CN-121998790-A - Crop planting method and system based on big data

CN121998790ACN 121998790 ACN121998790 ACN 121998790ACN-121998790-A

Abstract

The invention discloses a crop planting method and system based on big data, and relates to the technical field of agricultural planting. The method comprises the steps of collecting soil and historical crop data of an area to be planted, constructing an anti-disturbance characteristic field through clustering and adding Laplace noise to simulate environmental stress and protect privacy, calculating reverse neighbor count and environmental weight based on local neighborhood, further calculating weighted normalized centroid offset value, identifying an ecological instability grid by using a double statistical threshold, constructing a comprehensive evaluation function optimal dominant planting core, expanding a stable clustering area by adopting a sequential area growth algorithm limited by ecological niches, forcedly merging broken plaques into a dominant cluster, and finally implementing a differential planting method for a core planting area and an ecological buffer zone. The method effectively identifies microscopic ecological heterogeneity of farmlands, eliminates the fragmentation of operation plots, and remarkably improves crop yield and ecological stability.

Inventors

HAN YU
YUE RUNQING
WU XIAOHONG
YANG FEI
YAN WU
LIU LIQUN
WEN PENGXIANG

Assignees

利良华玫生物科技(成都)有限公司

Dates

Publication Date: 20260508
Application Date: 20260409

Claims (10)

1. The crop planting method based on big data is characterized by comprising the following steps: step S1, collecting soil data and historical crop data of a region to be planted; S2, performing cluster analysis on the soil data and the historical crop data to generate an anti-stress disturbance characteristic field, and calculating reverse neighbor counts, environment weights and local geometric centroids of the operation grids in the area to be planted based on the local space neighborhood; Step S3, calculating weighted normalized centroid offset values of all the operation grids by combining the local geometric centroids, the environmental weights and the reverse neighbor counts, setting a judgment threshold value based on the statistical distribution of the weighted normalized centroid offset values and the reverse neighbor counts, and identifying the ecological unstable operation grids; And S4, constructing a comprehensive evaluation function, selecting dominant planting kernels, expanding the dominant planting kernels into stable clustering areas by using an area growth algorithm, and respectively configuring differentiated planting methods for the stable clustering areas and the ecological unstable operation grids.
2. The big data based crop planting method as claimed in claim 1, wherein in the step S1, the area to be planted is divided into a plurality of operation grids, and soil data and historical crop data in each operation grid are collected; the soil data comprise soil organic matters, pH value and conductivity; The historical crop data comprises plant height and leaf area indexes of the historical crops; Step S1 is also used for collecting the space coordinates of the operation grid; And carrying out normalization processing on the soil data, the historical crop data and the space coordinates.
3. The method for growing crops based on big data according to claim 2, wherein in the step S2, the process of generating the stress-resistant disturbance characteristic field specifically includes: Splicing soil data and historical crop data in each operation grid in a region to be planted into feature vectors, and generating a plurality of soil feature center vectors from the feature vectors by using a clustering algorithm; And adding Laplacian noise on the soil characteristic center vector to generate an anti-stress disturbance characteristic field.
4. A method of growing crops based on big data according to claim 3, characterized in that the process of calculating the reverse neighbor count of the job grid comprises in particular: Selecting any operation grid in a region to be planted as a target grid, marking other operation grids in a local space adjacent region of the target grid as a first neighborhood grid set, and marking other operation grids in the local space adjacent region of the first neighborhood grid set as a second candidate grid set; Calculating Euclidean distance between the feature vector of the first neighborhood grid set and the feature vector of the second candidate grid set, sorting the second candidate grid sets in ascending order based on the Euclidean distance, and selecting k second candidate grid sets with the minimum Euclidean distance corresponding to each first neighborhood grid set to form neighbor sets corresponding to each first neighborhood grid set; counting the total frequency of the target grid in all neighbor sets, wherein the total frequency is the reverse neighbor count of the target grid; the environmental weight of the target grid is inversely related to the reverse neighbor count of the target grid; The method for calculating the local geometric centroid of the target grid comprises the step of calculating the mean value vector of the feature vector and the feature vector of the first neighborhood grid set in the target grid to serve as the local geometric centroid of the target grid.
5. The big data based crop planting method of claim 4, wherein the process of calculating the weighted normalized centroid offset value specifically comprises: calculating Euclidean distance between the feature vector of the target grid and the local geometric centroid to obtain the original centroid offset of the target grid; Taking the mean value of Euclidean distance between the feature vector of the target grid and the feature vector of the first neighborhood grid set as a normalization factor, and carrying out normalization processing on the original centroid offset by using the normalization factor to obtain a normalized centroid offset value; and correcting the normalized centroid offset value by using reverse neighbor counting to obtain a weighted normalized centroid offset value of the target grid.
6. The big data based crop planting method of claim 5, wherein identifying the ecologically unstable operation grid process specifically comprises: Calculating the average value and standard deviation of weighted normalized centroid offset values of all operation grids in the area to be planted, and constructing an offset judgment threshold value by using the average value and standard deviation; setting an isolation degree judgment threshold value based on the statistical distribution of the reverse neighbor counts; traversing each operation grid in the area to be planted, and judging that the operation grid is an ecological instability operation grid if the weighted normalized centroid offset value of the operation grid is larger than the offset judgment threshold or the reverse neighbor count is smaller than the isolation judgment threshold.
7. The big data based crop planting method of claim 6, wherein the process of constructing the comprehensive evaluation function and selecting dominant planting kernels specifically comprises: removing the operation grids marked as the ecological instability operation grids from the operation grids of the area to be planted, and forming a candidate set by the rest operation grids; Calculating Euclidean distance between the feature vector of each operation grid in the candidate set and the soil feature center vector in the stress-resistant disturbance feature field, and taking the minimum value of the Euclidean distance as a feature distance component; constructing a comprehensive evaluation function based on the characteristic distance component, wherein the mathematical expression of the comprehensive evaluation function is as follows: ; Wherein, the To work grid Is used for the comprehensive scoring of the (c), To work grid Is used for the feature vector of (a), Is the soil characteristic center vector in the stress disturbance resistant characteristic field, Is the index of the soil characteristic center vector, And The spatial coordinates of the grid are respectively, To work grid A set of the grid of jobs in the local spatial neighborhood, And The balance weight coefficient is the characteristic and space; And sequencing the operation grids in the candidate set according to the comprehensive scores, and selecting a plurality of operation grids with the minimum comprehensive scores as dominant planting cores.
8. The big data based crop planting method of claim 7, wherein the area growing algorithm specifically comprises: iteratively growing adjacent operation grids by taking the dominant planting cores as seed areas for area growth; Calculating a growth resistance difference value between an adjacent operation grid and a current seed area, and if the growth resistance difference value is smaller than a preset growth tolerance threshold value and the adjacent operation grid is not marked as an ecological unstable operation grid, merging the adjacent operation grid with the current seed area, and continuing to perform area growth by taking the adjacent operation grid as the seed area until growth converges; Counting the area of each communication area formed after growth convergence, and judging the communication area as a broken plaque if the communication area is smaller than a preset minimum operation granularity threshold; If the communicated area is larger than or equal to a preset minimum operation granularity threshold value, judging that the communicated area is a dominant cluster; calculating Euclidean distance between feature vectors of dominant planting cores corresponding to the broken patches and feature vectors of dominant planting cores corresponding to all dominant clusters, merging the broken patches into the dominant cluster with the minimum Euclidean distance, updating the area covered by the dominant cluster, and forming a stable clustering area by all the areas covered by the dominant clusters; The growth resistance difference value is the absolute value of the difference between the weighted normalized centroid offset value of the adjacent operation grid and the weighted normalized centroid offset value of the seed region.
9. The big data based crop planting method of claim 8, wherein configuring the differentiated planting method specifically comprises: Judging the stable clustering area as a core planting area, and configuring a high-density planting and high-fertility input method; and judging the ecological unstable operation grid as a ecological buffer zone, and configuring a low-density planting and protective farming method.
10. A big data based crop planting system applied to the big data based crop planting method according to any one of claims 1-9, comprising an acquisition module, a classification module, an identification module and a planting module; the acquisition module is used for acquiring soil data and historical crop data of a region to be planted; The classification module is used for carrying out cluster analysis on the soil data and the historical crop data, generating an anti-stress disturbance characteristic field, and calculating reverse neighbor counts, environment weights and local geometric centroids of the operation grids in the area to be planted based on the local space neighborhood; The identification module is used for calculating weighted normalized centroid offset values of all the operation grids by combining the local geometric centroid, the environmental weight and the reverse neighbor count, setting a judgment threshold value based on the statistical distribution of the weighted normalized centroid offset values and the reverse neighbor count, and identifying the ecological unstable operation grids; the planting module is used for constructing a comprehensive evaluation function, selecting dominant planting kernels, expanding the dominant planting kernels into stable clustering areas by using an area growth algorithm, and respectively configuring different planting methods for the stable clustering areas and the ecological unstable operation grids.

Description

Crop planting method and system based on big data Technical Field The invention relates to the technical field of agricultural planting, in particular to a crop planting method and system based on big data. Background Agriculture is the basis of national economy, along with the deep penetration of new generation information technologies such as Internet of things, cloud computing, artificial intelligence and the like, the traditional agriculture is accelerated to digital and intelligent transformation, in modern crop planting management, the whole agricultural production process is monitored and analyzed by utilizing a big data technology, the multi-source heterogeneous data such as meteorological hydrology, soil physicochemical properties, crop growth vigor, historical yield and the like in a planting area are widely collected, and the coupling relation between crop growth and environmental elements is revealed by utilizing a data mining algorithm, so that the method can assist an agricultural practitioner to formulate scientific seeding, fertilization, irrigation and pest control strategies, and has important practical significance for improving the utilization efficiency of agricultural resources, guaranteeing the grain safety and improving the market competitiveness of agricultural products. In order to achieve the above objective, the existing agricultural data analysis technology has already been tried to guide the current planting decision by using a historical data model, and chinese patent No. 117033810B discloses an agricultural data analysis management system and method based on big data, which mainly analyzes the average yield difference of the same kind of crops in different areas by acquiring historical agricultural data, and matches the optimal planting condition of the current area by combining the specific effects of temperature change, pest and disease occurrence degree and planting mode on yield; however, the prior art focuses on comparing the transverse output and the environmental mean value of different sections, lacks the fine recognition of the spatial heterogeneity of the ecological environment of soil in a land block, often does not uniformly distribute the soil properties of the land block, but has complex spatial variation, such as significant difference in the ecological stability of soil mutation areas and ditch adjacent areas, when the environmental data are utilized, the prior art carries out association analysis based on original observation values, lacks the robustness verification of the data under simulated environmental stress, does not fully consider the privacy protection and the feature stress resistance construction in the multi-source data fusion process, and also lacks a mechanism capable of automatically carrying out the region optimizing and the dynamic growth according to the similarity of ecological features and the spatial compactness on the demarcation of planting partitions, so that a standardized operation unit which is difficult to form a connecting piece, regular and uniform in attribute is difficult to limit the implementation effect and the depth of the fine management of the automation operation of a large farm machine. Disclosure of Invention The technical problems solved by the invention are that the prior art lacks of fine recognition of space heterogeneity of soil ecological environment in a land block, lacks of robustness verification of data under simulated environmental stress, does not fully consider privacy protection and characteristic stress resistance construction in a multi-source data fusion process, and also lacks of a mechanism capable of automatically carrying out region optimization and dynamic growth according to ecological characteristic similarity and space compactness, is difficult to form standardized operation units which are connected and regular and have uniform properties, and limits implementation effect of large-scale agricultural machinery automation operation and depth of fine management. In order to solve the technical problems, the invention provides the following technical scheme that the crop planting method based on big data comprises the following steps: step S1, collecting soil data and historical crop data of a region to be planted; S2, performing cluster analysis on the soil data and the historical crop data to generate an anti-stress disturbance characteristic field, and calculating reverse neighbor counts, environment weights and local geometric centroids of the operation grids in the area to be planted based on the local space neighborhood; Step S3, calculating weighted normalized centroid offset values of all the operation grids by combining the local geometric centroids, the environmental weights and the reverse neighbor counts, setting a judgment threshold value based on the statistical distribution of the weighted normalized centroid offset values and the reverse neighbor counts, and identifying the ecolog