CN-121095005-B - Multi-sensor fusion-based paddy field soil-rotation dynamic response analysis method and system

CN121095005BCN 121095005 BCN121095005 BCN 121095005BCN-121095005-B

Abstract

The invention discloses a paddy field soil-rotation dynamic response analysis method and system based on multi-sensor fusion, which relate to the technical field of agricultural information sensing, acquire soil response data from different depths and directions, acquire a high-correlation response feature subset after processing the soil response data based on a multi-layer information fusion network model, acquire a soil physical field excitation sequence under a multi-layer coupling boundary condition based on feature subset inversion, perform multi-time-domain parallel response analysis on a target area, acquire a stress-pore pressure coupling peak area and spatial distribution thereof in response output, construct a soil weak response fingerprint map, and finally propose optimization suggestions for a soil-rotation path, an operation rhythm and pressure parameters based on the map, so that dynamic pre-response evaluation and decision support before paddy field soil-rotation operation are realized.

Inventors

LIU CUIHONG
XU YAN
KONG AIJU
WU LIYAN
CUI HONGGUANG
ZHANG XINYUE
ZHANG HUI
WANG HUAN
XIN MINGJIN
YUE XIANG
FENG LONGLONG
SONG YUQIU

Assignees

沈阳农业大学

Dates

Publication Date: 20260508
Application Date: 20250901

Claims (8)

1. The paddy field soil-wheel dynamic response analysis method based on multi-sensor fusion is characterized by comprising the following steps: constructing a multi-layer information fusion network model M fused with earth surface stress disturbance signals, pore water pressure change trend, shear wave propagation speed and soil-water coupling nonlinear indexes, and representing dynamic evolution characteristics of multiple physical properties of soil; Synchronously collecting soil response data from different depths and spatial positions in a target paddy field region, wherein the soil response data comprises surface acceleration, pore water pressure, a shear modulus disturbance spectrum and a reflected wave time domain signal; Inputting the soil response data into a multi-layer information fusion network model M, and performing characteristic decoupling and mutual information optimization processing to obtain a characteristic subset T; mapping the spatial attributes in the feature subset T into simulation area grids, calculating the similarity of adjacent simulation area grids, and marking the similarity as a soil dynamic finite response unit if the similarity is larger than a preset threshold; Constructing a soil dynamic finite response unit, and generating a soil physical field excitation sequence under a multilayer coupling boundary condition according to physical characteristics in the characteristic subset T, wherein the soil physical field excitation sequence is used for driving the soil dynamic finite response unit; Applying a wheel mud load function in the mud dynamic finite response units, and performing multi-time-domain parallel response analysis on the target area to obtain time-varying stress and pore pressure response data in each response unit; Identifying a stress-pore pressure coupling peak area and a spatial distribution curve thereof in the response analysis result, and extracting a high-risk response unit to form a soil weak response map P; and providing a wheel mud path optimization, an operation rhythm adjustment and a pressure parameter setting suggestion based on the mud weak response map P, and realizing dynamic pre-response evaluation and operation parameter optimization before paddy field wheel mud operation.
2. The paddy field soil dynamic response analysis method based on multi-sensor fusion according to claim 1 is characterized by comprising the steps of collecting training sample data of different physical quantities in a paddy field area, wherein the physical quantities comprise stress disturbance signals, pore water pressure changes, shear wave propagation speeds and nonlinear coupling indexes, aligning various sample data according to time synchronism to form a multi-scale input sequence, and constructing a multi-layer information fusion model M through a graph neural network, wherein nodes represent sensing types, and edges represent physical coupling relations.
3. The paddy field soil dynamic response analysis method based on multi-sensor fusion according to claim 2 is characterized by comprising the steps of arranging multi-type sensor nodes in a target paddy field area according to a preset grid, collecting surface acceleration, underground pore pressure, shear wave and reflected wave data from different depths, denoising and normalizing original soil response data to form a multi-dimensional data set.
4. The paddy field soil dynamic response analysis method based on multi-sensor fusion according to claim 3 is characterized in that a multi-dimensional dataset is input into a multi-layer information fusion network model M, a coupling characteristic tensor is extracted, and a maximum correlation minimum redundancy strategy is adopted to screen high-correlation response variables to form a characteristic subset T.
5. The paddy field soil dynamic response analysis method based on multi-sensor fusion is characterized by comprising the steps of carrying out double-channel extremum tracking on stress-pore pressure response data output by each unit, marking spatial nodes where local peaks of stress and pore pressure occur simultaneously, mapping the high-coupling peak nodes into a spatial distribution curve, extracting high-risk response units, and constructing a soil weak response map P by taking the high-risk response units as indexes and the coupling strength as weights.
6. The method for analyzing the dynamic response of the soil in the paddy field based on the multi-sensor fusion is characterized in that a path avoidance graph is generated for a high-risk response unit in a soil weak response graph P, a graph risk level is taken as a path weight, and a minimum risk path algorithm is adopted to generate a working track.
7. The method for analyzing the dynamic response of the paddy field soil and the soil based on multi-sensor fusion according to claim 1, wherein the multi-layer information fusion network model M adopts a graph rolling network idea, and a node state updating formula is as follows: ; representing that in the layer i network, the new feature vector of node u is equal to the weighted sum of the feature vectors of its neighboring nodes, Representing that in the layer 1 network, the new feature vector of the node u is equal to the weighted sum of the feature vectors of the neighboring nodes; The representation is calculated by the similarity between the node v and the node u characteristics; representing the trainable weight matrix of layer i, σ is the activation function.
8. A multi-sensor fusion-based paddy field soil dynamic response analysis system for realizing the multi-sensor fusion-based paddy field soil dynamic response analysis method as set forth in any one of claims 1 to 7, characterized by comprising: the data driving modeling module is used for constructing a multi-layer information fusion network model M which fuses surface stress disturbance signals, pore water pressure change trend, shear wave propagation speed and soil-water coupling nonlinear indexes and is used for representing dynamic evolution characteristics of multiple physical properties of the soil; The multi-source sensing acquisition module synchronously acquires soil response data from different depths and spatial positions in a target paddy field region, wherein the soil response data comprises surface acceleration, pore water pressure, shear modulus disturbance spectrum and reflected wave time domain signals; The response modeling module inputs the soil response data into a multi-layer information fusion network model M, performs characteristic decoupling and mutual information optimization processing to obtain a high-correlation response characteristic subset T, maps the spatial attribute in the characteristic subset T into simulation area grids, calculates the similarity of adjacent simulation area grids, and marks the similarity as a soil dynamic finite response unit if the similarity is larger than a preset threshold value; The response simulation module is used for constructing a soil dynamic finite response unit, generating a soil physical field excitation sequence under a multi-layer coupling boundary condition according to physical characteristics in the characteristic subset T and driving the soil dynamic finite response unit; The response analysis module is used for applying a wheel mud load function in the mud dynamic finite response units, performing multi-time-domain parallel response analysis on the target area, and acquiring time-varying stress and pore pressure response data in each response unit; The response map generation module is used for identifying a stress-pore pressure coupling peak area and a spatial distribution curve thereof in the response analysis result, extracting a high-risk response unit and forming a soil weak response map P; And the feedback module is used for proposing wheel mud path optimization, operation rhythm adjustment and pressure parameter setting advice based on the mud weak response map P, and realizing dynamic pre-response evaluation and operation parameter optimization before paddy field wheel mud operation.

Description

Multi-sensor fusion-based paddy field soil-rotation dynamic response analysis method and system Technical Field The invention relates to the technical field of agricultural information perception, in particular to a paddy field soil-wheel dynamic response analysis method and system based on multi-sensor fusion. Background The paddy field mud-turning operation is a key link in the agricultural production of rice crops, and mainly realizes shaping, loosening and cement mixing of the soil of a cultivated layer through mechanical disturbance, and the execution effect of the paddy field mud-turning operation has obvious influence on the aspects of root growth, moisture retention, nutrient distribution, disease and pest prevention and control and the like of crops. However, because the soil in the paddy field has complex structure, large fluctuation of water content and uneven mechanical load effect, the traditional method is difficult to effectively evaluate the real-time dynamic response of the soil in the mud-turning process, often depends on experience or post observation, and lacks a quantifiable and predictable scientific basis. With the development of agricultural informatization, the state sensing of paddy field soil by introducing a sensor monitoring technology has been tried, but the sensor monitoring technology is limited by low information dimension obtained by a single sensing source, and the combined analysis of multiple physical quantities such as soil stress, pore water pressure, deformation and the like under dynamic load conditions cannot be realized, so that hysteresis and uncertainty exist in an evaluation result. The method for fusing multisource sensing information and combining physical modeling and dynamic response calculation is needed at present, and can be used for carrying out high-precision analysis and simulation on the real response process of the soil medium in paddy field wheel soil operation, so that support is provided for agricultural machinery parameter optimization and paddy field management strategies. Disclosure of Invention The invention aims to provide a paddy field soil dynamic response analysis method and system based on multi-sensor fusion, which are used for solving the defects in the background technology. In order to achieve the purpose, the invention provides the following technical scheme that the paddy field soil dynamic response analysis method based on multi-sensor fusion comprises the following steps: constructing a multi-layer information fusion network model M fused with earth surface stress disturbance signals, pore water pressure change trend, shear wave propagation speed and soil-water coupling nonlinear indexes, and representing dynamic evolution characteristics of multiple physical properties of soil; Synchronously collecting soil response data from different depths and spatial positions in a target paddy field region, wherein the soil response data comprises surface acceleration, pore water pressure, a shear modulus disturbance spectrum and a reflected wave time domain signal; inputting the soil response data into a multi-layer information fusion network model M, and performing characteristic decoupling and mutual information optimization processing to obtain a characteristic subset T; constructing a soil dynamic finite response unit, and inverting a soil physical field excitation sequence under a multi-layer coupling boundary condition according to physical characteristics in the characteristic subset T; Applying a wheel mud load function in the mud dynamic finite response units, and performing multi-time-domain parallel response analysis on the target area to obtain time-varying stress and pore pressure response data in each response unit; Identifying a stress-pore pressure coupling peak area and a spatial distribution curve thereof in the response analysis result, and extracting a high-risk response unit to form a soil weak response map P; and providing a wheel mud path optimization, an operation rhythm adjustment and a pressure parameter setting suggestion based on the mud weak response map P, and realizing dynamic pre-response evaluation and operation parameter optimization before paddy field wheel mud operation. Preferably, training sample data of different physical quantities in a paddy field area are collected, wherein the physical quantities comprise stress disturbance signals, pore water pressure changes, shear wave propagation speeds and nonlinear coupling indexes, various sample data are aligned according to time synchronism to form a multi-scale input sequence, a multi-layer information fusion model M is constructed through a graph neural network, nodes represent sensing types, and edges represent physical coupling relations. Preferably, a plurality of types of sensing nodes are distributed in a target paddy field area according to a preset grid, surface acceleration, underground pore pressure, shear wave and reflected wave data from diffe