Gain performance evaluation method and system for compound products

Abstract

The present invention discloses a gain performance evaluation method and system for a chemical product, and relates to the technical field of performance evaluation. The method comprises: obtaining multiple soil monitoring data and multiple plant status monitoring data; performing neighborhood similarity fusion on multiple fertilization grids to obtain multiple fused fertilization areas; fertilizing the target chemical product to multiple fused fertilization areas, and monitoring, determining multiple window soil monitoring data and multiple window plant status monitoring data; performing mapping gain analysis, determining multiple regional soil gain coefficients and multiple regional plant status gain coefficients; performing centralized screening analysis, determining centralized soil gain coefficients and centralized plant status gain coefficients; performing comprehensive analysis, determining target performance gain coefficients, and using the target performance gain coefficients as gain performance evaluation results. The method solves the technical problem of inaccurate performance evaluation of chemical products in the prior art, and achieves the technical effect of improving the accuracy of performance evaluation of chemical products.

Description

Technical Field

[0001] The present invention relates to the technical field of performance evaluation, and in particular to a method and system for evaluating the gain performance of a compound product. Background Art

[0002] In the field of modern agriculture, the use of chemical products is of great significance for increasing crop yields, improving soil quality and promoting sustainable agricultural development. However, the performance evaluation of chemical products has always been a complex and critical issue in agricultural production. Traditional methods of evaluating the effect of fertilization usually only judge based on a single indicator of soil or plants, which makes it difficult to fully reflect the dynamic changes in soil and plant status after fertilization. At the same time, there are large differences in soil conditions and plant growth status in different fertilization areas, and the existing fertilization management methods lack targeted analysis of conditions in different regions, resulting in uncertainty in the gain effect of chemical products. In addition, the frequency of data collection after fertilization is low and the data analysis method is simple, making it impossible to monitor and evaluate the fertilization effect in real time and accurately. Summary of the invention

[0003] The present application provides a method and system for evaluating the gain performance of a compound product, which solves the technical problem of inaccurate performance evaluation of a compound product in the prior art.

[0004] In view of the above problems, the present application provides a method and system for evaluating the gain performance of a compound product.

[0005] In a first aspect of the present application, a method for evaluating the gain performance of a compound product is provided, the method comprising:

[0006] The target fertilization area is equally divided into grids, and soil monitoring data and plant status monitoring data sets are collected for the divided multiple fertilization grids to obtain multiple soil monitoring data and multiple plant status monitoring data, wherein each fertilization grid corresponds to a grid position; based on the multiple soil monitoring data and the multiple plant status monitoring data, neighbor grid similarity analysis is performed, and the multiple fertilization grids are subjected to neighborhood similarity fusion to obtain multiple fused fertilization areas, wherein the multiple fused fertilization areas include multiple fused area soil monitoring data and multiple fused plant status monitoring data; at the fertilization node, the target chemical product is fertilized to the multiple fused fertilization areas according to the preset fertilization plan, and soil monitoring and plant status monitoring are performed on the multiple fused fertilization areas after fertilization in preset interval windows to determine multiple window soil monitoring data and multiple window plant status monitoring data; according to the multiple fused area soil monitoring data, multiple fused plant status monitoring data, the multiple fused area soil monitoring data and multiple fused area plant status monitoring data are obtained. The monitoring data, the soil monitoring data of multiple windows and the plant status monitoring data of multiple windows are mapped and analyzed for gain, and the soil gain coefficients of multiple regions and the plant status gain coefficients of multiple regions are determined; the steady-state evaluation is performed on the soil gain coefficients of the multiple regions and the plant status gain coefficients of the multiple regions respectively, and the soil screening step length and the plant screening step length are determined according to the evaluation results; based on the soil screening step length and the plant screening step length, as well as the screening starting points corresponding to the soil gain coefficients of the multiple regions and the plant status gain coefficients of the multiple regions, the soil gain coefficients of the multiple regions and the plant status gain coefficients of the multiple regions are concentratedly screened and analyzed to determine the concentrated soil gain coefficient and the concentrated plant status gain coefficient; the performance evaluation network layer is used to comprehensively analyze the concentrated soil gain coefficient and the concentrated plant status gain coefficient to determine the target performance gain coefficient of the target compound product, and the target performance gain coefficient is used as the gain performance evaluation result.

[0007] In a second aspect of the present application, a system for evaluating the gain performance of a compound product is provided, the system comprising:

[0008] A data acquisition module, the data acquisition module is used to equally divide the target fertilization area into grids, and collect soil monitoring data and plant status monitoring data sets for the divided multiple fertilization grids to obtain multiple soil monitoring data and multiple plant status monitoring data, wherein each fertilization grid corresponds to a grid position; a fusion module, the fusion module is used to perform neighbor grid similarity analysis based on the multiple soil monitoring data and the multiple plant status monitoring data, perform neighborhood similarity fusion on the multiple fertilization grids, and obtain multiple fused fertilization areas, wherein the multiple fused fertilization areas include multiple fused area soil monitoring data and multiple fused plant status monitoring data; a monitoring module, the monitoring module is used to fertilize the multiple fused fertilization areas with the target compound product according to the pre-set fertilization plan at the fertilization node, and perform soil monitoring and plant status monitoring on the multiple fused fertilization areas after fertilization in preset interval windows to determine multiple window soil monitoring data and multiple window plant status monitoring data; a gain analysis module, the gain analysis module is used to determine the multiple window soil monitoring data and multiple window plant status monitoring data based on the multiple fused area soil Monitoring data, multiple fused plant status monitoring data, multiple window soil monitoring data and multiple window plant status monitoring data are mapped and analyzed to determine multiple regional soil gain coefficients and multiple regional plant status gain coefficients; a screening analysis module, the screening analysis module is used to perform steady-state evaluation on the multiple regional soil gain coefficients and the multiple regional plant status gain coefficients respectively, determine the soil screening step and the plant screening step according to the evaluation results, based on the soil screening step and the plant screening step, and the screening starting points corresponding to the multiple regional soil gain coefficients and the multiple regional plant status gain coefficients respectively, perform centralized screening analysis on the multiple regional soil gain coefficients and the multiple regional plant status gain coefficients to determine the centralized soil gain coefficient and the centralized plant status gain coefficient; a comprehensive analysis module, the comprehensive analysis module is used to use the performance evaluation network layer to perform a comprehensive analysis on the centralized soil gain coefficient and the centralized plant status gain coefficient to determine the target performance gain coefficient of the target compound product, and use the target performance gain coefficient as the gain performance evaluation result.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] First, the target fertilization area is equally divided into grids, and soil monitoring data and plant status monitoring data sets are collected for the divided multiple fertilization grids to obtain multiple soil monitoring data and multiple plant status monitoring data, wherein each fertilization grid corresponds to a grid position. Next, neighbor grid similarity analysis is performed based on multiple soil monitoring data and multiple plant status monitoring data, and neighborhood similarity fusion is performed on multiple fertilization grids to obtain multiple fused fertilization areas, wherein the multiple fused fertilization areas include multiple fused area soil monitoring data and multiple fused plant status monitoring data. At the fertilization node, the target compound product is fertilized to multiple fused fertilization areas according to the pre-set fertilization scheme, and soil monitoring and plant status monitoring are performed on multiple fused fertilization areas after fertilization in a preset interval window to determine multiple window soil monitoring data and multiple window plant status monitoring data. Next, mapping gain analysis is performed based on multiple fused area soil monitoring data, multiple fused plant status monitoring data, multiple window soil monitoring data, and multiple window plant status monitoring data to determine multiple area soil gain coefficients and multiple area plant status gain coefficients. Then, a steady-state evaluation is performed on the soil gain coefficients of multiple regions and the plant state gain coefficients of multiple regions respectively, and the soil screening step length and the plant screening step length are determined according to the evaluation results. Based on the soil screening step length and the plant screening step length, as well as the screening starting points corresponding to the soil gain coefficients of multiple regions and the plant state gain coefficients of multiple regions, the soil gain coefficients of multiple regions and the plant state gain coefficients of multiple regions are centrally screened and analyzed to determine the central soil gain coefficient and the central plant state gain coefficient. Finally, the performance evaluation network layer is used to comprehensively analyze the central soil gain coefficient and the central plant state gain coefficient to determine the target performance gain coefficient of the target compound product, and the target performance gain coefficient is used as the gain performance evaluation result. The technical problem of inaccurate performance evaluation of compound products in the prior art is solved, and the technical effect of improving the accuracy of performance evaluation of compound products is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0012] Figure 1 A schematic flow chart of a method for evaluating the gain performance of a compound product provided in an embodiment of the present application;

[0013] Figure 2 A schematic diagram of the structure of a gain performance evaluation system for a compound product provided in an embodiment of the present application.

[0014] Explanation of the accompanying drawings: data acquisition module 11, fusion module 12, monitoring module 13, gain analysis module 14, screening analysis module 15, comprehensive analysis module 16. DETAILED DESCRIPTION

[0015] The present application solves the technical problem of inaccurate performance evaluation of chemical products in the prior art by providing a method and system for evaluating the gain performance of chemical products.

[0016] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0017] It should be noted that the terms "including" and "having" are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or are inherent to these processes, methods, products or devices.

[0018] Embodiment 1, as Figure 1 As shown, the present application provides a method for evaluating the gain performance of a compound product, wherein the method comprises:

[0019] The target fertilization area is equally divided into grids, and soil monitoring data and plant status monitoring data sets are collected for the divided multiple fertilization grids to obtain multiple soil monitoring data and multiple plant status monitoring data, wherein each fertilization grid corresponds to a grid position.

[0020] The target fertilization area is divided into equal grids to ensure that the fertilization area is divided into several small areas of the same size and shape. Each small area is called a fertilization grid. In each fertilization grid, soil and plant status monitoring data are collected. Specifically, soil monitoring data may include indicators such as soil moisture, pH value, and nutrient content (such as nitrogen, phosphorus, and potassium), while plant status monitoring data may cover parameters such as plant growth height, leaf color, and number of leaves; these monitoring data constitute an independent data set for each grid, reflecting the soil and plant growth conditions at that location.

[0021] Based on the multiple soil monitoring data and the multiple plant status monitoring data, a neighbor grid similarity analysis is performed, and the multiple fertilization grids are fused by neighborhood similarity to obtain multiple fused fertilization areas, wherein the multiple fused fertilization areas include multiple fused area soil monitoring data and multiple fused plant status monitoring data.

[0022] Based on the soil monitoring data and plant status monitoring data collected from each fertilization grid, the neighbor grid similarity analysis is used to identify and aggregate grids with similar soil and plant status characteristics. By calculating the similarity between grids (for example, based on Euclidean distance or correlation coefficient, etc.), the grids in adjacent grids that are similar in soil conditions, plant growth conditions, etc. are found so that they can be effectively merged. Through similarity analysis, these fertilization grids with similar characteristics are fused into fused fertilization areas; each fused fertilization area contains multiple original grids, and includes the comprehensive soil monitoring data and plant status monitoring data of the area, reflecting the overall situation in the fused area.

[0023] Furthermore, based on the plurality of soil monitoring data and the plurality of plant status monitoring data, a neighboring grid similarity analysis is performed, and the plurality of fertilization grids are subjected to neighborhood similarity fusion to obtain a plurality of fused fertilization areas, including:

[0024] A first fertilization grid is randomly selected from the multiple fertilization grids, and a first-order neighbor grid is screened and constructed according to the grid position of the first fertilization grid to determine a plurality of first-order neighbor grids; similarity coefficients of the first fertilization grid and the multiple first-order neighbor grids are respectively calculated in combination with the multiple soil monitoring data and the multiple plant status monitoring data to obtain a plurality of neighbor similarity coefficients; a neighbor similarity analysis is performed based on the multiple neighbor similarity coefficients to determine a first fused fertilization area; the first fused fertilization area is eliminated from the multiple fertilization grids, and a neighbor grid similarity analysis is performed on the remaining multiple fertilization grids to obtain a plurality of fused fertilization areas.

[0025] Specifically, one of the multiple fertilization grids is randomly selected as the first fertilization grid, and the first-order neighbor grids around it are determined according to the grid position of the first fertilization grid. The first-order neighbor grids usually include grids directly adjacent to the first fertilization grid, or neighbor grids selected according to a preset distance threshold. For the first fertilization grid and its first-order neighbor grids, the similarity coefficient between the first fertilization grid and each first-order neighbor grid is calculated respectively in combination with its soil monitoring data (such as soil moisture, pH, nutrient content, etc.) and plant status monitoring data (such as plant height, leaf color, etc.); the similarity coefficient can be calculated by a variety of methods, such as Pearson correlation coefficient, cosine similarity, Euclidean distance (the inverse is used as a similarity index), etc.; the calculation result will generate multiple neighbor similarity coefficients, each coefficient represents the degree of similarity between the first fertilization grid and the corresponding first-order neighbor grid. Similarity analysis is performed based on the above similarity coefficients, and a similarity threshold is set. The first-order neighbor grid with a similarity higher than the threshold and the first fertilization grid are classified as the same fused fertilization area, thereby obtaining the first fused fertilization area. The first determined fused fertilization area is removed from the fertilization grid set, and the remaining fertilization grids are retained; the above steps are repeated for these remaining grids to perform similarity screening, similarity coefficient calculation and fusion until all grids are assigned to the corresponding fused fertilization area.

[0026] Further, including:

[0027] When the multiple neighbor similarity coefficients are greater than or equal to the preset similarity coefficient, multiple first-order similar neighbor grids are obtained; based on the multiple first-order similar neighbor grids and the corresponding grid positions, the multiple fertilization grids are screened for second-order similar neighbor grids to obtain multiple second-order similar neighbor grid sets; after N-order similar neighbor grid screening and analysis, multiple N-order similar neighbor grid sets are obtained; the multiple first-order similar neighbor grids, the multiple second-order similar neighbor grid sets, and the multiple N-order similar neighbor grid sets are fused to obtain a first fused fertilization area.

[0028] When the neighbor similarity coefficient is greater than or equal to the preset similarity coefficient, these qualified neighbor grids are used as first-order similar neighbor grids, where the preset similarity coefficient is a threshold used to determine which grids are similar enough to the starting grid and can be classified into the same category. Based on the position of the first-order similar neighbor grids, the surrounding neighbor grids of these grids are screened for second-order similar neighbors, that is, the similarity coefficient calculation is continued for the neighbor grids of the first-order similar neighbor grids. When the similarity coefficient is greater than or equal to the preset similarity coefficient, these grids are identified as second-order similar neighbor grids and classified into the corresponding second-order similar neighbor grid set. Similarly, based on the similar neighbor grid set screened in the previous order, the next-order similar neighbor grid is further screened to form a second-order, third-order, and even N-order similar neighbor grid set. Each order of screening expands and screens the grids based on the similarity threshold, thereby constructing a larger and more uniform similar area. A plurality of first-order similar neighbor grids, a plurality of second-order similar neighbor grid sets, and a plurality of N-order similar neighbor grid sets are fused to generate a complete first fused fertilization area.

[0029] Furthermore, the second-order similar neighbor grid is a fertilization grid that is adjacent to the first-order similar neighbor grid and whose neighbor similarity coefficient is greater than or equal to a preset similarity coefficient.

[0030] The second-order similar neighbor grid is the grid that is directly adjacent to the first-order similar neighbor grid. For all grids adjacent to the first-order similar neighbor grid, the neighbor similarity coefficient between them and the initial fertilization grid is calculated. If the similarity coefficient of these grids is greater than or equal to the preset similarity coefficient threshold, they are classified into the second-order similar neighbor grid set.

[0031] At the fertilization node, the target chemical product is fertilized to the multiple fused fertilization areas according to the preset fertilization plan, and soil monitoring and plant status monitoring are performed on the multiple fused fertilization areas after fertilization in preset interval windows to determine multiple window soil monitoring data and multiple window plant status monitoring data.

[0032] A detailed pre-fertilization plan should be formulated based on the soil characteristics, plant needs and the properties of the target chemical products in the integrated fertilization area. The pre-fertilization plan should include the amount of fertilizer, time of fertilization, method of fertilization (such as broadcasting, hole fertilization, drip irrigation, etc.) and type and ratio of the target chemical products in each integrated fertilization area.

[0033] At the fertilization node, the target compound product is evenly distributed to each fertilization area according to the pre-set fertilization plan. After fertilization, each fertilization area is continuously monitored within the preset time interval window. Specifically, in each preset interval window, soil monitoring is performed on multiple fertilization areas after fertilization, and the monitoring content includes key indicators such as soil nutrient content, soil pH value, and soil moisture; at the same time, the status of plants in each fertilization area is monitored, and the monitoring content includes plant growth, leaf color, pests and diseases, etc.; by summarizing the monitoring data, multiple windows of soil monitoring data and multiple windows of plant status monitoring data are obtained.

[0034] Mapping gain analysis is performed based on the multiple fused regional soil monitoring data, the multiple fused plant status monitoring data, the multiple window soil monitoring data and the multiple window plant status monitoring data to determine multiple regional soil gain coefficients and multiple regional plant status gain coefficients.

[0035] Compare the soil monitoring data and plant status monitoring data of multiple fused regions with the soil monitoring data and plant status monitoring data of multiple windows after fertilization, and calculate the gain coefficient of each fused fertilization region, including the regional soil gain coefficient and the regional plant status gain coefficient. The regional soil gain coefficient reflects the improvement of the soil after fertilization, such as the increase of soil nutrient content and pH value adjustment; the regional plant status gain coefficient reflects the improvement of plant growth after fertilization, such as the increase of plant height and the number of leaves; specifically, the percentage change, logarithmic ratio change, etc. can be calculated by comparing the data changes before and after fertilization.

[0036] A steady-state evaluation is performed on the multiple regional soil gain coefficients and the multiple regional plant state gain coefficients respectively, and a soil screening step length and a plant screening step length are determined according to the evaluation results; based on the soil screening step length and the plant screening step length, and the screening starting points corresponding to the multiple regional soil gain coefficients and the multiple regional plant state gain coefficients respectively, a centralized screening analysis is performed on the multiple regional soil gain coefficients and the multiple regional plant state gain coefficients to determine a centralized soil gain coefficient and a centralized plant state gain coefficient.

[0037] The soil gain coefficient and plant state gain coefficient of each fused fertilization area are brought together for preliminary screening. The preliminary screening can be based on set standards, such as removing noise data or outliers, to ensure the accuracy of the data. Further centralized analysis of the data that have passed the preliminary screening and selection of representative gain coefficients can be based on statistical analysis methods (such as mean, median, weighted average, etc.) to extract the central trend of gain data in different regions to ensure that the screened gain coefficients can accurately reflect the overall fertilization effect. By conducting centralized screening and analysis of the soil gain coefficients of each region, a representative centralized soil gain coefficient is determined to reflect the overall soil improvement effect. Similarly, the plant state gain coefficients of each region are screened to determine a centralized plant state gain coefficient that represents the overall effect of plant growth improvement.

[0038] Further, steady-state evaluation is performed on the multiple regional soil gain coefficients and the multiple regional plant state gain coefficients respectively, and the soil screening step length and the plant screening step length are determined according to the evaluation results. Based on the soil screening step length and the plant screening step length, and the screening starting points corresponding to the multiple regional soil gain coefficients and the multiple regional plant state gain coefficients respectively, the multiple regional soil gain coefficients and the multiple regional plant state gain coefficients are subjected to centralized screening analysis to determine the centralized soil gain coefficient and the centralized plant state gain coefficient, including:

[0039] Perform steady-state evaluation on the multiple regional soil gain coefficients and the multiple regional plant state gain coefficients respectively to obtain soil gain steady-state factors and plant state gain steady-state factors; determine the soil screening step and the plant state screening step according to the deviation of the soil gain steady-state factors and the plant state gain steady-state factors from the preset gain steady-state factors respectively; calculate the means of the multiple regional soil gain coefficients and the multiple regional plant state gain coefficients respectively to obtain the mean of the regional soil gain coefficients and the mean of the regional plant state gain coefficients; use the mean of the regional soil gain coefficients and the mean of the regional plant state gain coefficients as screening starting points, conduct centralized screening and analysis on the multiple regional soil gain coefficients and the multiple regional plant state gain coefficients according to the soil screening step and the plant state screening step, and determine the centralized soil gain coefficient and the centralized plant state gain coefficient.

[0040] Specifically, steady-state analysis was performed on the soil gain coefficient and plant state gain coefficient of multiple regions to evaluate their stability of change after fertilization. During the steady-state evaluation process, the variance of the soil gain coefficient and the plant state gain coefficient was calculated to obtain the soil gain steady-state factor and the plant state gain steady-state factor. The soil gain steady-state factor is used to measure the stability of soil gain, that is, whether the soil state has achieved stable improvement after fertilization. The plant state gain steady-state factor is used to evaluate the stability of plant growth state and reflect the continuous improvement effect of fertilization on plants. The obtained soil gain steady-state factor and plant state gain steady-state factor are compared with the preset gain steady-state factor, and the deviation between them is calculated. The deviation reflects the degree of difference between the actual gain steady-state factor and the preset value; according to the size of the deviation, the soil screening step and the plant state screening step are determined. The mean of the soil gain coefficient and the plant state gain coefficient of multiple regions are calculated to obtain the regional soil gain coefficient mean and the regional plant state gain coefficient mean. These two means are used as the initial starting point of the centralized screening analysis, representing the overall gain performance of the entire region. Taking the mean value of regional soil gain coefficient and the mean value of regional plant state gain coefficient as the starting point for screening, the soil gain coefficient and plant state gain coefficient of multiple regions are screened step by step according to the soil screening step and plant state screening step. The screening process gives priority to retaining data with higher steady-state factors and smaller deviations, so that the final screened gain coefficient is more representative and stable. After screening, the centralized soil gain coefficient and centralized plant state gain coefficient are finally obtained.

[0041] Further, including:

[0042] Iterate the mean of the regional soil gain coefficient according to the soil screening step to determine the stage regional soil gain coefficient; judge whether the concentrated screening density of the stage regional soil gain coefficient is greater than or equal to the concentrated screening density of the mean of the regional soil gain coefficient; if so, update the stage regional soil gain coefficient to the starting point and continue iterative screening; if not, take the mean of the regional soil gain coefficient as the concentrated soil gain coefficient.

[0043] Preferably, the mean value of the regional soil gain coefficient is used as the initial screening starting point, and the gain coefficient is iteratively analyzed according to the determined soil screening step, and the screening results are gradually adjusted and converged; in each iteration, a new stage regional soil gain coefficient is calculated according to the current screening starting point and the soil screening step, and the stage regional soil gain coefficient reflects the soil gain situation in the region under the current screening step; the concentrated screening density of the stage regional soil gain coefficient is calculated and compared with the concentrated screening density of the initial regional soil gain coefficient mean, and the concentrated screening density can be defined as the number or proportion of soil gain coefficients within a certain range, which is used to measure the concentration of the gain coefficient; if the stage regional soil gain If the concentrated screening density of the coefficient is greater than or equal to the concentrated screening density of the mean regional soil gain coefficient, it means that the gain coefficient obtained in the current iteration is more concentrated or stable. Therefore, the stage regional soil gain coefficient is updated to a new starting point and the iterative screening continues. If the concentrated screening density of the stage regional soil gain coefficient is less than the concentrated screening density of the mean regional soil gain coefficient, it means that the iteration process has not increased the concentration of the gain coefficient. Therefore, the iteration is stopped and the mean regional soil gain coefficient is used as the final concentrated soil gain coefficient. When the conditions for stopping iteration are met (i.e., the concentrated screening density has not increased), the iteration process is terminated and the mean regional soil gain coefficient at this time is the concentrated soil gain coefficient.

[0044] The performance evaluation network layer is used to comprehensively analyze the concentrated soil gain coefficient and the concentrated plant state gain coefficient to determine the target performance gain coefficient of the target compound product, and the target performance gain coefficient is used as the gain performance evaluation result.

[0045] Construct a performance evaluation network layer for comprehensive analysis of soil and plant state gain data. The centralized soil gain coefficient and centralized plant state gain coefficient are used as input data and input into the performance evaluation network layer; the performance evaluation network layer identifies the mutual influence relationship between soil and plant state gain by analyzing and integrating these data, and calculates the overall gain performance effect; through the comprehensive analysis of the performance evaluation network layer, a target performance gain coefficient is generated. The target performance gain coefficient represents the overall gain effect of the chemical product in the current fertilization area and conditions, and reflects the comprehensive improvement of soil and plant state. The target performance gain coefficient is used as the gain performance evaluation result to provide data support for the comprehensive effect of chemical products in fertilization applications.

[0046] Furthermore, the performance evaluation network layer is used to comprehensively analyze the centralized soil gain coefficient and the centralized plant state gain coefficient to determine the target performance gain coefficient of the target compound product, including:

[0047] The soil gain coefficients in multiple sample sets, the plant state gain coefficients in multiple sample sets and the target performance gain coefficients in multiple sample sets are obtained as training sets; the framework constructed based on the feedforward neural network is supervised and trained using the training sets to learn the two-to-one mapping relationship between the concentrated soil gain coefficients and the concentrated plant state gain coefficients and the target performance gain coefficients until convergence, thereby obtaining the performance evaluation network layer that has been trained.

[0048] Specifically, multiple sample data are obtained from historical data, including the concentrated soil gain coefficient, concentrated plant state gain coefficient and corresponding target performance gain coefficient of the sample, and these data sets are used as training sets. Each sample in the training set contains the gain performance of soil and plant state and the actual gain effect of the compound product; a performance evaluation network layer is designed based on a feedforward neural network to establish a two-to-one mapping relationship between the concentrated soil gain coefficient and the concentrated plant state gain coefficient and the target performance gain coefficient. The feedforward neural network gradually extracts the characteristics of the input data through multiple hidden layers to achieve accurate mapping of the input gain data to the output performance gain coefficient; the feedforward neural network is supervised and trained using the prepared training set; by continuously inputting the soil gain coefficient and plant state gain coefficient in the training set as input, the network layer adjusts its weights and bias parameters to minimize the error between the output prediction value and the true target performance gain coefficient. The training process includes forward propagation (calculating the output prediction value based on the current parameters), error calculation (calculating the error between the predicted output and the true target performance gain coefficient of the sample), and back propagation (reversely adjusting the weights based on the error to optimize the network parameters). The training process is continuously iterated until the error between the predicted value of the network output and the actual target performance gain coefficient of the sample converges to the preset threshold, indicating that the model has fully learned the mapping relationship between input and output. At this point, the trained performance evaluation network layer can efficiently predict the target performance gain coefficient. The trained performance evaluation network layer can be applied to real-time gain performance evaluation. By inputting the new concentrated soil gain coefficient and concentrated plant state gain coefficient, the network layer can quickly calculate the target performance gain coefficient of the target compound product, providing intelligent support for fertilization effect evaluation.

[0049] In summary, the embodiments of the present application have at least the following technical effects:

[0050] First, the target fertilization area is equally divided into grids, and soil monitoring data and plant status monitoring data sets are collected for the divided multiple fertilization grids to obtain multiple soil monitoring data and multiple plant status monitoring data, wherein each fertilization grid corresponds to a grid position. Next, neighbor grid similarity analysis is performed based on multiple soil monitoring data and multiple plant status monitoring data, and neighborhood similarity fusion is performed on multiple fertilization grids to obtain multiple fused fertilization areas, wherein the multiple fused fertilization areas include multiple fused area soil monitoring data and multiple fused plant status monitoring data. At the fertilization node, the target compound product is fertilized to multiple fused fertilization areas according to the pre-set fertilization scheme, and soil monitoring and plant status monitoring are performed on multiple fused fertilization areas after fertilization in a preset interval window to determine multiple window soil monitoring data and multiple window plant status monitoring data. Next, mapping gain analysis is performed based on multiple fused area soil monitoring data, multiple fused plant status monitoring data, multiple window soil monitoring data, and multiple window plant status monitoring data to determine multiple area soil gain coefficients and multiple area plant status gain coefficients. Then, a steady-state evaluation is performed on the soil gain coefficients of multiple regions and the plant state gain coefficients of multiple regions respectively, and the soil screening step length and the plant screening step length are determined according to the evaluation results. Based on the soil screening step length and the plant screening step length, as well as the screening starting points corresponding to the soil gain coefficients of multiple regions and the plant state gain coefficients of multiple regions, the soil gain coefficients of multiple regions and the plant state gain coefficients of multiple regions are centrally screened and analyzed to determine the central soil gain coefficient and the central plant state gain coefficient. Finally, the performance evaluation network layer is used to comprehensively analyze the central soil gain coefficient and the central plant state gain coefficient to determine the target performance gain coefficient of the target compound product, and the target performance gain coefficient is used as the gain performance evaluation result. The technical problem of inaccurate performance evaluation of compound products in the prior art is solved, and the technical effect of improving the accuracy of performance evaluation of compound products is achieved.

[0051] Example 2, based on the same inventive concept as the gain performance evaluation method of the compound product in the previous example, Figure 2 As shown, the present application provides a gain performance evaluation system for a compound product, wherein the system comprises:

[0052] A data acquisition module 11, wherein the data acquisition module 11 is used to equally divide the target fertilization area into grids, and collect soil monitoring data and plant status monitoring data sets for the divided multiple fertilization grids to obtain multiple soil monitoring data and multiple plant status monitoring data, wherein each fertilization grid corresponds to a grid position; a fusion module 12, wherein the fusion module 12 is used to perform neighbor grid similarity analysis based on the multiple soil monitoring data and the multiple plant status monitoring data, and perform neighborhood similarity fusion on the multiple fertilization grids to obtain multiple fused fertilization areas, wherein the multiple fused fertilization areas include multiple fused area soil monitoring data and multiple fused plant status monitoring data; a monitoring module 13, wherein the monitoring module 13 is used to fertilize the multiple fused fertilization areas with the target compound product according to the pre-set fertilization plan at the fertilization node, and perform soil monitoring and plant status monitoring on the multiple fused fertilization areas after fertilization in preset interval windows to determine multiple window soil monitoring data and multiple window plant status monitoring data; a gain analysis module 14, wherein the gain analysis module 14 is used to obtain multiple window soil monitoring data and multiple window plant status monitoring data based on the multiple fusion areas. The regional soil monitoring data, multiple fused plant status monitoring data, multiple window soil monitoring data and multiple window plant status monitoring data are mapped and analyzed for gain, and multiple regional soil gain coefficients and multiple regional plant status gain coefficients are determined; a screening and analysis module 15, wherein the screening and analysis module 15 is used to respectively perform steady-state evaluation on the multiple regional soil gain coefficients and the multiple regional plant status gain coefficients, determine the soil screening step length and the plant screening step length according to the evaluation results, and based on the soil screening step length and the plant screening step length, as well as the screening starting points corresponding to the multiple regional soil gain coefficients and the multiple regional plant status gain coefficients, perform centralized screening and analysis on the multiple regional soil gain coefficients and the multiple regional plant status gain coefficients to determine the centralized soil gain coefficient and the centralized plant status gain coefficient; a comprehensive analysis module 16, wherein the comprehensive analysis module 16 is used to perform a comprehensive analysis on the centralized soil gain coefficient and the centralized plant status gain coefficient using a performance evaluation network layer, determine the target performance gain coefficient of the target compound product, and use the target performance gain coefficient as the gain performance evaluation result.

[0053] Furthermore, the fusion module 12 is used to perform the following method:

[0054] A first fertilization grid is randomly selected from the multiple fertilization grids, and a first-order neighbor grid is screened and constructed according to the grid position of the first fertilization grid to determine a plurality of first-order neighbor grids; similarity coefficients of the first fertilization grid and the multiple first-order neighbor grids are respectively calculated in combination with the multiple soil monitoring data and the multiple plant status monitoring data to obtain a plurality of neighbor similarity coefficients; a neighbor similarity analysis is performed based on the multiple neighbor similarity coefficients to determine a first fused fertilization area; the first fused fertilization area is eliminated from the multiple fertilization grids, and a neighbor grid similarity analysis is performed on the remaining multiple fertilization grids to obtain a plurality of fused fertilization areas.

[0055] Furthermore, the fusion module 12 is used to perform the following method:

[0056] When the multiple neighbor similarity coefficients are greater than or equal to the preset similarity coefficient, multiple first-order similar neighbor grids are obtained; based on the multiple first-order similar neighbor grids and the corresponding grid positions, the multiple fertilization grids are screened for second-order similar neighbor grids to obtain multiple second-order similar neighbor grid sets; after N-order similar neighbor grid screening and analysis, multiple N-order similar neighbor grid sets are obtained; the multiple first-order similar neighbor grids, the multiple second-order similar neighbor grid sets, and the multiple N-order similar neighbor grid sets are fused to obtain a first fused fertilization area.

[0057] Furthermore, the screening and analysis module 15 is used to perform the following method:

[0058] Perform steady-state evaluation on the multiple regional soil gain coefficients and the multiple regional plant state gain coefficients respectively to obtain soil gain steady-state factors and plant state gain steady-state factors; determine the soil screening step and the plant state screening step according to the deviation of the soil gain steady-state factors and the plant state gain steady-state factors from the preset gain steady-state factors respectively; calculate the means of the multiple regional soil gain coefficients and the multiple regional plant state gain coefficients respectively to obtain the mean of the regional soil gain coefficients and the mean of the regional plant state gain coefficients; use the mean of the regional soil gain coefficients and the mean of the regional plant state gain coefficients as screening starting points, conduct centralized screening and analysis on the multiple regional soil gain coefficients and the multiple regional plant state gain coefficients according to the soil screening step and the plant state screening step, and determine the centralized soil gain coefficient and the centralized plant state gain coefficient.

[0059] Furthermore, the screening and analysis module 15 is used to perform the following method:

[0060] Iterate the mean of the regional soil gain coefficient according to the soil screening step to determine the stage regional soil gain coefficient; judge whether the concentrated screening density of the stage regional soil gain coefficient is greater than or equal to the concentrated screening density of the mean of the regional soil gain coefficient; if so, update the stage regional soil gain coefficient to the starting point and continue iterative screening; if not, take the mean of the regional soil gain coefficient as the concentrated soil gain coefficient.

[0061] Furthermore, the comprehensive analysis module 16 is used to perform the following method:

[0062] The soil gain coefficients in multiple sample sets, the plant state gain coefficients in multiple sample sets and the target performance gain coefficients in multiple sample sets are obtained as training sets; the framework constructed based on the feedforward neural network is supervised and trained using the training sets to learn the two-to-one mapping relationship between the concentrated soil gain coefficients and the concentrated plant state gain coefficients and the target performance gain coefficients until convergence, thereby obtaining the performance evaluation network layer that has been trained.

[0063] Furthermore, the comprehensive analysis module 16 is used to perform the following method:

[0064] The second-order similar neighbor grid is a fertilization grid that is adjacent to the first-order similar neighbor grid and whose neighbor similarity coefficient is greater than or equal to a preset similarity coefficient.

[0065] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0066] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0067] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A method for evaluating the gain performance of a compound product, characterized in that: The method comprises: The target fertilization area is equally divided into grids, and soil monitoring data and plant status monitoring data sets are collected for the divided multiple fertilization grids to obtain multiple soil monitoring data and multiple plant status monitoring data, wherein each fertilization grid corresponds to a grid position; Based on the multiple soil monitoring data and the multiple plant status monitoring data, a neighboring grid similarity analysis is performed, and the multiple fertilization grids are subjected to neighborhood similarity fusion to obtain multiple fused fertilization areas, wherein the multiple fused fertilization areas include multiple fused area soil monitoring data and multiple fused plant status monitoring data; At the fertilization node, the target compound product is fertilized to the multiple fusion fertilization areas according to the pre-set fertilization plan, and soil monitoring and plant status monitoring are performed on the multiple fusion fertilization areas after fertilization in preset interval windows to determine multiple window soil monitoring data and multiple window plant status monitoring data; Perform mapping gain analysis based on the multiple fused regional soil monitoring data, the multiple fused plant status monitoring data, the multiple window soil monitoring data, and the multiple window plant status monitoring data to determine multiple regional soil gain coefficients and multiple regional plant status gain coefficients; Performing steady-state evaluations on the multiple regional soil gain coefficients and the multiple regional plant state gain coefficients respectively, determining a soil screening step length and a plant screening step length according to the evaluation results, and performing centralized screening analysis on the multiple regional soil gain coefficients and the multiple regional plant state gain coefficients based on the soil screening step length and the plant screening step length, and the screening starting points respectively corresponding to the multiple regional soil gain coefficients and the multiple regional plant state gain coefficients, to determine a centralized soil gain coefficient and a centralized plant state gain coefficient; The performance evaluation network layer is used to comprehensively analyze the concentrated soil gain coefficient and the concentrated plant state gain coefficient to determine the target performance gain coefficient of the target compound product, and the target performance gain coefficient is used as the gain performance evaluation result.

2. The method for evaluating the gain performance of a compound product according to claim 1, wherein: Based on the plurality of soil monitoring data and the plurality of plant status monitoring data, a neighboring grid similarity analysis is performed, and the plurality of fertilization grids are subjected to neighborhood similarity fusion to obtain a plurality of fused fertilization areas, including: Randomly selecting a first fertilization grid from the plurality of fertilization grids, performing first-order neighbor grid screening and construction according to a grid position of the first fertilization grid, and determining a plurality of first-order neighbor grids; Combining the plurality of soil monitoring data and the plurality of plant status monitoring data, respectively calculating similarity coefficients of the first fertilization grid and the plurality of first-order neighbor grids to obtain a plurality of neighbor similarity coefficients; Performing neighbor similarity analysis based on the multiple neighbor similarity coefficients to determine a first fusion fertilization area; The first fused fertilization area is removed from the multiple fertilization grids, and neighbor grid similarity analysis is performed on the remaining multiple fertilization grids to obtain multiple fused fertilization areas.

3. The method for evaluating the gain performance of a compound product as claimed in claim 2, characterized in that: include: When the plurality of neighbor similarity coefficients are greater than or equal to a preset similarity coefficient, a plurality of first-order similar neighbor grids are obtained; Based on the multiple first-order similar neighbor grids and the corresponding grid positions, the multiple fertilization grids are screened for second-order similar neighbor grids to obtain multiple second-order similar neighbor grid sets; After N-order similar neighbor grid screening and analysis, multiple N-order similar neighbor grid sets are obtained; The multiple first-order similar neighbor grids, the multiple second-order similar neighbor grid sets, and the multiple N-order similar neighbor grid sets are fused to obtain a first fused fertilization area.

4. The method for evaluating the gain performance of a compound product according to claim 1, wherein: The plurality of regional soil gain coefficients and the plurality of regional plant state gain coefficients are respectively subjected to steady-state evaluation, and a soil screening step length and a plant screening step length are determined according to the evaluation results; based on the soil screening step length and the plant screening step length, and the screening starting points respectively corresponding to the plurality of regional soil gain coefficients and the plurality of regional plant state gain coefficients, the plurality of regional soil gain coefficients and the plurality of regional plant state gain coefficients are subjected to centralized screening analysis to determine a centralized soil gain coefficient and a centralized plant state gain coefficient, including: Performing steady-state evaluation on the soil gain coefficients of the plurality of regions and the plant state gain coefficients of the plurality of regions respectively to obtain a soil gain steady-state factor and a plant state gain steady-state factor; Determining a soil screening step length and a plant state screening step length according to the deviations of the soil gain steady-state factor and the plant state gain steady-state factor from a preset gain steady-state factor respectively; Calculating the average values ​​of the plurality of regional soil gain coefficients and the plurality of regional plant state gain coefficients respectively to obtain the average value of the regional soil gain coefficient and the average value of the regional plant state gain coefficient; The mean value of the regional soil gain coefficient and the mean value of the regional plant state gain coefficient are respectively used as the screening starting points, and the multiple regional soil gain coefficients and the multiple regional plant state gain coefficients are centrally screened and analyzed according to the soil screening step and the plant state screening step to determine the central soil gain coefficient and the central plant state gain coefficient.

5. The method for evaluating the gain performance of a compound product as claimed in claim 4, characterized in that: include: Iterate the mean value of regional soil gain coefficient according to the soil screening step length to determine the regional soil gain coefficient of the stage; Determine whether the concentrated screening density of the regional soil gain coefficient in the stage is greater than or equal to the concentrated screening density of the mean value of the regional soil gain coefficient. If so, update the regional soil gain coefficient in the stage to the starting point and continue iterative screening; If not, the mean value of the regional soil gain coefficient is taken as the concentrated soil gain coefficient.

6. The method for evaluating the gain performance of a compound product according to claim 1, wherein: The performance evaluation network layer is used to comprehensively analyze the concentrated soil gain coefficient and the concentrated plant state gain coefficient to determine the target performance gain coefficient of the target compound product, including: Acquire soil gain coefficients in multiple sample sets, plant state gain coefficients in multiple sample sets, and target performance gain coefficients in multiple samples as training sets; The framework constructed based on the feedforward neural network is supervised and trained using the training set to learn the two-to-one mapping relationship between the concentrated soil gain coefficient and the concentrated plant state gain coefficient and the target performance gain coefficient until convergence, thereby obtaining the trained performance evaluation network layer.

7. The method for evaluating the gain performance of a compound product as claimed in claim 3, characterized in that: The second-order similar neighbor grid is a fertilization grid that is adjacent to the first-order similar neighbor grid and whose neighbor similarity coefficient is greater than or equal to a preset similarity coefficient.

8. A system for evaluating the gain performance of a compound product, characterized in that: A method for evaluating the gain performance of a compound product according to any one of claims 1 to 7, the system comprising: A data acquisition module, wherein the data acquisition module is used to equally divide the target fertilization area into grids, and collect a set of soil monitoring data and plant status monitoring data for the divided multiple fertilization grids to obtain multiple soil monitoring data and multiple plant status monitoring data, wherein each fertilization grid corresponds to a grid position; A fusion module, the fusion module is used to perform a neighboring grid similarity analysis based on the multiple soil monitoring data and the multiple plant status monitoring data, perform neighborhood similarity fusion on the multiple fertilization grids, and obtain multiple fused fertilization areas, wherein the multiple fused fertilization areas include multiple fused area soil monitoring data and multiple fused plant status monitoring data; A monitoring module, wherein the monitoring module is used to fertilize the target compound product to the multiple fused fertilization areas according to a pre-set fertilization plan at a fertilization node, and to perform soil monitoring and plant status monitoring on the multiple fused fertilization areas after fertilization in a preset interval window, and determine soil monitoring data in multiple windows and plant status monitoring data in multiple windows; A gain analysis module, the gain analysis module is used to perform mapping gain analysis based on the multiple fused regional soil monitoring data, the multiple fused plant status monitoring data, the multiple window soil monitoring data and the multiple window plant status monitoring data to determine multiple regional soil gain coefficients and multiple regional plant status gain coefficients; A screening and analysis module, wherein the screening and analysis module is used to respectively perform steady-state evaluations on the multiple regional soil gain coefficients and the multiple regional plant state gain coefficients, determine a soil screening step length and a plant screening step length according to the evaluation results, and based on the soil screening step length and the plant screening step length, as well as the screening starting points respectively corresponding to the multiple regional soil gain coefficients and the multiple regional plant state gain coefficients, perform centralized screening and analysis on the multiple regional soil gain coefficients and the multiple regional plant state gain coefficients to determine a centralized soil gain coefficient and a centralized plant state gain coefficient; A comprehensive analysis module is used to use the performance evaluation network layer to comprehensively analyze the concentrated soil gain coefficient and the concentrated plant state gain coefficient, determine the target performance gain coefficient of the target chemical product, and use the target performance gain coefficient as the gain performance evaluation result.

Images