A system for evaluating crop adaptability and providing planting recommendations for arable land

Abstract

This invention relates to the field of intelligent agricultural environmental monitoring technology, and discloses a system for evaluating crop adaptability and providing planting suggestions for arable land. The system comprises multiple distributed micro-environment decision-making units. Each unit collects soil moisture and light data in real time through environmental sensors, combines this data with slope information obtained from tilt sensors, and uses a local processor to calculate entropy values ​​representing crop stress risk based on a dynamic weight adjustment algorithm. Furthermore, it achieves entropy value interaction and collaborative decision-making among units through a low-power communication network. This invention combines micro-topographic features with distributed computing, and through a slope-driven entropy weight adaptive mechanism, enables the system to autonomously identify local hydrological stress risks and form a collaborative response network based on physical laws among units, achieving accurate perception and intelligent decision-making for complex farmland environments.

Description

Technical Field

[0001] This invention relates to a system for evaluating crop adaptability and providing planting recommendations for arable land, belonging to the field of intelligent agricultural environmental monitoring technology. Background Technology

[0002] In the field of intelligent decision-making in agricultural environment, data processing systems based on prediction mainly rely on satellite remote sensing and historical data from weather stations to build centralized models and generate crop planting suggestions through cloud computing. Such systems treat arable land as homogeneous units, and their static decision-making logic faces significant limitations in real-world scenarios with complex micro-topography.

[0003] Taking the terraced fields in the hilly areas of southern China as an example, when encountering sudden rainstorms, the unified corn planting suggestions given by the centralized model based on macro data cannot respond to local hydrological dynamics: waterlogging stress is caused by runoff accumulation in low-lying areas, while drought stress is caused by rapid water loss in steep slopes. The fundamental bottleneck lies in the mismatch between the scale of environmental perception and the decision-making response mechanism - the data cannot capture the coupling effect between micro-topography and short-term meteorological events, resulting in the omission of key stress risks.

[0004] Industry attempts to improve data accuracy by increasing sensor density have failed to address the core challenges: 1. Centralized processing architectures inherently suffer from latency, making it difficult to respond promptly to transient environmental changes; 2. Static model weights cannot adapt to slope-driven local hydrological heterogeneity; 3. The lack of a risk transmission mechanism between neighboring units in the decision-making logic prevents coordinated defense from being triggered by local disasters. These deficiencies lead to a sharp drop in the system's prediction confidence in complex farmland scenarios, and may even trigger cascading crop losses. Therefore, how to construct a crop adaptability evaluation system that can autonomously perceive micro-topographic features, respond to environmental dynamics in real time, and achieve distributed collaborative decision-making has become the technical problem that this invention aims to solve. Summary of the Invention

[0005] This invention provides a system for evaluating crop adaptability and providing planting recommendations for arable land. Its main purpose is to solve the problem of inaccurate crop adaptability evaluation caused by the neglect of micro-topography and dynamic changes in the local environment in existing centralized prediction systems.

[0006] To achieve the above objectives, the present invention provides a system for evaluating crop adaptability to arable land and providing planting recommendations, comprising:

[0007] Multiple microenvironment decision units are deployed within the cultivated land area. Each microenvironment decision unit includes: an environmental sensor for sensing local soil moisture and ambient light intensity; and a tilt sensor for sensing the deployment tilt angle of the microenvironment decision unit.

[0008] A local processor is configured to: calculate a local entropy value indicating the level of crop growth risk based on local soil moisture and ambient light intensity sensed by environmental sensors and built-in crop physiological threshold rules; and dynamically adjust the weights of environmental factors in the local entropy value calculation rules based on the deployment tilt angle sensed by tilt sensors, wherein when the deployment tilt angle is greater than a first angle threshold, the weights of environmental factors related to drought stress are increased; and when the deployment tilt angle is less than a second angle threshold, the weights of environmental factors related to waterlogging stress are increased.

[0009] A communication module is used to broadcast local entropy values ​​to neighboring microenvironment decision units at set time intervals; and a local processor is further configured to: receive neighborhood entropy values ​​broadcast from neighboring microenvironment decision units; and generate crop planting recommendations based on its own calculated local entropy values ​​and the received neighborhood entropy values, through set collaborative decision rules.

[0010] Preferably, the tilt sensor is a triaxial accelerometer, and the local processor is configured to: analyze the readings of the triaxial accelerometer to determine the deployment tilt angle when the microenvironment decision unit is stationary.

[0011] Preferably, the local processor is further configured to: activate a weight negotiation mode when the increase in local soil moisture detected by the environmental sensor exceeds a first humidity threshold within a set time period; in this mode, the microenvironment decision unit broadcasts a weight adjustment announcement containing its own deployment tilt angle information and the planned adjustment of entropy value calculation weight to neighboring microenvironment decision units through the communication module; and the local processor is further configured to: after receiving the weight adjustment announcement from a neighboring microenvironment decision unit, execute asymmetric adoption logic to correct its own entropy value calculation weight based on the difference between its own deployment tilt angle and the deployment tilt angle of the neighboring microenvironment decision unit.

[0012] Preferably, the asymmetric adoption logic executed by the local processor includes: when the deployment tilt angle of the microenvironment decision unit itself is less than the deployment tilt angle of the neighboring microenvironment decision units, the microenvironment decision unit adopts the suggestions of the neighboring microenvironment decision units regarding the adjustment of the water conservation weight by a proportion higher than that adopted by itself regarding the adjustment of the drainage weight.

[0013] Preferably, the local processor is configured to: when its own calculated local entropy value does not exceed a first entropy threshold, but the entropy values ​​of its received neighboring microenvironment decision units are used to calculate the neighborhood risk accumulation value through weighted summation. Exceeding the second entropy threshold At the same time, the risk level of crop planting recommendations is raised, and crop varieties with stronger stress resistance are recommended, including: ,in, Indicates the first The entropy value of each neighboring microenvironment decision-making unit. Indicates the relationship with the first Distance-related weights of neighboring microenvironment decision-making units This represents the number of neighboring micro-environment decision units that receive data from the micro-environment decision unit.

[0014] Preferably, the environmental sensor also includes a light sensor, and the local processor is configured to: switch the light sensor to a high-frequency sampling mode at night to monitor fluctuations in ambient light intensity; and exchange timestamps of the light intensity fluctuation characteristics monitored by each other with the adjacent micro-environment decision unit through the communication module to collaboratively calculate vector information of the speed and direction of cloud movement; when the speed of the vector information exceeds a first speed threshold and the direction deviation is less than a third angle threshold, trigger the micro-environment decision unit to enter the early warning working state.

[0015] Preferably, when the local processor is in the early warning working state, it increases the frequency of local entropy value calculation and local entropy value broadcasting.

[0016] Preferably, the local processor is further configured to: when a rainfall event is detected, operate its tilt sensor in vibration sampling mode to acquire vibration signals caused by rainwater runoff; evaluate the reliability of statically sensing the deployment tilt angle by analyzing the spectral characteristics of the vibration signals and comparing them with the expected spectral characteristics corresponding to the deployment tilt angle sensed by the tilt sensor when stationary; and adjust its own weight to adjust the priority of the declaration based on the reliability evaluation results.

[0017] Preferably, the communication module of the microenvironment decision unit is further configured to: drive the communication module to generate a standardized mechanical vibration as an active detection signal and transmit it to the soil; and the local processor is further configured to: use an inclination sensor to collect the echo signal returned after the detection signal is reflected by the soil; determine the soil compaction information by analyzing the transit time and energy attenuation rate of the echo signal; and use the compaction information to correct the weights in its local entropy calculation rules.

[0018] Compared with the prior art, the beneficial effects of the present invention are:

[0019] 1. Through real-time interaction between the slope information analyzed by the tilt sensor and the local environmental perception data, the system enables each micro-environment decision-making unit to gain autonomous cognition of local hydrological characteristics. The slope data dynamically adjusts the stress factor weight in the entropy calculation, automatically strengthening the waterlogging risk judgment logic in low-lying areas, while steep slope units prioritize responding to drought signals. This mechanism of transforming static terrain parameters into dynamic decision variables allows the system to naturally form a distributed risk map that conforms to physical laws during rainfall events, avoiding the inherent defect of traditional models that mismatch macro-climate data with micro-habitat responses. The micro-environment decision-making unit replaces the original data transmission with periodic broadcast single-byte entropy values, building a neighborhood risk transmission channel with extremely low communication overhead. When the unit's own entropy value does not reach the warning threshold but the accumulated entropy value of the upstream unit exceeds the critical level, the system spontaneously raises the risk level and recommends stress-resistant crops. This design enables local hydrological correlation. For example, the reception of steep slope runoff in low-lying areas is transformed into group collaborative decision-making through entropy value transmission, upgrading disaster response from individual adaptation to watershed-level joint prevention and control, forming an autonomous stress-resistant network that does not require central scheduling.

[0020] 2. At night, the light sensor switches to a high-frequency sampling mode to capture the temporal characteristics of light intensity fluctuations caused by cloud cover blocking starlight. It exchanges timestamp data with neighboring units and collaboratively calculates cloud motion vectors. When cloud speed and direction meet the conditions for rainstorm formation, the system triggers an increase in the entropy calculation frequency in advance. This mechanism transforms the idle hardware into a weather radar network, enabling each unit to obtain short-term rainfall prediction capabilities and avoiding the time barrier of passive response in traditional systems. During rainfall events, the tilt sensor switches to a vibration sampling mode to collect runoff impact spectrum and compares it with the expected spectrum characteristics corresponding to the static slope. When the spectrum deviation reveals sensor attitude drift, the system automatically reduces the priority of that unit in weight negotiation and sends a calibration request to the user. This mechanism of using natural rainfall to achieve hardware status diagnosis enables the system to have self-healing capabilities against gradual physical environmental interference, ensuring the long-term reliability of core decision parameters. Attached Figure Description

[0021] Figure 1 This is a functional architecture diagram of a system for evaluating crop adaptability and providing planting recommendations for arable land, as described in this invention.

[0022] Figure 2 This is a graph showing the change of risk value over time in a simulated rainfall event using the system of the present invention.

[0023] Figure 3 This is a time sequence diagram of data processing and collaborative interaction of a single microenvironment decision-making unit in the system of this invention.

[0024] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be described with reference to the accompanying drawings and specific embodiments. However, it should be understood that the specific embodiments described herein are only for explaining the present invention and are not intended to constitute any form of limitation on the scope of protection of the present invention.

[0026] The arable land crop adaptability evaluation and planting suggestion system disclosed in this application has an overall architecture consisting of multiple distributed micro-environment decision units deployed within the arable land area. Each micro-environment decision unit physically integrates an environmental sensor for sensing local soil moisture and ambient light intensity, a tilt sensor for accurately determining its own deployment tilt angle, a local processor responsible for executing all calculations and decision-making logic, and a communication module for enabling low-power data interaction between units. These spatially discrete units, through autonomous sensing, local computation, and neighborhood collaboration, jointly construct an intelligent evaluation network capable of accurately responding to micro-topographical differences and local environmental dynamics. In specific operations… In the operational procedures, the realization of system capabilities begins with the direct response to the heterogeneity of micro-topography in complex farmland environments. The reason why traditional homogenization models are insufficiently adaptable in scenarios such as hilly terraces is that they cannot transform static slope parameters into real-time predictions of dynamic hydrological stress risks. In view of this, the local processor of each micro-environment decision unit is configured to execute a dynamic weight adjustment algorithm driven by slope information. The procedures begin with the initialization phase after the deployment of the micro-environment decision unit. When the unit is in a static state, its built-in triaxial accelerometer, which acts as a tilt sensor, accurately calculates the deployment tilt angle characterizing the local slope by analyzing the component readings of the gravity vector along its three orthogonal axes. This data is then stored as a baseline parameter. Once the system enters continuous operation, the local processor calculates a basic local entropy value based on real-time soil moisture and ambient light intensity data sensed by environmental sensors, and refers to built-in specific crop physiological threshold rules. This entropy value quantifies the comprehensive growth pressure faced by the crop in this microenvironment. The core of this calculation rule lies in the fact that the weights of environmental factors are not constant, but rather adjusted by the deployment tilt angle. Dynamic modulation is performed; specifically, a first angle threshold is set internally by the local processor. With the second angle threshold These two thresholds are critical inflection points determined through offline calibration experiments on soil types in the target area. The slope value defined as the point at which the water runoff rate begins to significantly exceed the infiltration rate in that soil, while This corresponds to the slope value of depressions that are prone to surface water accumulation; when the processor detects... When drought stress occurs, it will automatically increase the weight of environmental factors associated with drought stress in the entropy calculation model. For example, it will increase the weight coefficient of the indicator of excessively low soil moisture. Conversely, when drought stress is detected, it will automatically decrease the weight of environmental factors associated with drought stress in the entropy calculation model. When the slope is sloped, the weight of environmental factors related to waterlogging stress is increased. In this way, the system no longer regards the slope as just a static geographic information, but transforms it into a dynamic decision variable that can actively adjust the sensitivity of risk perception. This makes units deployed on steep slopes naturally more alert to drought signals, while units in low-lying areas have a higher warning priority for waterlogging risk. This achieves autonomous, differentiated and accurate identification of local hydrological stress risks.

[0027] Furthermore, to address scenarios where local risks propagate and spread along topographic gradients—for example, the risk of flooding in depressions is often caused by rapid runoff from upstream slopes—the system establishes a two-tiered collaborative decision-making mechanism that includes passive risk accumulation and active weight negotiation. Under normal conditions, after calculating its local entropy value, each unit's local processor broadcasts a small local entropy value representing its own risk level to neighboring units at set time intervals via its communication module. Simultaneously, it continuously receives neighboring entropy values ​​for passive risk accumulation assessment. When a unit's local entropy value does not exceed a preset first entropy threshold, but the entropy values ​​it receives from multiple neighboring units are weighted and summed using a specific collaborative decision-making rule, the resulting neighborhood risk accumulation value is calculated. However, it exceeded the preset second entropy threshold. At that time, the unit will be forced to upgrade its risk level, and its specific collaborative decision-making rules will be deterministically expressed as follows: ,in, Indicates the first The entropy value of the neighboring units, The number of effective neighboring units, and the weight. The determination follows a physical model that integrates spatial distance and elevation differences, namely... Not only is it inversely proportional to the physical distance between units, but it is also strongly positively correlated with their relative elevation, ensuring that risk signals from upstream slopes are given higher decision-making weight. Under specific triggering conditions, the system enters a more proactive weight negotiation mode. For example, when the environmental sensor detects that the increase in local soil moisture exceeds a first humidity threshold characterizing a rainstorm event within a set time period, the local processor activates the weight negotiation mode. In this mode, the unit no longer broadcasts a single entropy value to neighboring units through the communication module, but a weight adjustment announcement that includes its own deployment tilt angle information and the planned adjustment weight calculated from the entropy value. When receiving a message from a neighboring unit... After this announcement, the local processor will execute an asymmetric adoption logic based on the difference between its own deployment tilt angle and the deployment tilt angle of the announcing unit. Specifically, when its own deployment tilt angle is less than that of the neighboring unit, i.e., when it is in a downstream position, the processor will adopt the neighboring unit's suggestion on adjusting the water conservation weight with a higher adoption coefficient, and adopt its suggestion on adjusting the drainage weight with a lower coefficient. This two-layer mechanism enables the cooperation between units to go from simple risk value transmission to a deeper level of decision-making logic negotiation, thereby building a distributed autonomous resilience network that can provide both macro-level early warning and micro-level adaptation without central scheduling.

[0028] To further enhance the system's environmental prediction capabilities and long-term operational reliability, the system also integrates a series of self-diagnostic and proactive sensing modules based on the reuse of existing hardware functions. For example, during nighttime, the local processor can switch the light sensor to a high-frequency sampling mode. By monitoring and recording the timestamps of subtle fluctuations in ambient light intensity caused by starlight or moonlight being blocked by clouds, and exchanging the timestamps of their respective fluctuation characteristics with neighboring units, the system collaboratively calculates the speed and direction vector information of cloud movement. When this vector information indicates the approach of a large-scale, high-speed cloud cluster, the system can enter early warning mode, correspondingly increasing the frequency of local entropy calculation and broadcasting, thus gaining valuable response time for impending rainfall events. When a rainfall event occurs, the local processor instructs the triaxial accelerometer, which acts as a tilt sensor, to switch to vibration sampling mode. By collecting and analyzing the spectral characteristics of vibration signals caused by rainwater runoff impacting the ground, and comparing them with a standard spectral model expected based on the static deployment tilt angle, if the deviation exceeds a preset confidence threshold, it indicates that the sensor's physical attitude may have shifted, and the system will then reduce the unit's [speed / direction]. The priority declared in the weight negotiation ensures the long-term reliability of core decision parameters. Furthermore, the system deeply integrates communication and sensing functions. The local processor drives the communication module to generate a standardized, weak mechanical vibration as an active detection signal, which is transmitted to the soil. A tilt sensor, acting as a highly sensitive vibration pickup, collects the echo signal reflected back from the soil. By analyzing the transit time and energy decay rate of the echo signal, the system can deduce information about soil compaction and use this information as a new physical dimension to correct the weights related to crop root stress in the local entropy calculation rules in real time. Finally, to ensure the sustainability of the aforementioned multimodal, high-energy-consuming operation, a power management module is introduced. Based on the local processor's current operating mode—including normal mode, emergency entropy weight negotiation mode, nighttime high-frequency sampling mode, and vibration sampling mode—and combined with real-time light intensity sensed by environmental sensors, it dynamically adjusts the power supply strategy for the communication module and various sensors. This optimizes the energy consumption of the entire microenvironment decision-making unit to the maximum extent while ensuring the response speed of critical tasks.

[0029] Meanwhile, to ensure that the microenvironment decision-making unit described in this invention can withstand the physical impact of rainstorms, runoff erosion, and routine agricultural activities when deployed in complex farmland areas such as hills and terraces, and to guarantee its long-term stability and reliability, the specific implementation of this invention may further include a deployment procedure for physically reinforcing the unit. This procedure may include one or more of the following aspects: the core components of the microenvironment decision-making unit are encapsulated in a robust shell with a high protection level (e.g., IP67 or above); and, depending on the soil quality and slope of the deployment point, the unit can be firmly anchored in a predetermined position in various ways, such as by integrating the unit with a long metal pile and driving the pile into a deeper stable soil layer during deployment; and before large-scale deployment, the target area can be surveyed to place the unit in an area that avoids major flood discharge paths and is far from the edge areas where large agricultural machinery frequently operates, so as to reduce the risk of it encountering extreme physical damage from the source. These are all extended implementation methods known to those skilled in the art.

[0030] Example 1: In this example, within a terraced farming environment located in the hilly region of southern China, multiple microenvironmental decision-making units (MICs) are deployed across a group of plots with significant elevation differences. One unit is situated on a steep slope, while another is located in a depression at the base of the slope where runoff easily accumulates. On a midsummer afternoon, a high-intensity, short-duration thunderstorm system is rapidly approaching the area, posing distinct but interconnected AC threats to these two microenvironmental plots. The night before the thunderstorm, the system has already utilized its built-in high-frequency sampling mode based on a light sensor... The cloud movement vector monitoring function, achieved in collaboration with neighboring units through timestamp exchange, determined that a cloud cluster meeting the conditions for heavy rain formation was moving towards the area. This prediction triggered all micro-environment decision-making units to enter early warning mode in advance, correspondingly increasing the frequency of their local entropy calculations and broadcasts. This demonstrates the synergistic efficiency of the system in responding to challenges. By utilizing the cross-scenario function reuse of idle hardware, the meteorological forecasting capability is transformed into a pre-alert state for the entire distributed network, gaining a response time advantage for subsequent accurate decision-making. When thunderstorms arrive and rainfall begins in the early stages, areas located on steep slopes... Both units in the low-lying area detected a rapid increase in local soil moisture within a short period, exceeding a preset first moisture threshold. This event simultaneously triggered them to enter a weight negotiation mode. At this point, the core mechanism of the system to resolve the technical dilemma was revealed. The unit on the steep slope, based on its pre-defined deployment tilt angle (greater than the first angle threshold), generated and broadcast a weight adjustment announcement. Its core intention was to increase the drought stress weight associated with rapid water loss. Upon receiving this announcement, the unit in the low-lying area downstream immediately executed asymmetric adoption logic. By comparing the deployment tilt angles of both units, it identified itself as being in the hydrological downstream region. Therefore, it accepted the upstream unit's announcement regarding the weight adjustment for strengthening water conservation, i.e., preventing waterlogging, with a preset, higher adoption coefficient. This process resolved the contradiction between global instructions and local realities in traditional agricultural decision-making. The system did not rely on a central server for unified scheduling but instead, through autonomous announcements and asymmetric adoption between units, spontaneously formed an interconnected risk perception chain between the upstream and downstream plots of the entire terraced field, conforming to physical hydrological laws.

[0031] Furthermore, the system's architectural advantage lies in its ability to make edge nodes intelligent enough to achieve efficient collaboration simply by exchanging the most refined risk insights. During the peak of this thunderstorm event, steep slope plots experienced a sharp increase in local entropy values ​​calculated based on dynamically adjusted high drought stress weights due to rapid soil moisture inflow. In contrast, low-lying areas, despite having local soil moisture levels below the flooding threshold and thus lower local entropy values, had their local processors calculate a cumulative neighborhood risk value by weighted summation of received neighborhood entropy values. Because it assigns extremely high weight to the upstream steep slope plots. ,lead to The value rapidly exceeded the collaborative early warning threshold. This collaborative decision-making based on entropy transfer forces an increase in the risk level of the low-lying area unit. The resulting crop planting recommendations are not based on its immediate state, but on an early response to the upcoming flood peak runoff from upstream, which is predicted by high entropy values. It recommends more resilient and flood-resistant crop varieties, thereby effectively avoiding the chain reaction of planting losses that may be caused by decision lag.

[0032] Example 2: This example was conducted on an indoor test platform with precisely controllable environmental parameters. The platform consisted of a soil chamber with adjustable slope and a sprinkler system capable of simulating different rainfall intensities. The soil chamber was filled with homogenized typical loam and pre-embedded with a high-precision soil moisture sensor as the baseline for this experiment. Two microenvironmental decision units, Unit A and Unit B, were deployed at two key locations on the test platform. Unit A was located in the middle of a steep slope area with a slope set at 20°, while Unit B was located in a flat slope bottom area with a slope of 0°. This setup aimed to simulate typical upstream and downstream hydrological relationships in a terraced field environment. The control group employed a centralized decision-making model that continuously collected raw soil moisture data from units A and B, generating a globally unified risk assessment result by calculating their average. The experimental group, however, fully implemented the distributed decision-making system of this invention. Key experimental parameters, such as the setting of simulated rainfall intensity, followed a rigorous decision-making logic chain. The fundamental technical trade-off in setting these parameters lay in ensuring that the rainfall intensity was sufficient to generate significant surface runoff and localized waterlogging within a limited experimental time, thereby effectively stimulating the system's decision response, while avoiding excessive intensity that would cause the hydrological process to be too rapid to be effectively captured by the sensor system. Therefore, the rainfall intensity... The value of is related to the saturated hydraulic conductivity of the soil. Relatedly, the specific decision-making rule is that, in order to stably generate surface runoff on loam with a slope greater than 15°, rainfall intensity... It should be set at to Within the specified range, based on pre-determined measurements of the soil used. To ensure the repeatability of the experiment and the validity of the data, a constant simulated rainfall intensity was ultimately selected in this experiment.

[0033] After the experiment started, the sprinkler system began to continuously rain at the set intensity, simultaneously recording the global risk assessment value of the control group, the local entropy values ​​independently calculated and broadcast by units A and B in the experimental group, and the true value of the soil moisture sensor as a benchmark. After 10 minutes of the experiment, the system exhibited significant differences. The control group, using an averaging algorithm, still had a low global risk level and failed to identify any localized stress risks. However, in the experimental group, unit A, located on a steep slope, had its local processor already increasing the weight of drought-related environmental factors based on its deployment tilt angle. Therefore, although its surface soil moisture was still high, its local entropy value had already reached a high level because the system anticipated rapid water loss. Meanwhile, unit B, located at the bottom of the slope, had its local processor, through collaborative decision-making rules, increase the weight of its neighborhood risk accumulation value. In the calculation, the high entropy value from upstream unit A was given extremely high weight. This led to a rapid escalation of its comprehensive risk assessment, as shown in Table 1, whose data records reveal this dynamic evolution process.

[0034] Table 1: Comparison of risk assessment status of the test system during rainfall events.

[0035]

[0036] Referring to Table 1, at the same time, the low-risk assessment value of the control group was based on a simple average of the steep slope humidity of 52% and the slope base humidity of 58%. This macroscopic data masked the contrasting physical processes occurring locally: water loss on the steep slope and water accumulation at the slope base. In contrast, the decision of experimental unit A was based on a slope-driven entropy weight adaptive mechanism, anticipating drought stress and outputting a high entropy value. The high entropy decision of unit B was mainly attributed to the neighborhood risk accumulation value in the collaborative decision-making rule. Dominated by the high entropy signal of upstream unit A, rather than its own unsaturated soil moisture, this data comparison shows that the system architecture of the present invention can combine static terrain information with dynamic neighborhood risk transmission through distributed intelligence, thereby making accurate and predictive decisions that exceed the perception capabilities of a single node.

[0037] Example 3: This example combines Figures 1 to 3 This document describes the implementation of a system for evaluating the adaptability of arable land crops and providing planting recommendations. Figure 1As shown in the figure, this diagram illustrates the overall functional architecture of the system of the present invention. This architecture is based on multiple distributed micro-environment decision-making units, such as micro-environment decision-making unit A located on a steep slope and micro-environment decision-making unit B located in a depression. Each unit integrates an environmental sensor for collecting soil moisture and ambient light intensity, and a tilt sensor for acquiring slope angle. The collected data is processed by a local processor for entropy calculation and weight adjustment, and entropy broadcasting and collaborative decision-making are achieved through a communication module. Different units interact through entropy broadcasting and weight negotiation, and information is incorporated into the core algorithm process through distributed collaboration. This process first collects environmental data, then enters the slope weight adjustment stage, the core of which is to increase the drought weight for steep slopes and the waterlogging weight for depressions based on terrain. Next, the system performs local entropy calculation, which is based on biological physiological thresholds, and then achieves neighbor unit communication through entropy broadcasting. The system further accumulates neighborhood risks (the formula is...). The system comprehensively assesses risks through collaborative decision-making (which can enter a weighted negotiation mode) and incorporates the asymmetric adoption logic used in risk level assessment to ultimately provide crop planting recommendations, enabling the recommendation of stress-resistant varieties. The entire process is also supported by a series of auxiliary functional modules, including a cloud movement monitoring module for nighttime light sampling, a vibration signal analysis module for sensor self-calibration, a soil compaction detection module for active acoustic detection, and a power management module for implementing dynamic power supply strategies. The real-time decisions made by the core algorithm process will generate comprehensive system outputs, specifically including: providing crop planting recommendations with variety adaptability evaluation and risk level warnings; generating a distributed risk map reflecting local hydrological response and collaborative defense networks; performing real-time environmental monitoring for dynamic state adjustment and early warning state management; and outputting system maintenance information including sensor calibration reminders and equipment status diagnostics.

[0038] like Figure 2As shown, the horizontal axis represents time (minutes), and the vertical axis represents risk value. The graph contains three key curves: The first curve (solid line) represents the local entropy value of the upstream unit. This value rises rapidly after the start of rainfall, reaches its peak in about 25 to 30 minutes, and then falls back, reflecting the rapid prediction of runoff risk by the steep slope unit; The second curve (thin dashed line) represents the local entropy value of the downstream unit. Its risk value increases slowly over time (minutes), indicating that the risk assessed by the depression unit based solely on its own environmental parameters (such as soil moisture) has a significant lag; The third curve (thick dashed line) represents the cumulative risk value R of the downstream unit's neighborhood. This value is given a high weight in the calculation of the risk signal from the upstream unit, so its trend is highly correlated with the local entropy value of the upstream unit. It rises to the high-risk range in a very short time, indicating that the downstream unit successfully foresaw and responded to the risk transmitted from the upstream through a collaborative decision-making mechanism. Its decision-making was not limited by its own lagging local entropy value, thus verifying the advanced nature of the distributed collaborative early warning of this invention.

[0039] like Figure 3 As shown in the diagram, this sequence diagram reveals in detail the core interaction process within and between individual microenvironment decision-making units in the system. This process involves five key roles: environmental sensor, tilt sensor, local processor, communication module, and neighboring units. The process begins in the data acquisition phase, where the environmental sensor sequentially sends soil moisture data and ambient light intensity data to the local processor. Simultaneously, the tilt sensor sends the deployment tilt angle to the local processor. After receiving all the sensor data, the local processor first performs internal calculations, adjusting the weights according to the tilt angle, and then calculates the local entropy value. After the calculation is completed, the local processor transmits the result, the local entropy value, to the communication module, which is responsible for broadcasting the local entropy value to neighboring units. In response, the neighboring units return their neighborhood entropy values ​​to the communication module. After receiving the neighborhood entropy values, the communication module transmits them to the local processor. Finally, after integrating its own information with that of neighboring units, the local processor executes the final decision-making step, which is to calculate the neighborhood risk accumulation value and generate crop planting recommendations based on this result.

[0040] Example 4: In a specific deployment scenario, to ensure that the internal decision-making model of a microenvironment decision-making unit is optimally adapted to the physical environment before long-term operation, the system is configured to execute an initialization calibration procedure that includes active environmental detection and adaptive parameter calibration. This procedure aims to address how to set a reliable initial value for the decision logic of each unit, particularly the weights related to hydrological response, that aligns with the physical characteristics of its micro-plot, in the absence of prior knowledge. To address this challenge, the initialization procedure first enters the active soil compaction detection stage. In this stage, the local processor instructs its communication module to use a preset, low-frequency pulse sequence... The device drives its built-in vibration motor to generate standardized mechanical vibrations as an active detection signal, which is then transmitted to the soil. Simultaneously, a triaxial accelerometer, acting as a tilt sensor, is switched to a high-frequency vibration sampling mode to collect echo signals reflected from soil layers at different depths. The local processor then executes a defined signal processing procedure. This procedure first applies a bandpass filter to the acquired raw echo signal sequence to remove environmental background noise and the device's own high-frequency electromagnetic interference. Next, a peak detection algorithm is used to identify the peak value of the main reflected echo in the signal, and the time delay of this echo peak value relative to the moment the detection signal was emitted, i.e., the transit time, is precisely calculated. And the attenuation rate of the energy of the echo peak relative to the energy of the initial probe signal. Subsequently, the system uses a preset conversion function to convert the measured data... and Mapped to a dimensionless soil compaction index The specific form of the transformation function is determined during the offline calibration phase. The calibration process involves taking representative soil samples from the target deployment area, preparing soil columns with multiple known compaction levels in the laboratory, and repeating the above active detection and signal processing process for each soil column level, recording a series of corresponding ( The data pairs were finally fitted using the least squares method. and and The stable functional relationship between them can be expressed as a linear model. The coefficients and This is the product of this offline calibration. Thus, each unit gains a quantitative understanding of the physical structure of the soil beneath it, including this compaction index. It was subsequently used as the initial weight related to root stress and water infiltration capacity in the local entropy calculation rule.

[0041] Next, to construct the expected spectral feature model in the sensor self-calibration mechanism and provide a setting procedure for the weight coefficients of the asymmetric adoption logic, the system utilizes the first full-area irrigation or the first effective natural rainfall event to perform a one-time model self-learning and parameter adaptive association. During this process, when the rainfall event is triggered, the local processor of each unit drives its tilt sensor to work in vibration sampling mode, continuously collecting vibration signals caused by rainwater runoff, and analyzing the spectral features of the signal through fast Fourier transform. Given that this is the first rainfall after the system deployment, the system assumes that the deployment tilt angle of each unit is accurate. Therefore, the processor takes the energy centroid frequency of the spectrum calculated at this time as a reference value and stores it as a reference value with respect to the current deployment tilt angle. By binding the expected spectral characteristics, this process completes the online construction of the self-verification model. Furthermore, the system links this process with parameter optimization of the collaborative decision-making logic. Specifically, the adoption ratio or weight coefficient in its asymmetric adoption logic is no longer a fixed value, but a function dynamically related to the physical environment parameters of the local and neighboring units. When downstream unit B receives a weight adjustment announcement from upstream unit A, it adopts the weight coefficient of the upstream water conservation proposal. It will be determined by a base value. Together with multiple correction terms, its correction logic is deterministically expressed as the adoption weight. Soil compaction index between upstream and downstream units The difference is positively correlated, meaning that when the downstream soil is denser than the upstream soil, The value will be increased to respond more proactively to upstream runoff warnings. This move will integrate proactive soil detection, sensor self-calibration model building, and collaborative decision parameter setting into a unified adaptive calibration process executed early in the system's lifecycle.

[0042] Example 5: To ensure the accuracy of crop adaptability evaluation for different crop varieties using the system of the present invention, the crop physiological threshold rules upon which the system is based are constructed and deployed according to a standardized procedure. This procedure begins with the cultivation of the target crop variety in a controlled environment plant growth chamber, and continuous, non-invasive quantitative monitoring of its key growth indicators is performed by a dedicated plant physiological monitoring device. During the experiment, by precisely controlling single environmental variables such as soil moisture and ambient light intensity, the crop undergoes a complete process from optimal growth state to significant stress state, and the continuous changes in environmental parameters and crop physiological indicators, such as leaf water potential and photosynthetic rate, are recorded. By statistically analyzing this dataset, the critical point of the environmental variable that leads to an irreversible statistically significant decrease in the key physiological indicators of the crop is determined. This critical point is defined as a physiological stress threshold for the crop variety. This process is repeated to obtain a complete set of thresholds covering multiple stress types such as drought, waterlogging, and low light. Finally, these verified quantitative threshold data are structured and stored as a rule file associated with a specific crop variety identifier, which can be directly queried and called by the local processor.

[0043] After generation, the rule file is packaged as part of the firmware update package. During the initialization phase of the microenvironment decision unit before it leaves the factory or during field deployment, it is burned into the non-volatile memory of the local processor of each unit through a dedicated interface to complete the loading of the core decision knowledge base. When the system needs to support a new crop variety, it only needs to use the new variety as input and execute the aforementioned offline calibration experimental procedure to generate a corresponding, standardized new rule file. This file can then be remotely distributed and loaded into all microenvironment decision units deployed in the cultivated area via wireless firmware upgrade, thereby expanding the system's crop adaptability evaluation capabilities. This procedure ensures that the fundamental basis of the system's decision logic, namely the crop physiological threshold, is always based on objective and reproducible experimental data, rather than relying on any form of experience setting or fuzzy inference.

[0044] Example 6: To ensure that the core decision model of the system of the present invention has a unified and reproducible initial baseline when deployed in any target area, the system is configured to execute a standardized pre-deployment model calibration procedure. The core of this procedure is to use the local entropy value, which characterizes crop growth risk, to... Defined as a normalized weighted summation model based on multiple stress factors, its deterministic expression is: ,in , and These represent the drought stress, waterlogging stress, and low light stress levels, calculated by combining soil moisture monitoring values ​​and light intensity monitoring values ​​with preset crop physiological thresholds. Their values ​​are normalized to between 0 and 1, while the weighting coefficients... and It is then set to the local deployment tilt angle. The deterministic function, specifically, the drought stress weights Labeled as Inland waterlogging stress weight It is then labeled as ,in and As the benchmark weight, The horizontal slope is the reference, while the coefficient is... and It is uniquely determined through regression analysis of a series of controlled slope infiltration-runoff experiments on typical soil samples in the target area. This procedure unambiguously transforms the static topographic parameter of slope into a dynamic and quantitative decision model weight, thereby eliminating the problem of prediction inconsistency caused by the ambiguity of model parameters.

[0045] To ensure that the collaborative logic among micro-environment decision-making units possesses rigorous physical meaning and reproducible mathematical expression, the system further incorporates a collaborative logic parameterization protocol. This protocol first incorporates the neighborhood risk accumulation value... Neighborhood weight Defined as a deterministic function that integrates spatial distance and hydrological potential energy, its expression is: ,in Representing the local unit and the first The elevation difference between neighboring units is positive only when the neighboring unit is higher than the local unit; otherwise, it is zero. Let be the Euclidean distance between the two, and be the coefficient. This is a calibration constant used to ensure the normalization of the sum of all weights. Furthermore, the protocol algorithmically encapsulates the asymmetric adoption logic in the weight negotiation mode, specifying the adoption coefficient of a downstream unit for the water conservation-related weight recommendations in the announcement when it receives a weight adjustment announcement from an upstream unit. It is determined by a linear model that includes a base trust level and a physical environment correction term, i.e. ,in This is a baseline adoption coefficient calibrated through a network-wide consistency test. This is the environmental sensitivity coefficient, and and These are the soil compaction indices obtained by active acoustic detection from local and neighboring units, respectively. This protocol ensures the determinism and consistency of the decision-making behavior of the entire distributed network in any deployment scenario by transforming abstract collaborative rules into closed mathematical equations driven by physically measurable parameters.

[0046] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0047] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A system for evaluating crop adaptability to arable land and providing planting recommendations, characterized in that, include: Multiple microenvironment decision units are deployed within the cultivated land area. Each microenvironment decision unit includes: an environmental sensor for sensing local soil moisture and ambient light intensity; and a tilt sensor for sensing the deployment tilt angle of the microenvironment decision unit. A local processor is configured to: calculate a local entropy value indicating the level of crop growth risk based on local soil moisture and ambient light intensity sensed by environmental sensors and built-in crop physiological threshold rules; dynamically adjust the weights of environmental factors in the local entropy value calculation rules based on the deployment tilt angle sensed by tilt sensors, wherein when the deployment tilt angle is greater than a first angle threshold, the weights of environmental factors related to drought stress are increased; when the deployment tilt angle is less than a second angle threshold, the weights of environmental factors related to waterlogging stress are increased. The local processor is also configured to: when its own calculated local entropy value does not exceed the first entropy value threshold, but the entropy values ​​of its received neighboring microenvironment decision units are used to calculate the neighborhood risk accumulation value through weighted summation. Exceeding the second entropy threshold At the same time, the risk level of crop planting recommendations is raised, and crop varieties with stronger stress resistance are recommended, including: ,in, Indicates the first The entropy value of each neighboring microenvironment decision-making unit. Indicates the relationship with the first Distance-related weights of neighboring microenvironment decision-making units The number of neighboring micro-environment decision units that receive data from the micro-environment decision unit; A communication module is used to broadcast local entropy values ​​to neighboring microenvironment decision units at set time intervals; and a local processor is further configured to: receive neighborhood entropy values ​​broadcast from neighboring microenvironment decision units; and generate crop planting recommendations based on its own calculated local entropy values ​​and the received neighborhood entropy values, through set collaborative decision rules.

2. The system for evaluating crop adaptability and providing planting recommendations for arable land as described in claim 1, characterized in that, The tilt sensor is a triaxial accelerometer, and the local processor is configured to analyze the triaxial accelerometer readings to determine the deployment tilt angle when the microenvironment decision unit is stationary.

3. The system for evaluating crop adaptability and providing planting recommendations for arable land as described in claim 1, characterized in that, The local processor is also configured to: activate a weight negotiation mode when the increase in local soil moisture detected by the environmental sensor exceeds a first humidity threshold within a set time period; in this mode, the microenvironment decision unit broadcasts a weight adjustment announcement containing its own deployment tilt angle information and the planned adjustment of entropy value calculation weight to neighboring microenvironment decision units through the communication module; and the local processor is also configured to: after receiving the weight adjustment announcement from a neighboring microenvironment decision unit, execute asymmetric adoption logic to correct its own entropy value calculation weight based on the difference between its own deployment tilt angle and the deployment tilt angle of the neighboring microenvironment decision unit.

4. The system for evaluating crop adaptability and providing planting recommendations for arable land as described in claim 3, characterized in that, The asymmetric adoption logic executed by the local processor includes: when the deployment tilt angle of the micro-environment decision unit itself is less than the deployment tilt angle of the neighboring micro-environment decision units, the micro-environment decision unit adopts the suggestions of the neighboring micro-environment decision units regarding the adjustment of the water conservation weight by a higher proportion than the proportion of its own adjustment of the drainage weight.

5. The system for evaluating crop adaptability and providing planting recommendations for arable land as described in claim 1, characterized in that, The environmental sensor also includes a light sensor. The local processor is configured to: switch the light sensor to a high-frequency sampling mode at night to monitor fluctuations in ambient light intensity; and exchange timestamps of the light intensity fluctuation characteristics monitored by each other with the neighboring micro-environment decision unit through the communication module to collaboratively calculate the vector information of the speed and direction of cloud movement; when the speed of the vector information exceeds a first speed threshold and the direction deviation is less than a third angle threshold, trigger the micro-environment decision unit to enter the early warning working state.

6. The system for evaluating crop adaptability and providing planting recommendations for arable land as described in claim 5, characterized in that, When the local processor is in alert mode, it increases the frequency of local entropy calculation and local entropy broadcast.

7. The system for evaluating crop adaptability to arable land and providing planting recommendations as described in claim 3, characterized in that, The local processor is also configured to: when a rainfall event is detected, operate its tilt sensor in vibration sampling mode to acquire vibration signals caused by rainwater runoff; evaluate the reliability of statically sensing the deployment tilt angle by analyzing the spectral characteristics of the vibration signal and comparing them with the expected spectral characteristics corresponding to the deployment tilt angle sensed by the tilt sensor when stationary; and adjust its own weight to adjust the priority of the declaration based on the reliability evaluation results.

8. The system for evaluating crop adaptability and providing planting recommendations for arable land as described in claim 1, characterized in that, The communication module of the microenvironment decision unit is also configured to: drive the communication module to generate a standardized mechanical vibration as an active detection signal and transmit it to the soil; and the local processor is also configured to: use an inclination sensor to collect the echo signal returned after the detection signal is reflected by the soil; determine the soil compaction information by analyzing the transit time and energy attenuation rate of the echo signal; and use the compaction information to correct the weights in its local entropy calculation rules.

Images