A device and method for precise application of liquid manure and synergistic soil improvement through deep tillage

Abstract

This invention relates to the field of agricultural machinery and soil improvement technology, specifically a device and method for precise application of liquid manure and deep tillage in conjunction with soil improvement. The device includes sensing (online NIR, TDR type soil moisture sensors, etc.), execution (tractor automatic driving, hydraulic depth control, etc.), control, and communication modules. The method uses three interconnected factors—manure characteristics-travel speed, soil moisture-deep tillage depth-fertilizer application rate, and historical yield-real-time growth—combined with normalized control to dynamically adjust operating parameters. This invention's device and method for precise application of liquid manure and deep tillage in conjunction with soil improvement can control the nutrient deviation per unit area within ±5%, increase crop yield by 10%-20%, increase nutrient utilization rate by 10%-15% and reduce waste by more than 25%, and extend equipment life by about 20%. It is suitable for precise deep application of manure and soil improvement operations for field crops such as corn and wheat.

Description

Technical Field

[0001] This invention relates to the field of agricultural machinery and soil improvement technology, specifically a device and method for precise application of liquid manure and deep loosening in conjunction with soil improvement. Background Technology

[0002] In current agricultural production, the application of liquid manure is often carried out in conjunction with deep tillage to improve soil conditions. This is typically achieved using a tractor-mounted system consisting of a fertilization mechanism and a deep tillage mechanism. Such systems are widely used in field crops (such as corn and wheat). Their core objective is to supplement soil nutrients through manure application, while deep tillage breaks up compacted layers and improves soil aeration and water retention. However, existing systems often remain at the level of single-parameter control or simple combination adjustments. For example, they may control the amount of fertilizer based on a preset fixed travel speed or set the deep tillage depth based on static soil fertility data. This fails to achieve deep coupling and dynamic response between manure characteristics, real-time soil conditions, crop growth, and historical production data. The overall operation mode remains primarily "experience-based" and "fixed," making it difficult to adapt to the complex and ever-changing environmental conditions and crop growth needs in the field.

[0003] Existing devices and methods for precise application of liquid manure and deep tillage in conjunction with soil improvement fail to establish a multi-dimensional linkage decision-making mechanism. This results in low nutrient supply precision, unstable fertilizer efficiency, high equipment wear and tear, and unreasonable spatiotemporal fertilization strategies during operation. Overall, they cannot meet the needs of agricultural production for precision, intelligence, and sustainability. Therefore, in view of the above situation, there is an urgent need to develop a device and method for precise application of liquid manure and deep tillage in conjunction with soil improvement to overcome the shortcomings in current practical applications. Summary of the Invention

[0004] The purpose of this invention is to provide a device and method for precise application of liquid manure and deep loosening in conjunction with soil improvement, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for precise application of liquid manure and synergistic soil improvement through deep tillage includes the following steps:

[0007] S1. Initialization phase: Load the variety parameters, growth period parameters and historical yield data of the target crop and the target plot, complete the self-check of the sensing module, and ensure that the sensor is in normal working condition.

[0008] S2, Travel and Sensing Stage: During the tractor's travel, the sensing module collects real-time data on manure characteristics, soil moisture, crop growth, deep tillage depth, and manure flow rate.

[0009] S3, Linkage Calculation Stage: Based on the real-time data collected in step S2, calculate the linkage parameters of manure characteristics-traveling speed, soil moisture-deep loosening depth-fertilizer application rate, and historical yield-real-time growth status respectively. Then, combine the normalized control model to calculate the final fertilizer application rate, target travel speed, and target deep loosening depth.

[0010] S4. Execution and Feedback Phase: The execution module adjusts the travel speed, deep tillage depth and fertilizer application amount according to the target parameters obtained in step S3. At the same time, the sensing module provides real-time feedback of actual operating data to dynamically correct deviations and ensure operational accuracy.

[0011] S5. Post-operation review stage: Generate operation data reports and operation parameter distribution charts, optimize the linkage calculation parameters in the control module based on the operation data, and provide data support for subsequent operations.

[0012] As a further aspect of the present invention: in step S2, the collected real-time data is preprocessed:

[0013] Outlier removal and smoothing are performed on the manure characteristic data collected by online near-infrared sensors to reduce the impact of environmental interference on data accuracy.

[0014] The soil moisture data collected by the TDR soil moisture sensor is corrected according to the soil texture of the target plot to improve the reliability of the soil moisture data.

[0015] As a further aspect of the present invention: In step S3, the calculation of the manure characteristics-traveling speed linkage parameter is achieved by constructing a comprehensive manure application index, weighting and integrating the manure viscosity and NPK nutrient concentration, and then mapping the index to the target traveling speed to achieve the matching of manure characteristics and traveling speed.

[0016] The calculation of the linkage parameters of soil moisture, deep loosening depth and fertilizer application rate is to first set the target deep loosening depth based on soil moisture, and then fine-tune the fertilizer application rate according to the actual deep loosening depth to ensure uniform nutrient density when the soil disturbance volume changes.

[0017] The calculation of the historical yield-real-time crop growth linkage parameter involves first classifying plot yield levels based on historical yields and setting corresponding benchmark fertilization coefficients, and then combining this with real-time crop growth to calculate dynamic correction factors, thereby enabling potential discovery and risk warning.

[0018] As a further aspect of the present invention: In step S4, the deviation correction adopts a PID control algorithm to perform closed-loop adjustment of the travel speed, deep loosening depth and fertilizer application amount, so that the deviation between the actual travel speed, actual deep loosening depth and actual fertilizer application amount and the corresponding target parameters is controlled within a preset range, thereby ensuring the stability and accuracy of the operation parameters.

[0019] As a further aspect of the present invention: in step S5, the operation data report includes at least the total operation area, total fertilizer application, average fertilizer application, average deep loosening depth, and average travel speed.

[0020] The operation parameter distribution map includes at least a deep tillage depth distribution map and a fertilizer application rate distribution map;

[0021] The operation results are presented intuitively through operation data reports and distribution charts, and the optimized linkage calculation parameters are stored in the data storage unit of the control module to improve the accuracy of subsequent similar operations.

[0022] A device for precise application of liquid manure and deep tillage combined with soil improvement, which implements the above-described method of precise application of liquid manure and soil improvement, includes a sensing module, an execution module, a control module and a communication module.

[0023] The sensing module is used to collect real-time data on manure characteristics, soil moisture, crop growth, deep tillage depth, and manure flow rate. It includes at least an online near-infrared sensor, a TDR-type soil moisture sensor, a crop growth detection sensor, a depth detection sensor, and a flow rate sensor.

[0024] The execution module is used to perform travel speed adjustment, deep tillage depth adjustment and fertilizer application rate adjustment, and includes at least a tractor speed control system, a hydraulic depth control system and a variable frequency fertilizer pump.

[0025] The control module is connected to the sensing module and the execution module through a communication module. It has a built-in data storage unit for storing crop variety parameters, crop growth period parameters, historical yield data of the plot, and sensor calibration parameters. Based on the data collected by the sensing module, it can perform linkage calculations of manure characteristics-traveling speed, soil moisture-deep loosening depth-fertilizer application rate, historical yield-real-time growth, and output control commands to the execution module.

[0026] The communication module is used to realize real-time data interaction between the sensing module, control module, and execution module, ensuring that the modules work together.

[0027] As a further aspect of the present invention: the crop growth detection sensor is an airborne NDVI sensor or a soil fertility sensor;

[0028] When the operation scenario is before sowing or during the seedling stage, the crop growth detection sensor is switched to a soil fertility sensor to collect real-time soil nutrient data to replace crop growth data and participate in the linkage calculation of historical yield and real-time growth.

[0029] As a further aspect of the present invention: the TDR-type soil moisture sensor is installed in front of the subsoiling plow to collect soil moisture data in advance, so as to avoid the interference of soil structure changes after subsoiling operations on the moisture detection results.

[0030] The depth detection sensor is installed on the side of the subsoiler to monitor the actual depth of the subsoiler in real time.

[0031] The flow sensor is installed at the outlet of the fertilizer pipeline to provide real-time feedback on the actual flow rate of the manure.

[0032] As a further aspect of the present invention: the hydraulic depth control system includes a hydraulic cylinder, a servo valve, and a displacement sensor;

[0033] The displacement sensor is used to detect the extension and retraction of the hydraulic cylinder. In conjunction with the commands output by the control module, it realizes closed-loop control of the deep tillage depth and ensures the accuracy of deep tillage depth adjustment.

[0034] The tractor speed control system changes the tractor's travel speed by adjusting the speed of the hydraulic motor, adapting to changes in the characteristics of manure.

[0035] As a further aspect of the present invention: the control module adopts an industrial-grade MCU, which has a multi-channel analog input interface and a PWM output interface, and can simultaneously process real-time data from multiple sensors and quickly output control commands;

[0036] The data storage unit adopts a dual storage design, which can store historical operation data, sensor calibration parameters and crop-related parameters for a long time, and supports data backtracking and operation review.

[0037] Compared with the prior art, the beneficial effects of the present invention are:

[0038] 1. By linking the characteristics of manure with the mechanical travel speed, the deviation of nutrient supply per unit area is controlled within ±5%, which is far superior to ±15% of traditional devices; at the same time, when operating with high-viscosity manure, the working pressure of the fertilizer pump can be reduced by 10% to 15%, extending the service life of the equipment by about 20%.

[0039] 2. By linking soil moisture, deep loosening depth, and fertilizer application rate, in plots with a soil moisture variation coefficient > 20%, the crop yield variation coefficient decreased from 15% to 8%, nutrient utilization increased by 10%–15%, nutrient loss was reduced by 20% under excessively wet conditions, and nutrient ineffectiveness was reduced by 30% under excessively dry conditions.

[0040] 4. By linking historical yield with real-time growth, the yield of historically low-yield plots can be increased by 10% to 20%, and the nutrient waste of historically high-yield plots can be reduced by more than 25%, thus avoiding the underutilization of potential in low-yield plots and the fertilization risks (such as excessive growth and lodging) in high-yield plots.

[0041] 5. Through normalized control, deep fusion of multi-source data is achieved, which can be adapted to a variety of field crops such as corn, wheat and rice. In the overall operation, crop yield is increased by 10% to 20% while nutrient waste is reduced by more than 25%, which has significant economic benefits (reduced fertilizer costs and increased yield) and ecological benefits (reduced nutrient loss and pollution). Attached Figure Description

[0042] Figure 1 This is a flowchart illustrating the process of the method for precise application of liquid manure and deep loosening combined with soil improvement in an embodiment of the present invention.

[0043] Figure 2 This is a block diagram showing the overall module connection of the liquid manure precision application and deep tillage synergistic soil improvement device in an embodiment of the present invention.

[0044] Figure 3 This is a soil potential fertilization baseline map in an embodiment of the present invention.

[0045] Figure 4 This is a deep tillage depth distribution map after the operation in an embodiment of the present invention.

[0046] Figure 5 This is a distribution diagram of the amount of fertilizer applied after the operation in an embodiment of the present invention. Detailed Implementation

[0047] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0048] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0049] Please see Figures 1-5 This invention provides a device and method for precise application of liquid manure and deep tillage in conjunction with soil improvement. By constructing three core linkage effects—manure characteristics-mechanical parameters, soil moisture-fertilization depth-fertilizer amount, and historical data-real-time crop growth—it achieves deep coupling decision-making and dynamic feedback adjustment of multi-source sensing information. It can adapt mechanical operation parameters in real time according to the physicochemical properties of manure, dynamically adjust fertilization depth and fertilizer amount according to soil moisture, and dialectically optimize fertilization strategies by combining historical yield and real-time crop growth. Ultimately, it achieves the synergistic goals of precise manure application, efficient utilization of soil nutrients, and optimized crop growth, significantly improving the intelligence level and sustainability of agricultural production, and effectively solving the pain points of traditional fertilization such as nutrient waste, unstable fertilizer efficiency, and high equipment wear and tear.

[0050] 1. Overall System Composition

[0051] This invention relies on the collaborative work of hardware actuators and software control modules to ensure the real-time implementation and precise execution of various linkage effects. Its core components can be divided into four categories: sensing module, execution module, control module, and communication module. The core components and functional division of each module are as follows:

[0052] (1) The sensing module includes an online near-infrared (NIR) sensor, a TDR soil moisture sensor, an airborne NDVI (normalized vegetation index) sensor, a laser rangefinder sensor, and a flow sensor.

[0053] Among them, the online near-infrared (NIR) sensor collects manure samples at a frequency of 10Hz and outputs manure viscosity (unit: mPa•s) and NPK nutrient concentration (based on N element content, range: 0.5%~5%) in real time through spectral analysis algorithm, providing core data for subsequent manure load assessment;

[0054] The TDR type soil moisture sensor is installed in front of the subtilist plow to collect soil moisture (volume moisture content, unit: %) in advance, avoiding the interference of soil structure changes after subtilist operation on moisture detection, and providing accurate soil moisture basis for depth adjustment;

[0055] The onboard NDVI sensor is installed on the top of the tractor and acquires the crop canopy growth (NDVI value, range: -1 to 1) in real time at a frequency of 2Hz. The vegetation index reflects the crop's photosynthetic capacity and nutritional status. In addition, if the operation scenario is before sowing or during the crop seedling stage (without an effective canopy), the soil fertility sensor (such as a soil nitrogen sensor) is switched to collect real-time soil nutrient data to replace the NDVI value in the historical-real-time linkage decision-making.

[0056] A laser rangefinder is installed on the side of the subsoiler to monitor the actual depth of the subsoiler in the soil in real time (unit: cm), ensuring the closed-loop control accuracy of the depth adjustment.

[0057] A flow sensor is installed at the outlet of the fertilizer pipeline to provide real-time feedback on the actual flow rate of manure (unit: L / min), providing real-time data support for the dynamic correction of fertilizer application.

[0058] (2) The execution module mainly includes the tractor automatic driving / hydraulic transmission system, hydraulic depth control system and variable frequency fertilizer pump;

[0059] The tractor's automatic driving / hydraulic transmission system can receive speed adjustment commands sent by the controller and change the tractor's travel speed (unit: km / h) by adjusting the speed of the hydraulic motor, thereby matching the working speed with the manure load;

[0060] The hydraulic depth control system consists of hydraulic cylinders, servo valves and displacement sensors. It can drive the subsoil plow to rise and fall according to the target depth command, accurately control the soil penetration depth, and ensure that the depth adjustment deviation is controlled within ±1cm.

[0061] The variable frequency fertilizer pump is installed between the manure storage tank and the fertilizer pipeline. It can adjust the pump speed according to the frequency command output by the controller, thereby changing the manure output and achieving precise control of the fertilizer application. The flow rate adjustment range covers 10-50L / min, which can adapt to the fertilization needs of different crops and different growth stages.

[0062] (3) The control module uses an industrial-grade MCU (such as the STM32H7 series) and a data storage unit;

[0063] The industrial-grade MCU integrates multi-sensor data acquisition, mathematical model calculation, and instruction execution output functions. It has 16 analog input interfaces and 8 PWM output interfaces, and can process real-time data from more than 5 sensors at the same time. The calculation response time is ≤100ms, ensuring the real-time performance of the linkage effect.

[0064] The data storage unit adopts a dual storage design of SD card and Flash, which can store the historical yield data of the target plot for nearly 3 years (unit: kg / hectare), variety parameters of different crops (such as benchmark fertilization amount and optimal soil moisture range), and sensor calibration parameters. The storage capacity is ≥16GB, which meets the needs of recording and backtracking long-term operation data.

[0065] (4) The communication module adopts CAN bus communication to realize real-time data interaction between sensors, controllers and actuators. The transmission rate is ≥250kbps and the communication delay is ≤50ms, ensuring the synchronization of multi-module collaborative work and avoiding control deviation caused by data transmission delay.

[0066] 2. Specific Implementation Steps

[0067] 2.1 The linkage between manure characteristics and mechanical travel speed;

[0068] By transforming the physicochemical properties of manure, such as viscosity and nutrient concentration, from interfering factors that need to be compensated for in traditional control into actively adjustable control parameters, and by constructing a comprehensive manure application index to quantify the manure application load, and then mapping this index to tractor travel speed, a matching logic of higher load and lower speed is achieved. This not only ensures a constant total amount of nutrients per unit area of ​​land (i.e., equal nutrient fertilization), but also reduces the working pressure on equipment under high load conditions and extends the service life of machinery. The specific implementation steps are as follows:

[0069] 2.1.1 Real-time sensing: Collection and preprocessing of manure characteristic parameters;

[0070] First, an online NIR sensor continuously collects manure samples at a frequency of 10 Hz, and the real-time manure viscosity (μ) and NPK concentration (C) are calculated using a spectral analysis algorithm. Due to interference from vibrations and temperature fluctuations in the field environment, the raw data needs to be preprocessed.

[0071] The first step is to remove outliers by discarding data that exceeds the sensor's range (viscosity 0–1000 mPa•s, concentration 0–10%) to avoid interference from extreme values ​​in subsequent calculations.

[0072] The second step is moving average filtering, which uses a moving average algorithm with a window size of 5 to smooth the effective data, obtaining the smoothed average viscosity (μ). avg ) and average concentration (C avg This step can reduce data fluctuations by more than 30%, ensuring the stability of manure characteristic parameters.

[0073] 2.1.2 Dynamic Decision-Making: Calculation of the Comprehensive Manure Application Index;

[0074] To comprehensively evaluate the load during manure application, a comprehensive manure application index (K) was constructed. fer The model, which uses a weighted fusion of viscosity and concentration parameters, transforms physical quantities of different dimensions into dimensionless exponents, as shown in the following formula:

[0075]

[0076] Where, μ std The standard viscosity of manure needs to be preset based on crop requirements and equipment capacity (e.g., μ when fertilizing corn). std =300 mPa•s), this value was determined through field calibration tests. Three different types of manure with varying viscosities were selected for operational testing, and the viscosity when the fertilizer pipeline was unblocked and the manure was evenly dispersed was used as the standard value; C tar The target NPK concentration is set according to the crop growth stage (e.g., C10 at the jointing stage of wheat). tar =2.0%), refer to the fertilization guidelines issued by the local agricultural technology extension department; w1 and w2 are weighting coefficients, satisfying w1+w2=1, calibrated through orthogonal experiments (usually w1=0.4, w2=0.6), giving priority to ensuring the influence of nutrient concentration on fertilization load, and avoiding insufficient or excessive nutrient supply due to concentration deviation.

[0077] K fer The value range of K is [0.5, 2.0]. fer When K > 1, it indicates that the viscosity or concentration of the manure is too high, and the applied load is greater than the standard load. In this case, the walking speed needs to be reduced to extend the fertilization time per unit area; when K ferWhen K < 1, it indicates that the current manure application load is lower than the standard load, and the travel speed can be appropriately increased to improve work efficiency; when K fer When μ = 1, the characteristics of the manure match the standard conditions, and the baseline operating speed can be maintained. For example, if μ avg =350mPa•s, C avg =2.5%, then K fer =0.4×(350 / 300)+0.6×(2.5 / 2.0)=1.28, indicating that the current manure load is higher than the standard and speed adjustment needs to be activated.

[0078] 2.1.3 Linked Execution: Dynamic Adjustment and Closed-Loop Control of Travel Speed;

[0079] First, set the baseline travel speed (v). base This value is preset based on crop row spacing and operational efficiency requirements (e.g., when the corn row spacing is 60cm, v). base =6km / h), ensuring that under standard load, it can meet the daily operating area requirement (about 15 hectares / day) and ensure that the manure is fully dispersed.

[0080] Subsequently, a speed regulation model was constructed, and K was used. fer Mapped to target travel speed (v) tar The formula is as follows:

[0081]

[0082] Where k is the speed adjustment coefficient, calibrated through field trials (usually k=0.3). The setting of this coefficient needs to balance adjustment sensitivity and system stability. An excessively large k value can easily lead to frequent speed fluctuations, while an excessively small k value will fail to respond promptly to load changes. For example, when K... fer When v = 1.28, tar =6×exp(-0.3×0.28)=5.47km / h, which is about 8.8% lower than the baseline speed. At this time, the fertilization time per unit area increases from 36 seconds / hectare at the baseline speed to 39.4 seconds / hectare. This ensures that the high viscosity manure has enough time to disperse in the deep loosening trench, avoids accumulation and clumping, and improves the soil's adsorption effect on nutrients.

[0083] Finally, the controller transmits v via the CAN bus. tar The command is sent to the tractor's automatic driving system, which uses a PID algorithm to adjust the speed of the hydraulic transmission mechanism, while simultaneously collecting the actual travel speed (v) in real time through wheel speed sensors. cur This forms a speed closed-loop control. The proportional coefficient K in the PID algorithm... p =0.8, integral coefficient K i =0.2, differential coefficient K d=0.1, and this algorithm can be used to convert v cur With v tar The deviation is controlled within ±0.2km / h to ensure the accuracy of speed adjustment and avoid uneven fertilization caused by speed fluctuations.

[0084] Through the above-mentioned linkage effect, on the one hand, the accuracy of nutrient application can be improved. By coordinating the adjustment of "concentration × flow rate × speed", the deviation of nutrient supply per unit area can be controlled within ±5%, which is far better than the ±15% deviation of traditional fixed-speed fertilization. On the other hand, it can protect the equipment. When operating high-viscosity manure, reducing the speed can reduce the working pressure of the fertilization pump by 10% to 15%, reduce pump wear, and extend its service life by about 20%.

[0085] 2.2 Three-level linkage between soil moisture, deep loosening depth, and fertilizer application rate;

[0086] In traditional fertilization control, soil moisture is only used to simply adjust the amount of fertilizer applied (e.g., reducing application when humidity is high), without considering the impact of moisture on fertilization depth. This results in manure being placed in unsuitable soil layers (e.g., deep application in dry soil leads to nutrient retention and ineffectiveness). This invention proposes a three-level linkage logic: moisture-driven depth and depth-feed feedback. First, the deep loosening depth is dynamically set based on moisture, and then the fertilizer application amount is fine-tuned based on the actual depth, ensuring that manure is always placed in the soil layer with the highest nutrient absorption efficiency. The specific implementation steps are as follows:

[0087] 2.2.1 Real-time sensing: Soil moisture data collection and correction;

[0088] A TDR-type soil moisture sensor is pre-installed 1.5m in front of the deep tillage plow, collecting soil moisture (θ, volumetric water content) at a frequency of 5Hz. Because different soil textures (such as sandy soil and clay soil) affect the TDR sensor's detection signal, the original soil moisture value needs to be corrected for soil texture. A correction coefficient table is established by measuring the texture parameters of soil samples from the target plot in the laboratory. The correction coefficient for sandy soil is 0.9 (because sandy soil has poor water retention, the original detection value is easily too high), the correction coefficient for clay soil is 1.1 (because clay soil has good water retention, the original detection value is easily too low), and the correction coefficient for loam is 1.0. The original soil moisture value is multiplied by the corresponding correction coefficient to obtain the corrected soil moisture value (θ). cal This step can reduce the soil moisture detection error from ±3% to ±1%, providing an accurate basis for depth adjustment.

[0089] 2.2.2 Dynamic Decision 1: Setting the Target Deep Tillage Depth;

[0090] Construct a soil moisture-target depth mapping model, based on θ cal Determine the target depth (h) of the subsoiling plow tar The formula is as follows:

[0091]

[0092] Among them, h base The baseline deep tillage depth is preset based on crop root characteristics (e.g., corn has deeper roots, h). base =25cm; wheat roots are relatively shallow, h base =20cm); θ opt To determine the optimal soil moisture level for the crop, set the appropriate level based on the crop variety (e.g., the optimal soil moisture level for corn is 18%–25%, with 21.5% as the midpoint). θ max Soil saturation moisture (θ for sandy soil) max =30%, clay soil θ max =35%); Δh is the depth adjustment range, preset to ±5cm (i.e., h tar The range of values ​​for is [h base -5cm,h base [+5cm]), to avoid nutrient evaporation due to being too shallow and root absorption due to being too deep.

[0093] When θ cal >θ opt When the soil is excessively wet, the deeper layers of the soil have more moisture, which will increase the humidity. tar Deepening the soil allows manure to be placed in deeper, moist soil, reducing nutrient loss due to surface runoff and promoting deeper root development in crops to enhance lodging resistance; when θ cal <θ opt When the soil is too dry, the shallow soil layer has relatively sufficient moisture, which will cause h tar Shallowing the soil allows manure to be placed in the main active layer of the root system, preventing it from being absorbed by the roots if it is placed in dry soil. For example, when working with corn, if θ cal =26% (too humid), then h tar =25+5×(26-21.5) / (35-21.5)=26.67cm, by increasing the depth, the manure is introduced into the deep moist soil; if θ cal =16% (too dry), then h tar =25+5×(16-21.5) / (35-21.5)=22.94cm, by adjusting the depth to ensure that the manure is close to the root system.

[0094] 2.2.3 Real-time control of deep tillage depth;

[0095] Laser rangefinders collect real-time data on the actual penetration depth (h) of the subsoil plow. cur The system then feeds the data back to the hydraulic depth control system. The system uses an incremental PID algorithm to adjust the extension and retraction of the hydraulic cylinder, with the control law as follows:

[0096]

[0097] Where Δu is the displacement adjustment amount of the hydraulic cylinder; K p =0.8, K i =0.2、K d =0.1 is the PID parameter, calibrated through a step response test. When a step depth command is input to the system (e.g., increasing from 25cm to 27cm), the PID parameter is adjusted to ensure the system has no overshoot and a response time ≤1 second; (h tar -h cur ) t d(h) represents the depth deviation at time t. tar -h cur ) / dt represents the rate of change of depth deviation. Using this algorithm, h can be... cur with h tar The deviation is controlled within ±1cm to ensure the accuracy of depth adjustment and avoid differences in fertilizer effect caused by depth deviation.

[0098] 2.2.4 Dynamic fine-tuning of fertilizer application rate;

[0099] Changing the depth of deep tillage leads to changes in soil disturbance volume (the deeper the soil, the larger the disturbance volume). If the fertilizer application rate remains constant, problems may arise such as insufficient nutrient density when applied deeply and excessively high nutrient density when applied shallowly. Therefore, based on the actual depth h... cur A depth-fertilizer application correction model is constructed to fine-tune the amount of manure applied (Q), as shown in the following formula:

[0100]

[0101] Among them, Q base The baseline fertilizer application rate is set according to the crop's growth stage (e.g., Q is the jointing stage of corn). base =30m 3 / hectare, wheat greening period Q base =20m 3 / hectare); α is the depth-fertilizer correction coefficient, calibrated through field fertilizer efficiency trials (usually α=0.3), selecting 3 different depths (h base -5cm, h base h base A fertilization experiment was conducted at a depth of +5cm to determine crop yield and nutrient utilization rate. It was found that when α=0.3, the difference in nutrient utilization rate at different depths was the smallest (≤5%).

[0102] When h cur h base As the depth increases, the volume of soil disturbance increases, and Q needs to be slightly increased to infiltrate the expanded space and ensure uniform nutrient density; when h cur <h baseWhen the soil depth is reduced, the volume of soil disturbance decreases, and Q needs to be slightly reduced to avoid nutrient overload. For example, during corn operations, if h cur =30cm (relative to h) base If the depth is increased by 5cm, then:

[0103] Q=30×(1+0.3×(30-25) / 25)=31.8m 3 / hectare;

[0104] By increasing fertilizer application by 6%, sufficient nutrient supply to the deeper soil layers is ensured; if h cur =20cm (relative to h) base (Adjust to a shallower depth of 5cm), then:

[0105] Q = 30 × (1 + 0.3 × (20 - 25) / 25) = 28.2 m³ / hectare;

[0106] By reducing fertilizer application by 6%, excessive vegetative growth in shallow soil can be avoided.

[0107] In areas with large fluctuations in soil moisture (such as soil moisture variation coefficient > 20%), after adopting this linkage control, the crop yield variation coefficient is reduced from 15% to 8% under traditional fixed-depth fertilization, nutrient utilization is increased by 10% to 15%, and nutrient loss due to unsuitable soil moisture is reduced (loss is reduced by 20% when the soil is too wet and ineffective amount is reduced by 30% when the soil is too dry).

[0108] 2.3 Spatiotemporal linkage between historical yield and real-time crop growth;

[0109] Traditional fertilization methods often rely on static prescription maps generated from historical yield data or make simple adjustments based solely on real-time crop growth, failing to consider the dialectical relationship between historical and real-time data. This leads to the problem of consistently low fertilization in low-yield areas and indiscriminately high fertilization in high-yield areas. This invention constructs a dynamic correction factor by comparing historical yields and real-time crop growth, achieving a dual function of potential exploration and risk warning. It increases fertilization for historically low-yield areas but with good real-time crop growth (potential exploration) and reduces fertilization for historically high-yield areas but with poor real-time crop growth (risk warning). The specific implementation steps are as follows:

[0110] 2.3.1 Baseline Establishment: Generation of Soil Potential Fertilization Map;

[0111] The first step is to load the historical yield data (Y) of the target plot for the past three years from the data storage unit. hist (Unit: kg / hectare) These data can be obtained through agricultural IoT platforms or recorded manually (e.g., data from harvester yield monitoring systems). Because historical yield data exhibits annual fluctuations, data smoothing is necessary: ​​calculate the average yield over the past 3 years (Y). avg ), comparing annual output with Y avgData with a deviation exceeding 10% are discarded, and Y is recalculated. avg This yields the revised historical average yield.

[0112] The second step is to adjust the Y... avg The target plots were divided into three yield levels: low-yield level (Y). avg <6000 kg / ha), medium-yield grade (6000≤Y) avg ≤8000 kg / ha), high-yield grade (Y avg >8000 kg / ha). A corresponding benchmark fertilization coefficient (F) is set for different grades. base Low-yield grade F base =0.8 (Initial application should be low to avoid wasting nutrients due to blind investment); Mid-level F base =1.0 (Regular application, maintaining stable yield); High-yield grade F base =1.1 (Appropriately apply higher amounts to maintain soil fertility and ensure high yield).

[0113] The third step involves combining the latitude and longitude information of the land parcel (obtained via the tractor's GPS module) with the Kriging interpolation method in ArcGIS software to calculate the F coordinates of each sampling point. base The interpolation value is a gridded soil potential fertilization baseline map, with the grid size set to 10m×10m (matching the detection resolution of the NDVI sensor) to ensure that each grid has a corresponding baseline fertilization coefficient, providing a baseline for subsequent real-time correction.

[0114] 2.3.2 Real-time correction: Crop growth data collection and comparison;

[0115] The airborne NDVI sensor collects the NDVI values ​​of the crop canopy at a frequency of 2 Hz. cur This value reflects the photosynthetic capacity of crops, NDVI cur The larger the value, the better the crop growth and the higher the chlorophyll content; NDVI cur The lower the NDVI value, the worse the crop growth, potentially indicating nutrient deficiency or disease. Since NDVI values ​​are significantly affected by the crop's growth stage (e.g., low NDVI values ​​during the seedling stage and high NDVI values ​​during the jointing stage), normalization based on the growth stage is necessary: ​​First, query the data storage unit for the optimal NDVI value for the corresponding growth stage of the crop (NDVI). opt For example, NDVI during the corn jointing stage opt =0.6, NDVI at wheat heading stage opt =0.7); then, calculate the Normalized Growth Index (NDVI). norm =NDVI cur / NDVI opt ), NDVI normThe value range is [0.5, 1.5], where NDVI norm >1.0 indicates that the plant is growing better than its best performance at the same time of year. NDVI norm <1.0 indicates that the growth is inferior to the best condition at the same time.

[0116] Simultaneously, extract the historical production level of the current work grid and load the historical average NDVI (NDVI) for that level. hist For example, the historical NDVI of low-yield grades hist =0.5, historical NDVI for high-yield grades hist =0.65). Calculate the growth deviation coefficient between real-time and historical data:

[0117] ΔNDVI=(NDVI cur -NDVI hist ) / NDVI hist );

[0118] ΔNDVI>0 indicates that the current growth is better than the same period in history, while ΔNDVI<0 indicates that the current growth is worse than the same period in history.

[0119] 2.3.3 Linked Execution: Dynamic Correction Factor Calculation and Fertilization Adjustment;

[0120] A historical-real-time dynamic correction factor (F3) model is constructed. Based on the combination of historical yield levels and ΔNDVI, the value of F3 is determined, thereby correcting the final fertilizer application rate. The core logic of this model is to dialectically handle the relationship between historical potential and real-time status. Specific scenarios and parameter settings are as follows:

[0121] Scenario 1: Historical High Yield Level (Y) avg >8000 kg / ha) + ΔNDVI ∈ [-0.1, 0.1] (normal growth). At this point, the soil fertility is good and the crop growth is stable, so there is no need to significantly adjust the amount of fertilizer applied. The value of F3 is 0.9 to 1.0. For example, when ΔNDVI = 0, F3 = 0.95. By slightly reducing the amount of fertilizer, we can avoid crop lodging or disease caused by excessive nutrients, while maintaining soil fertility.

[0122] Scenario 2: Historically high yield level + ΔNDVI < -0.1 (poor growth). In this case, there may be historically unknown problems (such as soil compaction, latent diseases, nutrient imbalance). Blindly applying fertilizer will exacerbate these problems or cause waste. An early warning should be issued and fertilization significantly reduced. F3 should be set between 0.2 and 0.5. For example, if ΔNDVI = -0.2 (significantly worse growth than historical levels), F3 = 0.3. Simultaneously, the controller will use audible and visual alarms to alert operators to investigate the problem area and prevent further losses due to improper fertilization.

[0123] Scenario 3: Historically Low Productivity Level (Y) avg<6000 kg / ha) + ΔNDVI ∈ [-0.1, 0.1] (normal growth). At this point, the soil potential has not been fully utilized, and a low-application strategy should be maintained to observe changes in growth. The F3 value should be 0.8–0.9. For example, ΔNDVI = 0.05 (slightly better growth), F3 = 0.85, to avoid premature application leading to nutrient accumulation.

[0124] Scenario 4: Historically low yield level + ΔNDVI > 0.1 (excellent growth). In this case, soil potential is activated (possibly due to favorable weather and proper early management), requiring significantly increased fertilizer application to support the crop in breaking historical low yield records. The F3 value is 1.2–1.5. For example, ΔNDVI = 0.2 (significantly better growth than historical levels), F3 = 1.4. By increasing the application of 140% of the baseline fertilizer amount, sufficient nutrients are provided to the crop, contributing to increased yield.

[0125] Scenario 5: Historical Middle Class (6000≤Y) avg ≤8000 kg / ha) + ΔNDVI ∈ [-0.2, 0.2] (any growth vigor). At this point, soil fertility is moderate, and fine-tuning is needed according to growth vigor. F3 is taken as 0.9 to 1.1. For example, ΔNDVI = 0.15 (excellent growth), F3 = 1.05; ΔNDVI = -0.15 (poor growth), F3 = 0.95. Stable yield is maintained through small adjustments.

[0126] Through the aforementioned synergistic effect, the limitations of traditional static prescriptions are overcome. This approach not only allows for assessment of soil potential based on historical data but also enables dynamic adjustments based on real-time plant growth. This approach taps into the potential of low-yield plots while mitigating fertilization risks in high-yield plots. In practical applications, this synergy can increase yields in low-yield plots by 10%–20% and reduce nutrient waste in high-yield plots by more than 25%.

[0127] 2.4 Normalized control: Multi-factor collaborative decision-making and final fertilizer application output;

[0128] Through the aforementioned three linkage effects, the system has obtained the manure concentration factor (F1), soil moisture-depth coupling factor (F2), and historical-real-time linkage factor (F3). These three factors need to be normalized, and the final fertilizer application rate (Q) will be output through a multiplicative effect. final This allows for simultaneous consideration and collaborative decision-making, avoiding parameter conflicts caused by the superposition of parameters after traditional individual control. The normalized control formula is as follows:

[0129]

[0130] Among them, F1 (manure concentration factor) achieves equal nutrient fertilization, and the formula is: F1=C tar / C avg ;

[0131] When the real-time concentration Cavg Higher than target concentration C tar When F1 < 1, the absolute nutrient content per unit area is kept stable by reducing the fertilizer volume; when C avg Below C tar When F1 > 1, nutrients are supplemented by increasing the volume of fertilizer application; for example, C avg =2.5%, C tar When the concentration is 2.0%, F1 = 0.8. Even if the fertilizer volume is reduced by 20%, the total amount of nutrients per unit area will still be consistent with the target, avoiding nutrient deviation caused by concentration fluctuations.

[0132] F2 (soil moisture-depth coupling factor) integrates the synergistic effects of soil moisture and deep tillage depth, and the formula is as follows:

[0133] F2=1+β•(θ cal -θ opt )+α•(h cur -h base ) / h base ;

[0134] Where β is the soil moisture correction coefficient (usually β=0.02), used to fine-tune the effect of soil moisture on fertilizer application, and θ cal >θ opt When, β•(θ) cal -θ opt If θ is positive, F2 is slightly increased to compensate for possible runoff loss; cal <θ opt When this value is negative, F2 decreases slightly to prevent nutrients from stagnating in dry soil areas; for example, θ cal =26%, h cur When the depth is 26.5cm, F2 = 1 + 0.02 × (26 - 21.5) + 0.3 × (26.5 - 25) / 25 = 1.108. By increasing the amount of fertilizer by 10.8%, it can meet the dual needs of overly wet soil and increased depth.

[0135] F3 (historical-real-time linkage factor) is a dynamic correction factor determined based on historical yield and real-time growth status. It is used to realize potential tapping and risk warning. Its value ranges from 0.2 to 1.5. This factor integrates spatiotemporal information into the final fertilizer application decision.

[0136] Example Calculation and Effect Verification: Taking precision fertilization during the corn jointing stage as an example, if Q base =30m 3 / hectare, F1=0.8 (C avg =2.5%), F2=1.108 (θ) cal =26%, h cur =26.5cm), F3=1.4 (historically low yield + current good growth), then Qfinal =30×0.8×1.108×1.4≈34.97m 3 / hectare. The controller will Q final Converted into speed command for variable frequency fertilizer pump (via flow-speed calibration curve, 35m) 3 / hectare corresponds to a pump speed of 1500 r / min), and the flow sensor provides real-time feedback on the actual flow rate. If the actual flow rate is lower than the target value (e.g., only 32 m³ / h), 3 / hectare), the controller automatically increases the pump speed to 1580 r / min to ensure the actual fertilizer application rate matches the Q. final The deviation is ≤ ±5%.

[0137] Through normalized control, the system achieves deep integration of multi-source information. Manure characteristics ensure nutrient accuracy, soil conditions ensure fertilizer effectiveness location, and spatiotemporal data ensure strategy optimization. The synergistic effect of these three factors ensures that the final fertilizer application meets crop needs and adapts to environmental conditions, avoiding the limitations of single-factor control.

[0138] 3. Key parameter calibration method

[0139] To ensure the control accuracy of each module of the system, the key parameters must be calibrated before operation. The calibration process should combine field trials and laboratory analysis. The specific methods are as follows:

[0140] 3.1 NIR sensor calibration;

[0141] Objective: To ensure the accuracy of manure viscosity and concentration detection and avoid F1 calculation deviations caused by sensor drift. The steps are as follows:

[0142] (1) Collect commonly used manure samples from the target plots and prepare 25 samples in combination with 5 different viscosities (100 mPa•s, 200 mPa•s, 300 mPa•s, 400 mPa•s, 500 mPa•s) and 5 different NPK concentrations (0.5%, 1.5%, 2.5%, 3.5%, 4.5%).

[0143] (2) In the laboratory, the actual viscosity of the sample was determined using a rotational viscometer, and the actual NPK concentration was determined using the Kjeldahl method;

[0144] (3) Place the samples one by one into the NIR sensor for detection and record the spectral values ​​output by the sensor;

[0145] (4) A partial least squares regression (PLSR) algorithm was used to establish a fitting model between the spectral values ​​and the actual viscosity and concentration;

[0146] (5) Verify the accuracy of the model and ensure that the viscosity detection error is ≤ ±5% and the concentration detection error is ≤ ±3%. If the error exceeds the standard, re-collect samples to optimize the model.

[0147] 3.2 Soil moisture sensor calibration;

[0148] Objective: To eliminate the influence of soil texture on soil moisture monitoring and ensure the accuracy of θ readings. cal The accuracy of h tar Provide reliable evidence. The steps are as follows:

[0149] (1) Collect soil samples of three typical textures (sandy soil, loam, and clay soil) from the target plot, and take 1000g of each sample;

[0150] (2) In the laboratory, the samples were added with water to a volume moisture content of 5%, 10%, 15%, 20%, 25%, 30%, and 35%, respectively, stirred evenly, and then left to stand for 24 hours.

[0151] (3) The original soil moisture value of each sample was detected by TDR sensor, and the actual soil moisture value was determined by drying method (drying at 105℃ to constant weight);

[0152] (4) For each soil texture, calculate the ratio of the original test value to the actual value, which is the correction coefficient for that texture (e.g., correction coefficient for sandy soil = actual value / original value = 0.9).

[0153] (5) Store the correction coefficient in the data storage unit and automatically retrieve it according to the soil texture during operation.

[0154] 3.3 Depth-fertilizer coefficient (α) calibration;

[0155] Objective: To determine the optimal value of α, ensuring that h cur When Q changes, fine-tuning minimizes the differences in nutrient utilization at different depths. The steps are as follows:

[0156] (1) Select a fertile experimental field and divide it into 9 plots (3m×10m), which are then divided into 3 groups, with 3 plots in each group;

[0157] (2) Set 3 deep tillage depths (h1=h base -5cm, h2=h base h3=h base +5cm), each group of cells corresponds to one depth;

[0158] (3) For each plot, fertilizer application corrections were set at three levels: α=0.2, 0.3, and 0.4. The fertilizer application amount for each plot was calculated (Q=Q base ×(1+α×(hh base ) / h base ));

[0159] (4) After the crop matures, the yield and nutrient utilization rate of each plot are measured (nutrient utilization rate = crop nutrient absorption / fertilizer nutrient absorption × 100%).

[0160] (5) Analyze the data and select the α value that minimizes the difference in nutrient utilization rate at the three depths (≤5%). Usually, α=0.3 is the optimal value.

[0161] 3.4 Calibrating the dynamic correction factor (F3);

[0162] Objective: To determine the value range of F3 under different scenarios, ensuring that F3 can effectively improve output or reduce risk. The steps are as follows:

[0163] (1) Collect data on the target plot’s “historical yield - real-time growth - fertilizer application - actual yield” for the past 5 years and establish a database;

[0164] (2) Divide the data into 5 scenarios according to “historical yield level - ΔNDVI” (such as high yield + normal growth, high yield + poor growth, etc.).

[0165] (3) For each scenario, select different F3 values ​​(such as 0.2, 0.3, 0.4, ..., 1.5) and analyze the relationship between F3 and actual yield and nutrient utilization rate;

[0166] (4) For each scenario, determine the optimal range of F3—the range that maximizes actual yield and nutrient utilization (e.g., in the low-yield + high-growth scenario, when F3 = 1.2 to 1.5, the yield increases by more than 15% and the nutrient utilization rate is ≥60%).

[0167] (5) Store the correspondence between the scene and the F3 range in the controller, and automatically match them during operation.

[0168] 4. System workflow example;

[0169] With corn jointing stage (h base =25cm, Q base =30m 3 / hectare, C tar =2.0%, θ opt Taking precision fertilization of a certain plot of land as an example (=21.5%), the complete workflow of the system is as follows, intuitively demonstrating the synergistic effect of each linkage:

[0170] 4.1 Initialization phase (10 minutes before the task begins);

[0171] The operator selects the crop variety (corn) and growth stage (jointing stage) via the touchscreen, and the controller automatically loads the corresponding parameters from the data storage unit: h base =25cm, Q base =30m 3 / hectare, Ctar =2.0%, θ opt =21.5%, μ std =300mPa•s,v base =6km / h;

[0172] Load the historical production data (Y) of the target plot for the past 3 years. hist =5800 kg / ha, belonging to the low-yield category), generate a soil potential fertilization benchmark map (such as Figure 3 As shown), the F of the current work area base =0.8;

[0173] The system automatically completes sensor self-tests: the NIR sensor collects a blank sample (air) and checks whether the spectral value is normal; the TDR sensor detects the soil moisture value in the air (which should be close to 0%); the laser rangefinder detects the initial height (which should be close to 50cm above the ground). If all sensors are normal, the system enters standby mode.

[0174] 4.2 Movement and Perception Stage (In-process);

[0175] The tractor starts, with v base At a speed of 6 km / h, all sensors begin collecting data in real time.

[0176] NIR sensors collected manure samples, yielding μ=350mPa•s and C=2.5%. After preprocessing, μ... avg =348 mPa•s, C avg =2.48%;

[0177] The TDR sensor collected soil moisture data; the original value was 28%. The soil texture was clayey, and the correction factor was 1.1. θ was then obtained. cal =28%×1.1=30.8% (overly humid);

[0178] NDVI sensors collect crop growth data. cur =0.72, NDVI during maize jointing stage opt =0.6, then:

[0179] NDVI norm =0.72 / 0.6=1.2;

[0180] Laser rangefinder sensor detects h cur =25cm (initial depth);

[0181] The flow sensor detects an initial flow rate of 25 L / min (corresponding to Q_base = 30 m). 3 / hectare).

[0182] 4.3 Linked Calculation Phase (Performed in Real Time);

[0183] Calculation of the linkage effect between manure properties and mechanical travel speed:

[0184] K fer =0.4×(348 / 300)+0.6×(2.48 / 2.0)=0.4×1.16+0.6×1.24=0.464+0.744=1.208;

[0185] v tar =6×exp(-0.3×(1.208-1))=6×exp(-0.3×0.208)=6×0.939≈5.63km / h;

[0186] The controller sends v tar The command was sent to the autonomous driving system, and the actual speed gradually decreased to 5.63 km / h, with a deviation of ≤ ±0.2 km / h.

[0187] Calculation of the three-level effect of soil moisture, deep loosening depth, and fertilizer application rate:

[0188] h tar =25+5×(30.8-21.5) / (35-21.5)=25+5×9.3 / 13.5≈25+3.44=28.44cm;

[0189] The hydraulic depth control system adjusts the deep loosening plow, h cur Gradually increased from 25cm to 28.4cm (deviation 0.04cm);

[0190] Q=30×(1+0.3×(28.4-25) / 25)=30×(1+0.3×3.4 / 25)=30×(1+0.0408)=31.224m 3 / hectare;

[0191] F2=1+0.02×(30.8-21.5)+0.3×(28.4-25) / 25=1+0.02×9.3+0.3×3.4 / 25=1+0.186+0.0408=1.2268.

[0192] Calculation of the spatiotemporal linkage effect between historical yield and real-time crop growth:

[0193] Historical production level is low (Y) hist =5800 kg / hectare), historical NDVI for the same period hist =0.5;

[0194] ΔNDVI = (0.72 - 0.5) / 0.5 = 0.44 > 0.1 (excellent growth);

[0195] Matching scenario four, F3=1.4 (take the midpoint of the scenario range of 1.2 to 1.5).

[0196] Normalized control calculation:

[0197] F1=C tar / C avg =2.0 / 2.48≈0.806;

[0198] Q final =30×0.806×1.2268×1.4≈30×0.806×1.7175≈30×1.384≈41.52m 3 / hectares.

[0199] 4.4 Execution and Feedback Phase (conducted in real time);

[0200] The controller will Q final =41.52m 3 / hectare converted to the speed command of the variable frequency fertilizer pump: According to the variable frequency fertilizer pump flow-speed calibration data table (suitable for corn jointing stage operations, as shown in the table below), 41.52m 3 / hectare corresponds to a pump speed of 1850 r / min;

[0201] Manure viscosity: 300 mPa•s; NPK concentration: 2.0%; Tractor speed: 6 km / h; Curve fitting method: Linear regression (R²) 2 =0.99);

[0202] Serial Number Pump speed (r / min) <![CDATA[Corresponding fertilization rate (m 3 / ha)]]> Remark 1 0 0.0 Initial standby state 2 200 4.3 - 3 400 8.6 - 4 600 12.9 - 5 800 17.2 - 6 1000 21.5 Corresponding to the low fertilizer application rate after F1 adjustment 7 1200 25.8 - 8 1400 30.1 - 9 1500 35.0 <![CDATA[Corresponding to Q base Conventional fertilization amount after fine-tuning]]> 10 1600 37.2 - 11 1800 40.8 - 12 1850 41.5 <![CDATA[Corresponding to the jointing stage of maize Q final > 13 2000 48.0 Maximum speed state

[0203] The variable frequency fertilizer pump was adjusted to 1850 r / min, and the flow sensor provided real-time feedback that the actual flow rate was 36.2 L / min (corresponding to Q=41.4 m). 3 / hectare), with Q final The deviation is 0.12m. 3 / hectare (≤±5%), no further adjustment is required;

[0204] The controller stores key data every 10 seconds: timestamp, latitude and longitude, Q. final h cur v cur NDVI cur θ cal This is used for subsequent work review and parameter optimization.

[0205] 4.5 Post-assignment review phase (after the assignment is completed);

[0206] Operators view operation data reports via touchscreen, including: total operation area (e.g., 10 hectares), total fertilizer application rate (e.g., 415m³). 3), average Q final (e.g., 41.5m) 3 / h, average h cur (e.g., 28.3cm), average v cur (e.g., 5.6 km / h);

[0207] like Figure 4 and Figure 5 As shown, this visually illustrates the differences in fertilization and depth across different areas within the plot. If a certain area Q... final Significantly higher (e.g., >50m) 3 / hectare), indicating that the area may be a potential plot of land with historically low yields but excellent growth, and should be closely monitored in the next quarter;

[0208] Based on the data from this operation, the system optimizes some parameters (such as adjusting α from 0.3 to 0.32) to improve the accuracy of the next operation.

[0209] This invention achieves a technological upgrade in manure application from static prescription to dynamic adaptation through deep coupling and dynamic feedback of multi-source information, solving core problems in traditional fertilization such as isolated factor control, nutrient waste, and unstable fertilizer efficiency. By proposing a manure characteristic-speed linkage, interfering factors are transformed into control parameters; by constructing a three-level linkage of soil moisture, depth, and fertilizer amount, precise matching of soil and fertilizer is achieved; and by establishing a historical-real-time dialectical decision-making model, the system possesses potential exploration and risk warning capabilities. This invention can be widely applied to the precise deep application of manure to field crops such as corn, wheat, and rice, increasing crop yield by 10%–20% while reducing nutrient waste by more than 25%, demonstrating significant economic and ecological benefits.

[0210] It should be noted that, in this invention, although the specification describes the embodiments, not every embodiment contains only one independent technical solution. This way of describing the specification is only for clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A method for precise application of liquid manure and synergistic soil improvement through deep tillage, characterized in that, Includes the following steps: S1. Initialization phase: Load the variety parameters, growth period parameters and historical yield data of the target crop and the target plot, complete the self-check of the sensing module, and ensure that the sensor is in normal working condition. S2, Travel and Sensing Stage: During the tractor's travel, the sensing module collects real-time data on manure characteristics, soil moisture, crop growth, deep tillage depth, and manure flow rate. S3, Linkage Calculation Stage: Based on the real-time data collected in step S2, calculate the linkage parameters of manure characteristics-traveling speed, soil moisture-deep loosening depth-fertilizer application rate, and historical yield-real-time growth status respectively. Then, combine the normalized control model to calculate the final fertilizer application rate, target travel speed, and target deep loosening depth. S4. Execution and Feedback Phase: The execution module adjusts the travel speed, deep tillage depth and fertilizer application amount according to the target parameters obtained in step S3. At the same time, the sensing module provides real-time feedback of actual operating data to dynamically correct deviations and ensure operational accuracy. S5. Post-operation review stage: Generate operation data reports and operation parameter distribution charts, optimize the linkage calculation parameters in the control module based on the operation data, and provide data support for subsequent operations; In step S3, the calculation of the manure characteristics-travel speed linkage parameter is achieved by constructing a comprehensive manure application index, weighting and integrating manure viscosity and NPK nutrient concentration, and then mapping the index to the target travel speed to achieve the matching of manure characteristics and travel speed. The calculation of the linkage parameters of soil moisture, deep loosening depth and fertilizer application rate is to first set the target deep loosening depth based on soil moisture, and then fine-tune the fertilizer application rate according to the actual deep loosening depth to ensure uniform nutrient density when the soil disturbance volume changes. The calculation of the historical yield-real-time crop growth linkage parameter involves first classifying plot yield levels based on historical yields and setting corresponding benchmark fertilization coefficients, and then combining this with real-time crop growth to calculate dynamic correction factors, thereby enabling potential discovery and risk warning.

2. The method for precise application of liquid manure and synergistic soil improvement via deep tillage as described in claim 1, characterized in that, In step S2, the collected real-time data is preprocessed: Outlier removal and smoothing are performed on the manure characteristic data collected by online near-infrared sensors to reduce the impact of environmental interference on data accuracy. The soil moisture data collected by the TDR soil moisture sensor is corrected according to the soil texture of the target plot to improve the reliability of the soil moisture data.

3. The method for precise application of liquid manure and synergistic soil improvement via deep tillage as described in claim 1, characterized in that, In step S4, the deviation correction adopts a PID control algorithm to perform closed-loop adjustment of the travel speed, deep loosening depth and fertilizer application amount, so that the deviation of the actual travel speed, actual deep loosening depth and actual fertilizer application amount from the corresponding target parameters is controlled within a preset range, ensuring the stability and accuracy of the operation parameters.

4. The method for precise application of liquid manure and synergistic soil improvement via deep tillage as described in claim 1, characterized in that, In step S5, the operation data report shall include at least the total operation area, total fertilizer application, average fertilizer application, average deep tillage depth, and average travel speed. The operation parameter distribution map includes at least a deep tillage depth distribution map and a fertilizer application rate distribution map; The operation results are presented intuitively through operation data reports and distribution charts, and the optimized linkage calculation parameters are stored in the data storage unit of the control module to improve the accuracy of subsequent similar operations.

5. A device for precise application of liquid manure and deep tillage in conjunction with soil improvement, implementing the method of precise application of liquid manure and deep tillage in conjunction with soil improvement as described in claim 1, characterized in that, It includes a sensing module, an execution module, a control module, and a communication module; The sensing module is used to collect real-time data on manure characteristics, soil moisture, crop growth, deep tillage depth, and manure flow rate. It includes at least an online near-infrared sensor, a TDR-type soil moisture sensor, a crop growth detection sensor, a depth detection sensor, and a flow rate sensor. The execution module is used to perform travel speed adjustment, deep tillage depth adjustment and fertilizer application rate adjustment, and includes at least a tractor speed control system, a hydraulic depth control system and a variable frequency fertilizer pump. The control module is connected to the sensing module and the execution module through a communication module. It has a built-in data storage unit for storing crop variety parameters, crop growth period parameters, historical yield data of the plot, and sensor calibration parameters. Based on the data collected by the sensing module, it can perform linkage calculations of manure characteristics-traveling speed, soil moisture-deep loosening depth-fertilizer application rate, historical yield-real-time growth, and output control commands to the execution module. The communication module is used to realize real-time data interaction between the sensing module, control module, and execution module, ensuring that the modules work together.

6. The device for precise application of liquid manure and synergistic soil improvement via deep tillage as described in claim 5, characterized in that, The crop growth detection sensor is an airborne NDVI sensor or a soil fertility sensor. When the operation scenario is before sowing or during the seedling stage, the crop growth detection sensor is switched to a soil fertility sensor to collect real-time soil nutrient data to replace crop growth data and participate in the linkage calculation of historical yield and real-time growth.

7. The device for precise application of liquid manure and synergistic soil improvement via deep tillage as described in claim 5, characterized in that, The TDR-type soil moisture sensor is installed in front of the subtilist to collect soil moisture data in advance, avoiding interference from soil structure changes after subtilizing on the moisture detection results. The depth detection sensor is installed on the side of the subsoiler to monitor the actual depth of the subsoiler in real time. The flow sensor is installed at the outlet of the fertilizer pipeline to provide real-time feedback on the actual flow rate of the manure.

8. The device for precise application of liquid manure and synergistic soil improvement via deep tillage as described in claim 5, characterized in that, The hydraulic depth control system includes hydraulic cylinders, servo valves, and displacement sensors. The displacement sensor is used to detect the extension and retraction of the hydraulic cylinder. In conjunction with the commands output by the control module, it realizes closed-loop control of the deep tillage depth and ensures the accuracy of deep tillage depth adjustment. The tractor speed control system changes the tractor's travel speed by adjusting the speed of the hydraulic motor, adapting to changes in the characteristics of manure.

9. The device for precise application of liquid manure and deep loosening of soil for synergistic improvement according to claim 5, characterized in that, The control module uses an industrial-grade MCU, which has a multi-channel analog input interface and a PWM output interface. It can process real-time data from multiple sensors simultaneously and output control commands quickly. The data storage unit adopts a dual storage design, which can store historical operation data, sensor calibration parameters and crop-related parameters for a long time, and supports data backtracking and operation review.

Images