CN-121978307-A - Intelligent orchard soil moisture content real-time monitoring and early warning system

CN121978307ACN 121978307 ACN121978307 ACN 121978307ACN-121978307-A

Abstract

The invention discloses a real-time monitoring and early warning system for soil moisture content of an intelligent orchard. According to the invention, a multi-depth sensor node network is deployed through a low-power-consumption wide-area network structure, real-time soil moisture content data of a main distribution layer covering a root system of a fruit tree is periodically acquired, industrial information processing is carried out on the soil moisture content data, synchronous meteorological data and soil attribute data to generate a standard data set, the data set is further input into a pre-trained water-required dynamic evaluation model for outputting a dynamic water shortage grade signal in combination with a fruit tree fertility stage, soil texture and future meteorological forecast, and finally an accurate regulation and control instruction is generated according to the signal and an irrigation strategy rule, so that the problems of low accuracy and poor monitoring and irrigation decision linkage timeliness caused by the fact that the actual moisture distribution deviation of a multipoint interpolation or single-depth monitoring and root system is difficultly balanced in power consumption and transmission stability and the early warning model are not combined with the meteorological information are solved, and the accuracy of orchard soil moisture monitoring and early warning accuracy and irrigation regulation and the timeliness are improved.

Inventors

CHANG YUANSHENG
WANG SEN
ZHENG WENYAN
HE XIAOWEN
WANG YONGXU
WANG HAIBO
HE PING
LI LINGUANG

Assignees

山东省果树研究所

Dates

Publication Date: 20260505
Application Date: 20260126

Claims (10)

1. Wisdom orchard soil moisture content real-time supervision and early warning system, its characterized in that, the system includes: the data acquisition module is used for periodically acquiring real-time soil moisture content data of a main distribution layer covering the root system of the fruit tree through a multi-depth sensor node network deployed based on a low-power-consumption wide area network architecture; the data processing module is used for carrying out industrial information processing on the real-time soil moisture content data, the weather data which are synchronous with the real-time soil moisture content data in a space-time mode and the soil attribute data, and generating a standard soil moisture content data set containing the moisture distribution characteristics of the root system layer; the dynamic early warning module is used for inputting the standard soil moisture content data set, the current fruit tree growth stage identification, the soil texture classification information and the future weather forecast data into a pre-trained fruit tree water-demand dynamic assessment model and outputting a dynamic water shortage grade signal associated with a specific irrigation area; And the intelligent decision module is used for generating an accurate regulation and control instruction pointing to a specific irrigation execution terminal according to the dynamic water shortage grade signal and a preset irrigation strategy rule.
2. The intelligent orchard soil moisture content real-time monitoring and early warning system according to claim 1, wherein the industrial information processing executed by the data processing module comprises a space-time kriging interpolation process, and the space-time kriging interpolation process is used for complementing space missing data generated by node failure and time breakpoint data generated by communication delay in the multi-depth sensor node network.
3. The intelligent orchard soil moisture content real-time monitoring and early warning system according to claim 1, wherein the multi-depth sensor node network is configured with an adaptive scheduling strategy, and the adaptive scheduling strategy dynamically adjusts the periodically acquired data acquisition frequency according to the precipitation probability in the future weather forecast data.
4. The intelligent orchard soil moisture content real-time monitoring and early warning system according to claim 1, wherein the fruit tree water-required dynamic assessment model is an integrated learning model, and the integrated learning model fuses output results of a first sub-model taking the standard soil moisture content data set as input and a second sub-model taking the future weather forecast data as input.
5. The intelligent orchard soil moisture content real-time monitoring and early warning system according to claim 4, wherein the first sub-model is a convolutional neural network and is used for extracting spatial correlation characteristics of the real-time soil moisture content data among different soil layers, and the second sub-model is a time sequence prediction network and is used for processing time sequence changes of the future weather forecast data.
6. The intelligent orchard soil moisture content real-time monitoring and early warning system according to claim 1, wherein the preset irrigation strategy rule is a multi-objective optimization rule, and the multi-objective optimization rule simultaneously considers the dynamic water shortage level signal, a preset total irrigation water constraint and the evaporation power parameter in the future weather forecast data.
7. The intelligent orchard soil moisture content real-time monitoring and early warning system according to claim 1, further comprising a visual interaction module, wherein the visual interaction module is used for receiving and fusing output data of the dynamic early warning module and the intelligent decision module, and generating a comprehensive situation map comprising soil moisture content space-time distribution, early warning areas and recommended irrigation schemes.
8. The intelligent orchard soil moisture content real-time monitoring and early warning system according to claim 4, further comprising an offline training module, wherein the offline training module uses historical data to train the fruit tree water-required dynamic assessment model, and a loss function adopted during training is as follows: Wherein, the Representing a loss value; Representing the total number of training samples; represent the first True water-required labels of the samples; Represents the pair of the dynamic evaluation model for fruit tree water demand Predicting water demand values of the samples; representing the number of the samples adjacent to the space in the standard soil moisture content data set; represent the first Predicting a water demand gradient for spatially adjacent samples; the weighting coefficients of the regularization term are spatially smoothed.
9. The intelligent orchard soil moisture content real-time monitoring and early warning system according to claim 1, wherein the dynamic early warning module introduces a soil water stress index when calculating the dynamic water shortage level signal As a core criterion, the calculation formula of the soil water stress index is as follows: Wherein, the Representing a soil water stress index; Representing the effective water content of the current root system layer obtained by calculation according to the standard soil moisture content data set and the soil texture classification information; a water critical threshold value related to the current fruit tree growth stage identification; is a sensitivity parameter; representing the actual evapotranspiration calculated from the recent environmental data; representing potential amounts of evapotranspiration from the future weather forecast data; Is an exponential function.
10. The intelligent orchard soil moisture content real-time monitoring and early warning system according to claim 6, wherein the mathematical expression of the multi-objective optimization rule is a minimization objective function : The constraint conditions are as follows: Wherein, the Representing the objective function value that needs to be minimized; representing predicted total irrigation water consumption, which is a decision variable (Irrigation duration), (Initial soil moisture content characterized by the Standard soil moisture content dataset) and A function of (predicted evapotranspiration from the future weather forecast data); an ideal water consumption for meeting the dynamic water shortage grade signal release requirement; representing the square of the euclidean norm; Punishment coefficients for the duration; The maximum allowable duration of single irrigation is set; Constraint on the total amount of irrigation water.

Description

Intelligent orchard soil moisture content real-time monitoring and early warning system Technical Field The invention relates to the technical field of intelligent agricultural monitoring and early warning, in particular to a real-time monitoring and early warning system for soil moisture content of an intelligent orchard. Background Along with the rapid development of intelligent agriculture, orchard accurate management provides higher requirements for soil moisture content monitoring. The prior art mainly relies on a multipoint interpolation method or a single depth sensor in the aspect of soil moisture content data acquisition, so that a monitoring result has deviation with actual water distribution of a root system of a fruit tree, and the equipment power consumption and the data real-time transmission stability are difficult to balance when the high-frequency continuous monitoring is performed, and particularly when the large-area orchard is deployed, the problems of signal coverage blind areas and node failure obviously influence monitoring continuity. In the early warning mechanism level, the current system mostly adopts a fixed threshold judgment or simple linear model, and cannot be effectively combined with water demand characteristics, weather forecast information and soil texture differences of different growth stages of the fruit trees to carry out dynamic adjustment, so that early warning accuracy is low, and false warning omission occurs frequently. In addition, the traditional monitoring data and irrigation decisions lack effective linkage, the timeliness of the early warning information converted into the regulation and control instruction is insufficient, and the requirement for fine water and fertilizer management is difficult to support. Accordingly, there is a need to provide a solution to the above-mentioned problems. Disclosure of Invention In order to solve the technical problems, the invention provides a real-time monitoring and early warning system for soil moisture content of an intelligent orchard, which has the following technical scheme: the data acquisition module is used for periodically acquiring real-time soil moisture content data of a main distribution layer covering the root system of the fruit tree through a multi-depth sensor node network deployed based on a low-power-consumption wide area network architecture; the data processing module is used for carrying out industrial information processing on the real-time soil moisture content data, the weather data which are synchronous with the real-time soil moisture content data in a space-time mode and the soil attribute data, and generating a standard soil moisture content data set containing the moisture distribution characteristics of the root system layer; the dynamic early warning module is used for inputting the standard soil moisture content data set, the current fruit tree growth stage identification, the soil texture classification information and the future weather forecast data into a pre-trained fruit tree water-demand dynamic assessment model and outputting a dynamic water shortage grade signal associated with a specific irrigation area; And the intelligent decision module is used for generating an accurate regulation and control instruction pointing to a specific irrigation execution terminal according to the dynamic water shortage grade signal and a preset irrigation strategy rule. Further, the industrial information processing executed by the data processing module comprises a space-time kriging interpolation process, wherein the space-time kriging interpolation process is used for complementing space missing data generated by node failure and time breakpoint data generated by communication delay in the multi-depth sensor node network. Further, the multi-depth sensor node network is configured with an adaptive scheduling strategy, and the adaptive scheduling strategy dynamically adjusts the periodically acquired data acquisition frequency according to the precipitation probability in the future weather forecast data. Furthermore, the fruit tree water-required dynamic assessment model is an integrated learning model, and the integrated learning model fuses the output result of a first sub-model taking the standard soil moisture content data set as input and a second sub-model taking the future weather forecast data as input. Furthermore, the first sub-model is a convolutional neural network and is used for extracting spatial correlation characteristics of the real-time soil moisture content data between different soil layers, and the second sub-model is a time sequence prediction network and is used for processing time sequence changes of the future weather forecast data. Further, the preset irrigation strategy rule is a multi-objective optimization rule, and the multi-objective optimization rule simultaneously considers the dynamic water shortage grade signal, the preset total irrigation water constraint and