CN-121981323-A - Multi-machine intelligent agriculture combined operation optimization method, device and medium

CN121981323ACN 121981323 ACN121981323 ACN 121981323ACN-121981323-A

Abstract

A multi-machine intelligent agriculture combined operation optimization method, device and medium relate to the intelligent agriculture field. The method comprises the steps of controlling a first agricultural machine to execute first-stage operation on a target land block to obtain a first-stage result, controlling a second agricultural machine to execute second-stage operation based on the first-stage result to generate a process operation parameter set, controlling a third agricultural machine to harvest the target land block to acquire real yield data, dividing each geographical subarea into different parameter operation subareas, calculating a representative yield value for each parameter operation subarea, identifying and marking the disturbed geographical subarea, correcting the yield value of the disturbed geographical subarea into the representative yield value of the corresponding parameter operation subarea to obtain a corrected yield data set, and generating an optimized operation strategy based on the process operation parameter set and the corrected yield data set. By implementing the technical scheme provided by the application, the accuracy of intelligent agricultural decision making is improved.

Inventors

LING XIUZE
HAN LV
HU BO
LI MING
YANG BO

Assignees

上海星联智创智能科技股份有限公司

Dates

Publication Date: 20260505
Application Date: 20260113

Claims (10)

1. The multi-machine intelligent agriculture combined operation optimization method is characterized by comprising the following steps of: Controlling a first agricultural machine to execute first-stage operation on a target land block to obtain a first-stage result, and controlling a second agricultural machine to execute second-stage operation on a plurality of geographic subareas in the target land block based on the first-stage result to generate a process operation parameter set corresponding to each geographic subarea; Controlling a third agricultural machine to harvest the target land block, and collecting actual harvest yield data corresponding to each geographical subarea; Dividing each geographical subarea into different parameter operation subareas based on the controllable operation parameter values in the process operation parameter set, wherein all geographical subareas belonging to the same parameter operation subarea correspond to the same controllable operation parameter value; Calculating a representative yield value for each of the parametric job partitions based on the real-time yield data for all geographic sub-areas within the parametric job partition; comparing the actual yield data of each geographical subarea with corresponding representative yield values in a deviation manner, and identifying and marking a disturbed geographical subarea; Correcting the yield value of the disturbed geographical subarea into a representative yield value of a corresponding parameter operation subarea, and keeping the yield value of the geographical subarea which is not marked as the real yield data to obtain a corrected yield data set; An optimized job strategy is generated based on the process job parameter set and the corrected yield data set.
2. The method according to claim 1, wherein the controlling the second agricultural machine performs a second stage of operations on the plurality of geographical sub-areas within the target parcel based on the first stage results, generating a set of process operation parameters corresponding to each of the geographical sub-areas, specifically comprising: Acquiring soil property parameters of each geographical subarea in the first stage result, and determining the soil type of each geographical subarea according to the soil shape parameters and a preset soil texture classification table; based on the soil type, acquiring a corresponding standard operation scheme from a preset agronomic knowledge base, wherein the standard operation scheme comprises a reference operation parameter value corresponding to the soil type; controlling the second agricultural machine to sequentially develop operations according to the sequence of the geographical subareas, and determining a target reference operation parameter value corresponding to the geographical subarea where the second agricultural machine is currently located; setting a controllable parameter of the second agricultural machine to the target reference operation parameter value when the second agricultural machine starts operation; during the operation process of the second agricultural machine, monitoring the actual working state parameters of the second agricultural machine in real time through a sensor; Performing deviation comparison on the actual working state parameter and the target reference working parameter value to obtain a deviation value; If the deviation value is greater than or equal to a preset allowable range, adjusting the working state of the working part, and recording the adjusted controllable working parameter value; after the second agricultural machine completes the operation of the target geographical subarea, constructing a controllable operation parameter value time sequence and an actual working state parameter time sequence recorded in the target geographical subarea into the process operation parameter set, wherein the target geographical subarea is any one geographical subarea of a plurality of geographical subareas.
3. The method according to claim 1, wherein said dividing each of said geographical sub-areas into different parametric job partitions based on controllable job parameter values in said process job parameter set, in particular comprises: Calculating the mean value and standard deviation of each controllable operation parameter in the process operation parameter set in all geographic subareas to obtain a parameter overall distribution benchmark; carrying out standardization processing on each controllable operation parameter value in the process operation parameter set corresponding to each geographical subarea based on the parameter overall distribution standard to obtain a standardized parameter score; weighting the normalized parameter fraction and a preset operation parameter weight vector to obtain a partition characteristic index; Performing cluster analysis on the partition characteristic indexes of all the geographic subareas to obtain a plurality of cluster clusters; and dividing the geographical subareas belonging to the same cluster into the parameter operation subareas.
4. The method according to claim 1, wherein said calculating, for each of said parametric job partitions, a representative yield value based on said real-world yield data for all geographical sub-areas within said parametric job partition, comprises: Acquiring actual yield data of all geographic subareas in the parameter operation partition, and calculating an area weighted average value of the actual yield data as an area weighted yield average value of the parameter operation partition according to the area occupation ratio of each geographic subarea; Calculating the sum of squares of the deviations of the actual yield data relative to the area weighted yield average value to obtain a yield deviation measure of the parameter operation partition; If the yield deviation measure is smaller than a preset threshold value, determining the area weighted yield average value as a representative yield value of the parameter operation partition; if the yield deviation measure is greater than or equal to a preset threshold, clustering and grouping the yield data of the parameter operation partition, and identifying a main yield level cluster; and respectively calculating the area weighted yield average value of each main yield level cluster, and determining the area weighted yield average value corresponding to the main yield level cluster with the largest sample size in each main yield level cluster as the representative yield value of the parameter operation partition.
5. The method according to claim 1, characterized in that said comparing the actual yield data of each of said geographical sub-areas with corresponding representative yield values, identifies and marks a disturbed geographical sub-area, in particular comprising: Calculating the yield deviation degree between the actual yield data of each geographical subarea and the representative yield value of the corresponding parameter operation subarea; determining a plurality of adjacent geographic subareas in a preset space neighborhood of each geographic subarea; calculating the space autocorrelation index of the actual yield data of the geographic subareas and the actual yield data of all corresponding adjacent geographic subareas; Based on the yield bias and the spatial autocorrelation index, a disturbed geographical sub-area is identified and marked.
6. The method according to claim 5, wherein the spatial autocorrelation index is a molan index, and wherein identifying and marking the disturbed geographical sub-area based on the yield bias degree and the spatial autocorrelation index specifically comprises: If the yield deviation degree is larger than a preset deviation threshold value and the Morlan index is a negative value, determining that the geographical subarea is a space abnormal cold point, and marking the geographical subarea as a negative disturbed geographical subarea; And if the yield deviation degree is larger than a preset deviation threshold value and the Morlan index is a positive value, determining that the geographical subarea is a space abnormal hot spot, and marking the geographical subarea as a forward disturbed geographical subarea.
7. The method according to claim 1, wherein generating an optimized job strategy based on the correspondence between the process job parameter set and the corrected yield data set, in particular comprises: Extracting input features from the set of process operating parameters and the set of corrected yield data; taking the input characteristics as input, taking the yield value in the corrected yield data set as a training target, and training a yield prediction model; Dividing the target land block into a plurality of job partitions, and generating a group of to-be-selected parameter sets containing different parameter combinations for each job partition; Calculating a predicted yield value under each parameter combination in a corresponding to-be-selected parameter set by using the yield prediction model for each operation partition, and calculating an operation cost value based on a cost consumption associated with the parameter combination; constructing an optimization model taking the predicted yield value and the operation cost value as variables, and solving the optimization model for each operation partition to obtain a determined partition optimal parameter; and associating the geographical information of each operation partition with the partition optimal parameters to generate a variable operation prescription chart for forming the optimized operation strategy.
8. A multi-machine intelligent agricultural joint operation optimization device, comprising one or more processors and a memory, the memory coupled with the one or more processors, the memory to store computer program code, the computer program code comprising computer instructions, the one or more processors to invoke the computer instructions to cause the multi-machine intelligent agricultural joint operation optimization device to perform the method of any of claims 1-7.
9. A computer readable storage medium comprising instructions that, when run on a multi-machine intelligent agricultural joint operation optimization device, cause the multi-machine intelligent agricultural joint operation optimization device to perform the method of any one of claims 1-7.
10. A computer program product, characterized in that the computer program product, when run on a multi-machine intelligent agricultural joint operation optimization device, causes the multi-machine intelligent agricultural joint operation optimization device to perform the method according to any one of claims 1-7.

Description

Multi-machine intelligent agriculture combined operation optimization method, device and medium Technical Field The application relates to the field of intelligent agriculture, in particular to a multi-machine intelligent agriculture combined operation optimization method, device and medium. Background In modern agricultural production, the application of intelligent agricultural machinery such as tractors, rice transplanting machines, harvesters and the like is increasingly popular, and becomes key equipment for promoting agricultural modernization. When the advanced machines execute various operation links such as cultivation, seed, pipe and harvest, various sensors can be carried, and massive operation process data and environment sensing data with high space-time resolution can be acquired and recorded in real time. For example, the tractor can acquire the information of soil compactness, water content, organic matter content and the like during tillage, the transplanter or seeder can accurately record the parameters of operation track, seeding/seedling discharging quantity, plant row spacing, operation depth and the like, and the combine harvester can generate a refined yield distribution map and record the quality information of grain water content, protein content and the like. However, in the conventional operation mode, these advanced machines often operate individually, and mass data generated during the operation, such as soil information, transplanting parameters, harvest yield, etc., are usually stored in a separate on-board system or cloud platform to form an information island. The state of data rupture and lack of effective linkage among all operation links leads to the failure of the follow-up operation to fully utilize the information of the front operation and the failure to systematically feed back and optimize the front end cultivation and seed links according to the final output result, thereby limiting the accuracy of agricultural production decisions. Disclosure of Invention The application provides a multi-machine intelligent agriculture combined operation optimization method, a multi-machine intelligent agriculture combined operation optimization device and a multi-machine intelligent agriculture combined operation medium, and the accuracy of agricultural decisions is improved. The first aspect of the application provides a multi-machine intelligent agricultural combined operation optimization method, which comprises the steps of controlling a first agricultural machine to execute first-stage operation on a target land block to obtain a first-stage result, controlling a second agricultural machine to execute second-stage operation on a plurality of geographical subareas in the target land block based on the first-stage result to generate a process operation parameter set corresponding to each geographical subarea, controlling a third agricultural machine to harvest the target land block to acquire actual yield data corresponding to each geographical subarea, dividing each geographical subarea into different parameter operation subareas based on controllable operation parameter values in the process operation parameter set, wherein all geographical subareas belonging to the same parameter operation subarea correspond to the same controllable operation parameter values, calculating a representative yield value based on the actual yield data of all geographical subareas in each parameter operation subarea for each parameter operation subarea, comparing the actual yield data of each geographical subarea with a corresponding representative yield value, acquiring actual yield data corresponding to each geographical subarea, identifying the actual yield value as a corrected geographical yield value, correcting the geographical yield value of each geographical subarea based on the actual yield value, and a geographical yield value of the corrected subarea is generated based on the actual yield value of the corrected geographical yield value. By adopting the technical scheme, the intelligent agricultural combined operation optimizing method for multiple varieties is provided, and the intelligent management and control of the whole agricultural production process is realized by controlling the cooperative operation of the first agricultural machine, the second agricultural machine and the third agricultural machine. The first agricultural machine and the second agricultural machine execute the first stage operation and the second stage operation respectively, a process operation parameter set reflecting the growth conditions of soil and crops is obtained, and a data basis is provided for the subsequent optimization decision. And the third agricultural machine performs harvesting operation, acquires refined actual yield data, and provides support for yield difference analysis. The method comprises the steps of dividing a land block into parameter operation partitions based on process paramete