry planting. BACKGROUND

[0002] Blueberry, whose scientific name is Vaccinium, belongs to the Vaccinium genus of the Ericaceae family, and is a perennial deciduous or evergreen shrub. It is rich in anthocyanins and other antioxidants and has been widely used in health food, medicine, cosmetics and other fields, with a growing market demand. As consumers pay more attention to healthy food, the market demand for blueberries is showing a rapid growth trend, which is prompting the blueberry planting industry to develop towards scale, intensification and high efficiency.

[0003] Blueberry is an oligotrophic plant, and good fertilization can not only improve yield and promote plant growth, but also improve blueberry quality.

[0004] In the prior art, there is a lack of scientific irrigation time planning, and irrigation is often based on the experience of the grower, and the irrigation time and irrigation amount cannot be adjusted in time according to the water requirement law of different growth stages of blueberry and the soil moisture condition. For example, during the flowering and fruit enlargement periods of blueberry, the water requirement is relatively sensitive, and if irrigation is not timely or the irrigation amount is insufficient, it will affect flower bud differentiation and fruit development, resulting in a decrease in yield; and during the dormancy period of blueberry, if irrigation is too much, it may cause diseases, affect plant wintering, and there is a lack of scientific fertilization amount calculation method, and the fertilization amount is often determined based on experience, which has the problems of over-fertilization or under-fertilization. Over-fertilization not only causes waste of fertilizer and increases production cost, but also may cause soil salinization and harm to the root system of blueberry; under-fertilization cannot meet the nutrient requirements of blueberry growth and development, affecting the yield and quality of blueberry. At the same time, the traditional water and fertilizer management method often separates irrigation and fertilization, and water and fertilizer cannot be supplied simultaneously, resulting in low fertilizer utilization rate. During irrigation, the fertilizer cannot be dissolved in time and transported to the surrounding of the root system of blueberry with water, and part of the fertilizer will be left on the soil surface or washed away by rain, causing waste of fertilizer and environmental pollution.

[0005] Therefore, it is necessary to provide a water and fertilizer management method and system for the whole life cycle of blueberry planting for realizing precise supply and intelligent control of water and fertilizer and improving the utilization rate of water and fertilizer. SUMMARY

[0006] The application provides a water and fertilizer management method for the whole life cycle of blueberry planting, comprising: obtaining water and fertilizer management experimental data of multiple blueberry growth stages, wherein the water and fertilizer management experimental data comprises water and fertilizer management schemes of multiple experimental planting areas, soilless cultivation substrate detection data and blueberry growth data, and the water and fertilizer management schemes of any two experimental planting areas are different; based on the water and fertilizer management experimental data of the multiple blueberry growth stages, a growth-related weighted undirected graph of the multiple blueberry growth stages is established; based on the water and fertilizer management experimental data of the multiple blueberry growth stages and the growth-related weighted undirected graph of the multiple blueberry growth stages, key growth factors and key water and fertilizer management factors of each blueberry growth stage are determined; for each blueberry growth stage, based on the water and fertilizer management experimental data of the blueberry growth stage, the key growth factors and the key water and fertilizer management factors, a candidate water and fertilizer management scheme of the blueberry growth stage is determined; and based on the candidate water and fertilizer management scheme of each blueberry growth stage, a water and fertilizer management scheme for the whole life cycle of blueberry planting is generated.

[0007] Further, based on the water and fertilizer management experimental data of the multiple blueberry growth stages, the growth-related weighted undirected graph of the multiple blueberry growth stages is established, comprising: determining growth monitoring factors of each blueberry growth stage; for any two blueberry growth stages, based on the water and fertilizer management experimental data of the two blueberry growth stages, factor values of each growth monitoring factor of the two blueberry growth stages in each experimental planting area are determined, and based on the factor values of each growth monitoring factor of the two blueberry growth stages in each experimental planting area, a growth correlation coefficient of any one growth monitoring factor of one blueberry growth stage and any one growth monitoring factor of another blueberry growth stage is calculated; and based on the growth correlation coefficient of any one growth monitoring factor of one blueberry growth stage and any one growth monitoring factor of another blueberry growth stage corresponding to any two blueberry growth stages, the growth-related weighted undirected graph of the multiple blueberry growth stages is established.

[0008] Further, based on the growth correlation coefficient of any one growth monitoring factor of one blueberry growth stage and any one growth monitoring factor of another blueberry growth stage, a growth correlation weighted undirected graph of multiple blueberry growth stages is established, including: for any two blueberry growth stages, based on the growth correlation coefficient of any one growth monitoring factor of one blueberry growth stage and any one growth monitoring factor of another blueberry growth stage, a growth monitoring factor pair corresponding to the two blueberry growth stages is determined, wherein the growth monitoring factor pair includes one growth monitoring factor of one blueberry growth stage and one growth monitoring factor of another blueberry growth stage; based on the growth monitoring factor pair corresponding to any two blueberry growth stages, a growth correlation weighted undirected graph of any two blueberry growth stages is established, wherein one vertex in the growth correlation weighted undirected graph represents one growth monitoring factor, and the weight of the edge connecting the two vertices of the growth monitoring factor pair is the absolute value of the growth correlation coefficient of the growth monitoring factor pair.

[0009] Further, based on the water and fertilizer management experimental data of multiple blueberry growth stages and the growth correlation weighted undirected graph of multiple blueberry growth stages, the key growth factor and the key water and fertilizer management factor of each blueberry growth stage are determined, including: based on the growth correlation weighted undirected graph of multiple blueberry growth stages, the weight of the growth monitoring factor of each blueberry growth stage is determined; based on the weight of the growth monitoring factor of each blueberry growth stage, the key growth factor of each blueberry growth stage is determined; for each blueberry growth stage, based on the water and fertilizer management experimental data of the blueberry growth stage and the key growth factor of the blueberry growth stage, the key water and fertilizer management factor of the blueberry growth stage is determined.

[0010] Further, based on the growth-related weighted undirected graph of the plurality of blueberry growth stages, the weight of the growth monitoring factor of each blueberry growth stage is determined, comprising: S11, determining a weight optimization path based on the growth-related weighted undirected graph of the plurality of blueberry growth stages; S12, for each blueberry growth stage, initializing the weight of each growth monitoring factor based on the number of growth monitoring factors; S13, determining the current growth monitoring factor according to the weight optimization path; S14, determining the weight correction value of the current growth monitoring factor according to the weight of the edge connected to the vertex corresponding to the current growth monitoring factor in the growth-related weighted undirected graph of the plurality of blueberry growth stages and the weight of the growth monitoring factor connected to the edge, and correcting the weight of the current growth monitoring factor according to the weight correction value of the current growth monitoring factor; S15, judging whether the traversal of the weight optimization path is completed, if yes, executing S16, if not, executing S13; S16, calculating the total weight correction amount; S17, judging whether the correction end condition is met according to the total weight correction amount, if yes, outputting the weight of the growth monitoring factor of each blueberry growth stage, if not, executing S18; S18, resetting the weight optimization path and executing S13.

[0011] Further, based on the water and fertilizer management experimental data of the blueberry growth stage and the key growth factor of the blueberry growth stage, the key water and fertilizer management factor of the blueberry growth stage is determined, comprising: determining a key soilless cultivation substrate state factor based on the water and fertilizer management experimental data of the blueberry growth stage and the key growth factor of the blueberry growth stage; determining the key water and fertilizer management factor of the blueberry growth stage based on the water and fertilizer management experimental data of the blueberry growth stage and the key soilless cultivation substrate state factor.

[0012] Further, based on the water and fertilizer management experimental data of the blueberry growth stage, the key growth factor and the key water and fertilizer management factor, the candidate water and fertilizer management scheme of the blueberry growth stage is determined, comprising: obtaining blueberry yield and quality detection data of a plurality of experimental planting areas; determining the optimal value interval of the key growth factor of each blueberry growth stage based on the blueberry yield and quality detection data of the plurality of experimental planting areas; determining the optimal value interval of the key water and fertilizer management factor based on the optimal value interval of the key growth factor of the blueberry growth stage, and generating the candidate water and fertilizer management scheme of the blueberry growth stage.

[0013] Further, based on the candidate water and fertilizer management scheme of each blueberry growth stage, the water and fertilizer management scheme for the whole life cycle of blueberry planting is generated, comprising: constructing a multi-objective optimization function; generating a plurality of candidate water and fertilizer management schemes for the whole life cycle of blueberry planting based on the candidate water and fertilizer management scheme of each blueberry growth stage; calculating the multi-objective optimization function value of each candidate water and fertilizer management scheme for the whole life cycle of blueberry planting, and generating the water and fertilizer management scheme for the whole life cycle of blueberry planting.

[0014] Further, the multi-objective optimization function is at least related to blueberry yield, blueberry quality, and water and fertilizer cost.

[0015] The application provides a water and fertilizer management system for the whole life cycle of blueberry planting, which applies the water and fertilizer management method for the whole life cycle of blueberry planting, and comprises: a data acquisition module for acquiring water and fertilizer management experimental data of multiple blueberry growth stages, wherein the water and fertilizer management experimental data comprises water and fertilizer management schemes of multiple experimental planting areas, soilless cultivation substrate detection data, and blueberry growth data, and the water and fertilizer management schemes of any two experimental planting areas are different; a data analysis module for establishing a growth-related weighted undirected graph of any two blueberry growth stages based on the water and fertilizer management experimental data of the multiple blueberry growth stages, and determining key growth factors and key water and fertilizer management factors of each blueberry growth stage based on the water and fertilizer management experimental data of the multiple blueberry growth stages and the growth-related weighted undirected graph of the multiple blueberry growth stages; and a scheme optimization module for determining a candidate water and fertilizer management scheme of each blueberry growth stage based on the water and fertilizer management experimental data of the blueberry growth stage, the key growth factors, and the key water and fertilizer management factors, and generating a water and fertilizer management scheme for the whole life cycle of blueberry planting based on the candidate water and fertilizer management scheme of each blueberry growth stage.

[0016] Compared with the prior art, the water and fertilizer management method and system for the whole life cycle of blueberry planting provided by the application have at least the following beneficial effects:

[0017] 1. The traditional method is mainly used for independently designing water and fertilizer strategies for a single growth stage (such as the swelling period and the flower bud differentiation period), which is easy to cause parameter conflicts between stages (such as the contradiction between high potassium in the swelling period and nitrogen control in the flower bud differentiation period). The application establishes a cross-stage parameter linkage model through whole-cycle data integration to ensure the synergy of water and fertilizer schemes in the time dimension. Through the collection and analysis of multi-region and differentiated water and fertilizer management experimental data, the experience bias is eliminated, and the scientificity and universality of the scheme are improved. Based on the growth-related weighted undirected graph, the correlation of growth factors in each stage is quantified, the key growth factors are accurately positioned, and the stage-based precise management is realized. Combined with experimental data and graph models, the key growth factors and water and fertilizer management elements of each growth stage are dynamically selected to avoid resource waste and improve management efficiency. The candidate schemes of each stage are integrated to form a water and fertilizer management strategy covering the whole life cycle of blueberries, ensuring the continuity of growth and improving the overall planting efficiency.

[0018] 2、By constructing a growth-related weighted undirected graph, the correlation of multi-stage growth monitoring factors is quantified in the form of edge weight, eliminating the problem of ignoring the implicit relationship between factors in traditional methods, and realizing the dynamic visualization of factor influence. The iterative correction algorithm is used to calculate the factor weight, and through the closed-loop process of "initialization-path traversal-correction value calculation-total amount verification", the subjective assignment deviation is avoided, and the weight distribution is more in line with the actual growth response law. Based on the weight ordering, the key growth factors are screened, and further traced to the key substrate state factors, and finally the key water and fertilizer management factors are locked, forming a "growth performance-substrate state-water and fertilizer operation" causal tracing chain, and the management target is enhanced by more than 50%.

[0019] 3、Based on multi-region yield and quality detection data, the optimal value interval of key growth factors is determined through statistical analysis, avoiding single region data deviation, making the management target value more in line with the actual production demand. Through the causal chain of key growth factors and key water and fertilizer management factors, the optimal interval of water and fertilizer parameters is reversely deduced, forming a "growth target-water and fertilizer operation" precise mapping, reducing resource waste caused by blind control. Based on the candidate schemes of each stage, a multi-objective optimization function is used to generate a whole cycle scheme, realizing the maximization of whole growth period benefit. BRIEF DESCRIPTION OF DRAWINGS

[0020] The present specification will be further illustrated in the form of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting, and in these embodiments, the same numbers represent the same structures, wherein:

[0021] Figure 1 is a flowchart of a water and fertilizer management method for the whole life cycle of blueberry planting according to some embodiments of the present specification;

[0022] Figure 2 is a schematic diagram of a growth-related weighted undirected graph according to some embodiments of the present specification;

[0023] Figure 3 is a flowchart of determining the weight of the growth monitoring factor of each blueberry growth stage according to some embodiments of the present specification;

[0024] Figure 4 is a flowchart of determining the weight optimization path according to some embodiments of the present specification;

[0025] Figure 5 is a module schematic diagram of a water and fertilizer management system for the whole life cycle of blueberry planting according to some embodiments of the present specification. DETAILED DESCRIPTION

[0026] In order to more clearly illustrate the technical solutions of the embodiments of the present specification, the drawings needed to be used in the embodiment description will be briefly introduced below. Obviously, the drawings in the following description are only some examples or embodiments of the present specification, and for those skilled in the art, the present specification can also be applied to other similar scenarios without creative labor on the basis of these drawings. Unless it is obvious from the language environment or otherwise stated, the same reference numbers in the drawings represent the same structure or operation.

[0027] Figure 1 is a flowchart of a water and fertilizer management method for the whole life cycle of blueberry planting according to some embodiments of the present specification, as shown in Figure 1 The water and fertilizer management method for the whole life cycle of blueberry planting can include the following steps.

[0028] Step 110, obtaining water and fertilizer management experimental data of multiple blueberry growth stages.

[0029] The water and fertilizer management experimental data includes water and fertilizer management schemes of multiple experimental planting areas, soilless culture substrate detection data and blueberry growth data, and the water and fertilizer management schemes of any two experimental planting areas are different.

[0030] Specifically, the multiple blueberry growth stages to which the present method is applied can at least include:

[0031] Nutritional growth period (1-2 years after planting): rapid development of root system and branches and leaves, high nitrogen fertilizer is needed to promote growth.

[0032] Flower bud differentiation period (late autumn to early winter): water and nitrogen control is needed, increase phosphorus and potassium fertilizer to promote flower bud formation.

[0033] Flowering and fruit setting period (spring): stable water supply is needed, supplement boron, calcium and other trace elements to prevent fruit drop.

[0034] Fruit enlargement period (30-60 days after flowering): high potassium fertilizer is needed to improve fruit quality, and moderate water control is needed to increase sugar accumulation.

[0035] Post-harvest recovery period: nitrogen and phosphorus fertilizer is needed to restore tree vigor and reserve nutrients for next year's growth.

[0036] For the same blueberry growth stage, the water and fertilizer management scheme of each experimental planting area needs to have at least one core variable different (such as irrigation frequency, nitrogen, phosphorus and potassium ratio, organic and inorganic fertilizer ratio, etc.), and the rest of the conditions (such as light, temperature, soilless culture substrate type) are as consistent as possible. The experimental planting areas need to be randomly assigned management schemes to avoid location effects (such as light difference in edge area) to interfere with the results.

[0037] The water and fertilizer management scheme can be a comprehensive strategy for irrigation and fertilization formulated for different growth stages of blueberries, including irrigation strategy (e.g., frequency and single water volume (e.g., 2-3 L / plant) of drip irrigation, micro-sprinkling or flooding irrigation), fertilizer ratio (e.g., ratio of nitrogen, phosphorus and potassium), concentration, fertilization frequency, etc.

[0038] For the same blueberry growth stage, only 1-2 core parameters (such as irrigation frequency or fertilizer ratio) of the water and fertilizer management scheme are changed in each experimental planting area, and the rest of the conditions (such as substrate type, light) remain the same. The parameter settings of the water and fertilizer management scheme need to cover the actual application range (such as setting 3 days / time, 5 days / time and 7 days / time for irrigation frequency). The experimental planting area number is randomly assigned to avoid location effects (such as differences in ventilation in the edge area).

[0039] For example, the water and fertilizer management scheme of the three experimental planting areas in the fruit enlargement stage can be as shown in Table 1.

[0040] Table 1

[0041] ;

[0042] The water and fertilizer management scheme of multiple experimental planting areas in multiple blueberry growth stages can be determined in any way. For example, by artificial experience setting. For another example, by orthogonal experiment to set the water and fertilizer management scheme of multiple experimental planting areas in multiple blueberry growth stages.

[0043] The soilless culture substrate (e.g., single or mixed substrate such as coconut husk, peat, perlite, rock wool, sawdust, etc.) is a solid cultivation medium to replace soil and provide physical support and nutrient supply for blueberry root system. The soilless culture substrate detection data can include factor values of soilless culture substrate state factors (e.g., PH value, electrical conductivity, porosity, etc.).

[0044] Blueberry growth data is used to quantitatively describe the morphology, physiological state and yield of blueberry plants. The blueberry growth data can include factor values of multiple growth factors, wherein the multiple growth factors can include morphology factors (e.g., plant height, crown width, new shoot length, leaf number, etc.), physiological factors (e.g., photosynthetic rate, stomatal conductance, leaf nutrient content (e.g., nitrogen, phosphorus, potassium concentration), etc.), yield factors (e.g., number of fruits per plant, single fruit weight, total yield, etc.), etc. The growth factors corresponding to different blueberry growth stages can be different, for example, as shown in Table 2.

[0045] Table 2

[0046] ;

[0047] In step 120, a growth-related weighted undirected graph of multiple blueberry growth stages is established based on the water and fertilizer management experiment data of multiple blueberry growth stages.

[0048] Specifically comprising:

[0049] determining a growth monitoring factor of each blueberry growth stage;

[0050] For any two blueberry growth stages, based on the water and fertilizer management experimental data of the two blueberry growth stages, determining the factor value of each growth monitoring factor of the two blueberry growth stages in each experimental planting area, and based on the factor value of each growth monitoring factor of the two blueberry growth stages in each experimental planting area, calculating the growth correlation coefficient of any one growth monitoring factor of one blueberry growth stage and any one growth monitoring factor of another blueberry growth stage;

[0051] Based on the growth correlation coefficient of any one growth monitoring factor of one blueberry growth stage and any one growth monitoring factor of another blueberry growth stage corresponding to any two blueberry growth stages, a growth correlation weighted undirected graph of multiple blueberry growth stages is established.

[0052] Specifically, the factor value of each growth monitoring factor of the blueberry growth stage in each experimental planting area can be the value at a certain moment (for example, the end moment) of the blueberry growth stage. Any one growth monitoring factor of one blueberry growth stage and any one growth monitoring factor of another blueberry growth stage can be taken as two dependent variables, and the factor values of the dependent variables in each experimental planting area are substituted into the correlation coefficient (for example, Pearson correlation coefficient, Spearman rank correlation coefficient, etc.) calculation formula to calculate the growth correlation coefficient of any one growth monitoring factor of one blueberry growth stage and any one growth monitoring factor of another blueberry growth stage.

[0053] For example, only as an example, taking the length of new shoots in the vegetative growth period and the fruit transverse diameter in the fruit swelling period as an example, the length of new shoots in the vegetative growth period and the fruit transverse diameter in the fruit swelling period of each experimental planting area are substituted into the correlation coefficient formula to calculate the growth correlation coefficient of the length of new shoots in the vegetative growth period and the fruit transverse diameter in the fruit swelling period.

[0054] As preferred, based on the growth correlation coefficient of any one growth monitoring factor of one blueberry growth stage and any one growth monitoring factor of another blueberry growth stage corresponding to any two blueberry growth stages, a growth correlation weighted undirected graph of multiple blueberry growth stages is established, including:

[0055] For any two blueberry growth stages, based on the growth correlation coefficient of any one growth monitoring factor of one blueberry growth stage and any one growth monitoring factor of another blueberry growth stage, a growth monitoring factor pair corresponding to the two blueberry growth stages is determined, wherein the growth monitoring factor pair includes one growth monitoring factor of one blueberry growth stage and one growth monitoring factor of another blueberry growth stage, for example, if the absolute value of the growth correlation coefficient of any one growth monitoring factor of one blueberry growth stage and any one growth monitoring factor of another blueberry growth stage is greater than a preset threshold (for example, 0.5), the two growth monitoring factors are taken as the growth monitoring factor pair corresponding to the two blueberry growth stages, for example, the absolute value of the growth correlation coefficient of the new shoot length of the vegetative growth period and the fruit transverse diameter of the fruit swelling period is greater than the preset threshold (for example, 0.5), and the growth monitoring factor pair of the vegetative growth period and the fruit swelling period is: new shoot length-fruit transverse diameter. It can be understood that there can be at most one same growth monitoring factor in any two growth monitoring factor pairs.

[0056] Based on the growth monitoring factor pair corresponding to any two blueberry growth stages, a growth correlation weighted undirected graph of any two blueberry growth stages is established, wherein, as shown in Figure 2 , one vertex in the growth correlation weighted undirected graph represents one growth monitoring factor, and the weight of the edge connecting the two vertices of the growth monitoring factor pair is the absolute value of the growth correlation coefficient of the growth monitoring factor pair.

[0057] Step 130, based on the water and fertilizer management experimental data of the plurality of blueberry growth stages and the growth correlation weighted undirected graph of the plurality of blueberry growth stages, determining the key growth factor and the key water and fertilizer management factor of each blueberry growth stage.

[0058] Specifically includes:

[0059] Based on the growth correlation weighted undirected graph of the plurality of blueberry growth stages, the weight of the growth monitoring factor of each blueberry growth stage is determined;

[0060] Based on the weight of the growth monitoring factor of each blueberry growth stage, the key growth factor of each blueberry growth stage is determined;

[0061] For each blueberry growth stage, based on the water and fertilizer management experimental data of the blueberry growth stage and the key growth factor of the blueberry growth stage, the key water and fertilizer management factor of the blueberry growth stage is determined.

[0062] Figure 3 is a flowchart of determining the weight of the growth monitoring factor of each blueberry growth stage according to some embodiments of the present specification, as shown in Figure 3As shown, as preferred, based on the growth-related weighted undirected graph of multiple blueberry growth stages, the weight of the growth monitoring factor of each blueberry growth stage is determined, including:

[0063] S11, based on the growth-related weighted undirected graph of multiple blueberry growth stages, the weight optimization path is determined;

[0064] S12, for each blueberry growth stage, based on the number of growth monitoring factors, the weight of each growth monitoring factor is initialized, for example, the initial value of the weight of each growth monitoring factor = 1 / n, wherein n is the total number of growth monitoring factors of the blueberry growth stage;

[0065] S13, according to the weight optimization path, the current growth monitoring factor is determined;

[0066] S14, according to the weight of the edge connected to the vertex corresponding to the current growth monitoring factor and the weight of the growth monitoring factor connected to the edge in the growth-related weighted undirected graph of multiple blueberry growth stages, the weight correction value of the current growth monitoring factor is determined, and the weight of the current growth monitoring factor is corrected according to the weight correction value of the current growth monitoring factor. Specifically, the weight of the edge connected to the vertex corresponding to the current growth monitoring factor and the weight of the growth monitoring factor connected to the edge can be weighted and summed to obtain the weight correction value of the current growth monitoring factor. The weight of the current growth monitoring factor and the weight correction value are summed to obtain the corrected weight of the current growth monitoring factor;

[0067] S15, it is judged whether the traversal of the weight optimization path is completed, that is, the last growth monitoring factor in the weight optimization path is accessed, if yes, S16 is executed, if no, S13 is executed;

[0068] S16, the total weight correction amount is calculated, wherein the total weight correction amount is the sum of the absolute values of the differences between the uncorrected weight and the corrected weight of each growth monitoring factor;

[0069] S17, according to the total weight correction amount, it is judged whether the correction end condition is met, if yes, the weight of the growth monitoring factor of each blueberry growth stage is output, if no, S18 is executed, wherein the correction end condition can be that the number of iterations reaches a maximum value (for example, 20 times), the total weight correction amount of the two consecutive iterations is less than a weight correction threshold (for example, 2), etc.

[0070] S18, one iteration is completed, the weight optimization path is reset, that is, the weight optimization path is traversed from the beginning, and S13 is executed.

[0071] For example, the weight correction value of the current growth monitoring factor can be calculated based on the following formula:

[0072] ;

[0073] wherein, is a weight correction value of the current growth monitoring factor (i.e., the ith growth monitoring factor), is a weight of the nth edge of the vertex corresponding to the current growth monitoring factor in the growth-related weighted undirected graph, is a weight of the vertex pair connected by the nth edge of the vertex corresponding to the current growth monitoring factor in the growth-related weighted undirected graph, is a total number of edges of the vertex corresponding to the current growth monitoring factor in the growth-related weighted undirected graph.

[0074] It can be understood that the above process outputs the weight value directly quantifying the contribution of each growth monitoring factor to the growth of blueberries through the edge weight in the weighted undirected graph and the iterative correction mechanism. Based on the growth-related weighted undirected graph, the direct correlation between monitoring factors is directly quantified by the edge weight, avoiding subjective assignment bias. The mean initialization weight (1 / n ) ensures fairness and provides a neutral starting point for subsequent optimization. The neighbor factor weighting correction mechanism: the weight correction value of the current factor is calculated by weighted summation of the neighbor factor weight and the connection edge weight, which comprehensively considers the direct correlation (edge weight) and indirect regulation (neighbor factor weight transmission), avoiding the bias caused by relying on a single path. Multiple iterations gradually smooth the data noise, making the weight converge to the global optimum and enhancing the robustness of the results. The process realizes the objectivity, accuracy and adaptability of weight allocation through quantification of correlation, dynamic optimization and stage adaptation.

[0075] Figure 4 is a flowchart of determining a weight optimization path according to some embodiments of the present specification, as shown in Figure 4 determining a weight optimization path can include the following steps:

[0076] S21, generating a plurality of weighted undirected graphs, and determining an optimal weight optimization path for each weighted undirected graph;

[0077] S22, for each sample weighted undirected graph, determining the value of a plurality of graph feature factors (for example, the number of connected edges, the sum of the weights of the connected edges, the average number of edges of the connected vertices, the average sum of the weights of the edges of the connected vertices, etc.) of each vertex and the sequence number of the vertex in the optimal weight optimization path, for each graph feature factor, substituting the value of the graph feature factor of each vertex and the sequence number of the vertex in the optimal weight optimization path into the calculation formula of the correlation coefficient (for example, Pearson correlation coefficient, Spearman rank correlation coefficient, etc.), and calculating to obtain the correlation coefficient of the graph feature factor and the sequence number corresponding to the sample weighted undirected graph;

[0078] S23, for each graph feature factor, calculate the average of the correlation coefficient of the graph feature factor corresponding to each sample weighted undirected graph and the serial number as the correlation coefficient average of the graph feature factor;

[0079] S24, the absolute value of the correlation coefficient average is greater than the average threshold (for example, 0.4, etc.) as the key graph feature factor, and the correlation coefficient average is taken as the weight of the key graph feature factor;

[0080] S25, for each vertex in the growth-related weighted undirected graph, determine the factor value of the key graph feature factor of the vertex, and perform weighted summation on the factor value of the key graph feature factor of the vertex according to the weight of the key graph feature factor, and calculate the ranking value of the vertex;

[0081] S26, according to the ranking value of each vertex in the growth-related weighted undirected graph, the vertices in the growth-related weighted undirected graph are ranked in descending order, and the weight optimization path is generated.

[0082] It can be understood that the above process generates multiple weighted undirected graphs and determines the optimal path respectively, avoids the limitation of single graph structure or optimization strategy through multi-path comparison, improves the global search ability, and ensures the robustness and adaptability of the final path. The correlation coefficient average of each graph feature and the optimal path serial number is calculated, the key features with absolute value greater than the threshold are selected, the factors with strong correlation with path ranking are retained, and the noise features are removed. Through multi-graph exploration, feature quantization, correlation screening and dynamic sorting, the scientific, accurate and interpretable weight optimization path is realized, which provides a complete solution from graph structure analysis to path generation for blueberry growth monitoring factor weight distribution.

[0083] Specifically, the optimal weight optimization path of each weighted undirected graph can be determined according to the following process:

[0084] S31, according to the weighted undirected graph, randomly generate multiple weight optimization paths, specifically, randomly determine the serial number of each vertex in the weight optimization path from the weighted undirected graph, thereby generating the weight optimization path;

[0085] S32, set multiple path evaluation indexes, for example, iteration number index, average value index of weight correction total amount of single iteration, etc.;

[0086] S33, based on the process of S13-S18, determine the correction result of each weight optimization path, wherein the correction result includes iteration number, average value of weight correction total amount of single iteration;

[0087] S34, for each weight optimization path, determine the value of the weight optimization path in each path evaluation index, wherein the more the number of iterations, the smaller the score of the number of iterations index, the larger the mean value of the total weight correction of single iteration, the smaller the score of the mean value index of the total weight correction of single iteration, sum the values of the weight optimization path in each path evaluation index to obtain the optimization value of the weight optimization path;

[0088] S35, the weight optimization path with the maximum optimization value is taken as the optimal weight optimization path.

[0089] It can be understood that generating multiple paths by randomly assigning vertex sequence numbers breaks through the limitation of a single initial path, covers a wider solution space, avoids falling into local optimization, and improves the global optimization probability. Setting the number of iterations, the mean value of the total correction of single iteration, and other indicators evaluates the path from the efficiency and stability dimensions. For example, a large number of iterations may reflect slow convergence of the path, and a large mean value of the total correction may indicate that the weight fluctuates sharply, both of which need to be suppressed. By comprehensive scoring, the path with the maximum optimization value is selected to ensure that the final path performs optimally in global search and provides a data basis for subsequent determination of the weight optimization path.

[0090] For each blueberry growth stage, the weight of each growth monitoring factor for each blueberry growth stage can be normalized, and the growth monitoring factor with a normalized weight greater than a weight threshold (e.g., 0.2) is taken as a key growth monitoring factor.

[0091] Preferably, based on the water and fertilizer management experimental data of the blueberry growth stage and the key growth factors of the blueberry growth stage, the key water and fertilizer management factors of the blueberry growth stage are determined, including:

[0092] Based on the water and fertilizer management experimental data of the blueberry growth stage and the key growth factors of the blueberry growth stage, the key soilless cultivation substrate state factors are determined.

[0093] Based on the water and fertilizer management experimental data of the blueberry growth stage and the key soilless cultivation substrate state factors, the key water and fertilizer management factors of the blueberry growth stage are determined.

[0094] Specifically, referring to the calculation method of the growth correlation coefficient of two growth monitoring factors, the correlation analysis of the soilless cultivation substrate state factors and the key growth factors is performed, and the correlation coefficient of the soilless cultivation substrate state factors and the key growth factors is calculated.

[0095] For each soilless cultivation substrate state factor, the mean value of the absolute value of the correlation coefficient of the soilless cultivation substrate state factor and each key growth factor is calculated, and the soilless cultivation substrate state factor with a mean value greater than a mean value threshold (e.g., 0.4) is taken as a key soilless cultivation substrate state factor.

[0096] The correlation between the water and fertilizer management factors and the key soilless cultivation substrate state factors is analyzed according to the calculation manner of the growth correlation coefficient of the two growth monitoring factors, and the correlation coefficient of the soilless cultivation substrate state factors and the key growth factors is calculated.

[0097] For each water and fertilizer management factor, the mean of the absolute values of the correlation coefficients of the soilless cultivation substrate state factors and each key growth factor is calculated, and the water and fertilizer management factor with a mean greater than a mean threshold value (for example, 0.4) is taken as a key water and fertilizer management factor.

[0098] In step 140, for each blueberry growth stage, a candidate water and fertilizer management scheme for the blueberry growth stage is determined based on the water and fertilizer management experimental data of the blueberry growth stage, the key growth factors and the key water and fertilizer management factors.

[0099] Specifically, it includes:

[0100] The blueberry yield and quality detection data of the multiple experimental planting areas are obtained.

[0101] Based on the blueberry yield and quality detection data of the multiple experimental planting areas, the optimal value interval of the key growth factors of each blueberry growth stage is determined. Specifically, the blueberry yield and quality detection data of the multiple experimental planting areas are used to determine the optimal experimental planting area, for example, the top 20% of the data in terms of yield / quality is taken as the optimal experimental planting area, the factor values of the key growth factors of the optimal experimental planting area at each blueberry growth stage are obtained, and the maximum and minimum values of the key growth factors are taken to form the optimal value interval of the key growth factors.

[0102] Based on the optimal value interval of the key growth factors of the blueberry growth stage, the optimal value interval of the key water and fertilizer management factor is determined, and the candidate water and fertilizer management scheme for the blueberry growth stage is generated. Specifically, the experimental planting area with the factor value of the key growth factor of the blueberry growth stage within the optimal value interval is taken as the candidate experimental planting area, and the maximum and minimum values of the key water and fertilizer management factor in the water and fertilizer management scheme of the candidate experimental planting area are used to determine the optimal value interval of the key water and fertilizer management factor.

[0103] For each key water and fertilizer management factor, a value is sampled from the optimal value interval of the key water and fertilizer management factor as the value of the key water and fertilizer management factor in the candidate water and fertilizer management scheme, thereby generating multiple candidate water and fertilizer management schemes for the blueberry growth stage.

[0104] In step 150, a water and fertilizer management scheme for the whole life cycle of blueberry planting is generated based on the candidate water and fertilizer management schemes for each blueberry growth stage.

[0105] Specifically, it includes:

[0106] constructing a multi-objective optimization function, wherein the multi-objective optimization function is related to at least blueberry yield, blueberry quality, and water and fertilizer cost;

[0107] generating a plurality of candidate water and fertilizer management schemes for the whole life cycle of blueberry planting based on the candidate water and fertilizer management schemes for each blueberry growth stage, wherein for each blueberry growth stage, a candidate water and fertilizer management scheme is sampled from the plurality of candidate water and fertilizer management schemes for the blueberry growth stage, and the sampled candidate water and fertilizer management schemes corresponding to each blueberry growth stage are combined to generate a candidate water and fertilizer management scheme for the whole life cycle of blueberry planting;

[0108] calculating the multi-objective optimization function value of each candidate water and fertilizer management scheme for the whole life cycle of blueberry planting to generate a water and fertilizer management scheme for the whole life cycle of blueberry planting.

[0109] Specifically, the blueberry yield is quantified by fruit weight per unit area (kg / acre), and the quality is quantified by comprehensive sugar content (°Brix), hardness (kg / cm²), and anthocyanin content (mg / 100g), etc. The comprehensive score is calculated by principal component analysis. The water and fertilizer cost includes direct costs such as fertilizer procurement, irrigation energy consumption, and manual application.

[0110] A normalization weighted summation method is used to construct the multi-objective function to eliminate dimensional differences.

[0111] The yield and quality of blueberries can be predicted according to the candidate water and fertilizer management scheme for the whole life cycle of blueberry planting by using a result prediction model. The result prediction model can be a long short-term memory network model. The result prediction model uses a bidirectional structure to capture the forward influence (such as early nitrogen fertilizer promoting vegetative growth) and backward feedback (such as later quality constraints on early water and fertilizer) of water and fertilizer input. The hidden layer dimension is set to 64, and the output dimension is 128 (bidirectional splicing). The yield and quality (such as sugar content, hardness, and anthocyanin content) are predicted by independent fully connected layers. 200 groups of historical planting data are collected to generate 800 groups of simulation data, which are divided into training set, validation set, and test set in the ratio of 8:1:1. The yield task uses mean square error, and the quality task uses weighted mean square error of multi-dimensional indicators (such as sugar content weight 0.5, hardness 0.3, and anthocyanin 0.2). The total loss is the weighted sum of the two tasks (weight 1:0.8). The Adam optimizer (learning rate 0.001) is used with learning rate decay (decaying to 0.9 times every 10 rounds) and batch size set to 32. The training rounds are 200, and early stopping (Patience=20) is used to prevent overfitting.

[0112] Based on the output of the result prediction model, the price, the amount of use and the amount of water of the fertilizer used by the candidate water and fertilizer management scheme, the multi-objective optimization function value of the candidate water and fertilizer management scheme for the whole life cycle of blueberry planting is calculated, and the candidate water and fertilizer management scheme for the whole life cycle of blueberry planting with the maximum multi-objective optimization function value is taken as the water and fertilizer management scheme for the whole life cycle of blueberry planting.

[0113] Figure 5 is a schematic diagram of a module of a water and fertilizer management system for the whole life cycle of blueberry planting according to some embodiments of the present specification, as shown in Figure 5 As shown, the water and fertilizer management system for the whole life cycle of blueberry planting can include a data acquisition module, a data analysis module and a scheme optimization module.

[0114] The data acquisition module is configured to acquire water and fertilizer management experimental data of multiple blueberry growth stages, wherein the water and fertilizer management experimental data includes water and fertilizer management schemes of multiple experimental planting areas, soilless cultivation substrate detection data and blueberry growth data, and the water and fertilizer management schemes of any two experimental planting areas are different.

[0115] The data analysis module is configured to establish a growth-related weighted undirected graph of any two blueberry growth stages based on the water and fertilizer management experimental data of the multiple blueberry growth stages, and determine key growth factors and key water and fertilizer management factors of each blueberry growth stage based on the water and fertilizer management experimental data of the multiple blueberry growth stages and the growth-related weighted undirected graph of the multiple blueberry growth stages.

[0116] The scheme optimization module is configured to determine a candidate water and fertilizer management scheme for each blueberry growth stage based on the water and fertilizer management experimental data of the blueberry growth stage, the key growth factors and the key water and fertilizer management factors, and generate a water and fertilizer management scheme for the whole life cycle of blueberry planting based on the candidate water and fertilizer management scheme of each blueberry growth stage.

[0117] The water and fertilizer management system for the whole life cycle of blueberry planting can apply the water and fertilizer management method for the whole life cycle of blueberry planting described above, which will not be repeated here.

[0118] Finally, it should be understood that the embodiments described in the present specification are only used to illustrate the principles of the embodiments of the present specification. Other variations can also belong to the scope of the present specification. Therefore, as an example but not limitation, alternative configurations of the embodiments of the present specification can be considered consistent with the teachings of the present specification. Accordingly, the embodiments of the present specification are not limited to the embodiments explicitly introduced and described in the present specification.

Claims

1. A water and fertilizer management method for the entire life cycle of blueberry cultivation, characterized by, The method comprises the following steps: Obtain water and fertilizer management experimental data of multiple blueberry growth stages, wherein the water and fertilizer management experimental data comprises water and fertilizer management schemes of multiple experimental planting areas, soilless cultivation substrate detection data and blueberry growth data, and the water and fertilizer management schemes of any two experimental planting areas are different; Based on the water and fertilizer management experimental data of the multiple blueberry growth stages, a growth-related weighted undirected graph of the multiple blueberry growth stages is established; Based on the water and fertilizer management experimental data of the multiple blueberry growth stages and the growth-related weighted undirected graph of the multiple blueberry growth stages, key growth factors and key water and fertilizer management factors of each blueberry growth stage are determined; For each blueberry growth stage, based on the water and fertilizer management experimental data of the blueberry growth stage, the key growth factors and the key water and fertilizer management factors, a candidate water and fertilizer management scheme of the blueberry growth stage is determined; Based on the candidate water and fertilizer management scheme of each blueberry growth stage, a water and fertilizer management scheme for the whole life cycle of blueberry planting is generated; Based on the water and fertilizer management experimental data of the multiple blueberry growth stages, a growth-related weighted undirected graph of the multiple blueberry growth stages is established, comprising: Determine the growth monitoring factors of each blueberry growth stage; For any two blueberry growth stages, based on the water and fertilizer management experimental data of the two blueberry growth stages, the factor values of each growth monitoring factor of the two blueberry growth stages in each experimental planting area are determined, and based on the factor values of each growth monitoring factor of the two blueberry growth stages in each experimental planting area, the growth correlation coefficients of any one growth monitoring factor of one blueberry growth stage and any one growth monitoring factor of another blueberry growth stage are calculated; Based on the growth correlation coefficients of any one growth monitoring factor of one blueberry growth stage and any one growth monitoring factor of another blueberry growth stage corresponding to any two blueberry growth stages, a growth-related weighted undirected graph of the multiple blueberry growth stages is established; Based on the growth correlation coefficients of any one growth monitoring factor of one blueberry growth stage and any one growth monitoring factor of another blueberry growth stage corresponding to any two blueberry growth stages, a growth-related weighted undirected graph of the multiple blueberry growth stages is established, comprising: For any two blueberry growth stages, based on the growth correlation coefficients of any one growth monitoring factor of one blueberry growth stage and any one growth monitoring factor of another blueberry growth stage, the growth monitoring factor pairs corresponding to the two blueberry growth stages are determined, wherein the growth monitoring factor pair comprises one growth monitoring factor of one blueberry growth stage and one growth monitoring factor of another blueberry growth stage; Based on the growth monitoring factor pairs corresponding to any two blueberry growth stages, a growth-related weighted undirected graph of the two blueberry growth stages is established, wherein one vertex in the growth-related weighted undirected graph represents one growth monitoring factor, and the weight of the edge connecting the two vertices of the growth monitoring factor pair is the absolute value of the growth correlation coefficient of the growth monitoring factor pair.

2. The water and fertilizer management method for the whole life cycle of blueberry cultivation according to claim 1, characterized by, Based on the water and fertilizer management experimental data of multiple blueberry growth stages and the growth-related weighted undirected graph of multiple blueberry growth stages, the key growth factors and key water and fertilizer management factors of each blueberry growth stage are determined, including: Based on the growth-related weighted undirected graph of multiple blueberry growth stages, the weight of the growth monitoring factor of each blueberry growth stage is determined. Based on the weight of the growth monitoring factor of each blueberry growth stage, the key growth factor of each blueberry growth stage is determined. For each blueberry growth stage, based on the water and fertilizer management experimental data of the blueberry growth stage and the key growth factor of the blueberry growth stage, the key water and fertilizer management factor of the blueberry growth stage is determined.

3. The water and fertilizer management method for the whole life cycle of blueberry cultivation according to claim 2, characterized by, Based on the growth-related weighted undirected graph of multiple blueberry growth stages, the weight of the growth monitoring factor of each blueberry growth stage is determined, including: S11, based on the growth-related weighted undirected graph of multiple blueberry growth stages, the weight optimization path is determined; S12, for each blueberry growth stage, based on the number of growth monitoring factors, the weight of each growth monitoring factor is initialized; S13, according to the weight optimization path, the current growth monitoring factor is determined; S14, according to the weight of the edge connected by the vertex corresponding to the current growth monitoring factor in the growth-related weighted undirected graph of multiple blueberry growth stages, and the weight of the growth monitoring factor connected by the edge, the weight correction value of the current growth monitoring factor is determined, and the weight of the current growth monitoring factor is corrected according to the weight correction value of the current growth monitoring factor; S15, judge whether the traversal of the weight optimization path is completed, if yes, execute S16, if not, execute S13; S16, calculate the total weight correction amount; S17, according to the total weight correction amount, judge whether the correction end condition is met, if yes, output the weight of the growth monitoring factor of each blueberry growth stage, if not, execute S18; S18, reset the weight optimization path, execute S13.

4. The water and fertilizer management method for the whole life cycle of blueberry cultivation according to claim 2, characterized by, Based on the water and fertilizer management experimental data of the blueberry growth stage and the key growth factor of the blueberry growth stage, the key water and fertilizer management factor of the blueberry growth stage is determined, including: Based on the water and fertilizer management experimental data of the blueberry growth stage and the key growth factor of the blueberry growth stage, the key soilless cultivation substrate state factor is determined; Based on the water and fertilizer management experimental data of the blueberry growth stage and the key soilless cultivation substrate state factor, the key water and fertilizer management factor of the blueberry growth stage is determined.

5. The water and fertilizer management method for the whole life cycle of blueberry cultivation according to any one of claims 1 to 4, characterized by, Based on the water and fertilizer management experimental data of the blueberry growth stage, the key growth factor and the key water and fertilizer management factor, the candidate water and fertilizer management scheme of the blueberry growth stage is determined, including: Obtain the blueberry yield and quality detection data of multiple experimental planting areas; Based on the blueberry yield and quality detection data of multiple experimental planting areas, the optimal value interval of the key growth factor of each blueberry growth stage is determined; Based on the optimal value interval of the key growth factor of the blueberry growth stage, the optimal value interval of the key water and fertilizer management factor is determined, and the candidate water and fertilizer management scheme of the blueberry growth stage is generated.

6. The water and fertilizer management method for the whole life cycle of blueberry cultivation according to claim 5, characterized by, Based on the candidate water and fertilizer management scheme of each blueberry growth stage, the water and fertilizer management scheme for the whole life cycle of blueberry planting is generated, including: Construct a multi-objective optimization function; generate a plurality of candidate water and fertilizer management schemes for the whole life cycle of blueberry planting based on the candidate water and fertilizer management schemes of each blueberry growth stage; calculate the multi-objective optimization function value of each candidate water and fertilizer management scheme for the whole life cycle of blueberry planting, and generate a water and fertilizer management scheme for the whole life cycle of blueberry planting.

7. The water and fertilizer management method for the whole life cycle of blueberry cultivation according to claim 6, characterized by, The multi-objective optimization function is at least related to blueberry yield, blueberry quality and water and fertilizer cost.

8. A water and fertilizer management system for the entire life cycle of blueberry cultivation, characterized by, The application of the water and fertilizer management method for the whole life cycle of blueberry planting according to any one of claims 1-7 comprises: a data acquisition module for acquiring water and fertilizer management experimental data of a plurality of blueberry growth stages, wherein the water and fertilizer management experimental data comprises water and fertilizer management schemes of a plurality of experimental planting areas, soilless cultivation substrate detection data and blueberry growth data, and the water and fertilizer management schemes of any two experimental planting areas are different; a data analysis module for establishing a growth-related weighted undirected graph of any two blueberry growth stages based on the water and fertilizer management experimental data of a plurality of blueberry growth stages, and determining key growth factors and key water and fertilizer management factors of each blueberry growth stage based on the water and fertilizer management experimental data of a plurality of blueberry growth stages and the growth-related weighted undirected graph of a plurality of blueberry growth stages; a scheme optimization module for determining, for each blueberry growth stage, a candidate water and fertilizer management scheme based on the water and fertilizer management experimental data, the key growth factors and the key water and fertilizer management factors of the blueberry growth stage, and generating a water and fertilizer management scheme for the whole life cycle of blueberry planting based on the candidate water and fertilizer management scheme of each blueberry growth stage.

Images