Water and fertilizer integrated control system and control method based on Internet of Things

Abstract

The invention relates to the technical field of water and fertilizer control, in particular to a water and fertilizer integrated control system and method based on the Internet of Things. The method comprises the following steps of obtaining environment measured data, historical environment data and regional geographic information of a farmland region, performing spatial modeling and time sequence learning on the farmland region, and constructing a farmland digital twinborn model; simulating the state of an area without sensors by using a farmland digital twinborn model to obtain virtual sensing node data; collecting multispectral image data of crops, and extracting physiological feature data of the crops; and constructing a crop physiological state model by utilizing the environment measured data, the virtual sensing node data and the crop physiological feature data, and generating crop physiological state evaluation data. Through multi-source data fusion and three-dimensional visualization integration, data-driven accurate regulation and control and whole-process intelligent management of the water and fertilizer integrated system are realized, and the intelligent level of irrigation and fertilization and the resource utilization efficiency are comprehensively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of water and fertilizer control, and in particular to a water and fertilizer integrated control system and a control method based on the Internet of Things. Background Art

[0002] Traditional water and fertilizer systems rely heavily on manual experience or timed control, making it difficult to adapt to the dynamic needs of crop growth. The introduction of IoT technology has enabled agricultural production to be equipped with environmental perception, remote control, and data-driven capabilities, driving the development of integrated water and fertilizer management towards intelligence and automation. Currently, more advanced technical solutions generally integrate soil moisture, meteorological, and water quality sensors to control irrigation and fertilization processes through preset rules. However, such methods still suffer from problems such as insufficient data coverage, delayed response mechanisms, and rough environmental modeling. This is especially true in areas with limited sensor distribution or complex farmland terrain, where it is difficult to fully perceive environmental conditions, leading to deviations in control strategies and affecting water and fertilizer utilization efficiency.

[0003] Although existing integrated water and fertilizer technology has a certain level of intelligence, it still has the following defects: First, there is a lack of accurate perception of areas where sensors are not deployed, resulting in inaccurate control of irrigation and fertilization ranges; second, there is a lack of in-depth modeling of crop physiological states, and only relies on soil parameters to judge crop needs, resulting in inaccurate fertilizer allocation decisions; third, the system is mostly rule-driven and lacks a dynamic response mechanism based on machine learning and expert models; fourth, there is a lack of visualization support, making it difficult for operation and maintenance personnel to grasp the overall status of farmland in real time. Summary of the Invention

[0004] Based on this, it is necessary for the present invention to provide a water-fertilizer integrated control system and control method based on the Internet of Things to solve at least one of the above technical problems.

[0005] To achieve the above objectives, a water-fertilizer integrated control method based on the Internet of Things includes the following steps:

[0006] Step S1: Obtaining measured environmental data, historical environmental data, and regional geographic information of the farmland area, and performing spatial modeling and time series learning on the farmland area to construct a farmland digital twin model; using the farmland digital twin model to simulate the state of the area without sensor deployment to obtain virtual sensor node data;

[0007] Step S2: Collect multispectral image data of crops and extract crop physiological characteristic data; use environmental measurement data, virtual sensor node data and crop physiological characteristic data to build a crop physiological state model and generate crop physiological state assessment data;

[0008] Step S3: real-time collection of water quality monitoring data; dynamic adjustment of the fertilizer-liquid ratio based on crop physiological status assessment data and water quality monitoring data, and control of the operation of the multi-channel fertilizer pump to generate fertilizer control parameters;

[0009] Step S4: Determine control decision data for irrigation duration, flow rate, and fertilizer amount based on virtual sensor node data and fertilizer control parameters;

[0010] Step S5: driving the solenoid valve, irrigation pump, and fertilizer pump to perform irrigation and fertilization operations based on the control decision data, and generating fertilization and irrigation execution log data;

[0011] Step S6: Build a three-dimensional visual smart farm twin dashboard based on the virtual sensor node data set, crop physiological status assessment data, water quality monitoring data, and fertilization and irrigation execution log data to generate a twin visual data view.

[0012] The present invention overcomes the key problems of insufficient data coverage and delayed response of traditional water and fertilizer systems by introducing multi-source data fusion, digital twin reconstruction, multispectral perception, dynamic water quality control and three-dimensional visualization integration, and realizes continuous perception and dynamic completion of farmland environmental status. Even when sensors cannot be fully deployed, the regional environmental distribution can be accurately reconstructed, effectively improving the spatial accuracy of regulation. Combined with crop canopy multispectral image analysis, it can obtain intuitive performance parameters of crop growth status, avoid relying solely on soil data to judge crop fertilizer requirements, and thus improve the pertinence and scientific nature of fertilizer-liquid ratio. By adjusting the fertilization plan through the linkage between crop health indicators and water quality status, the simultaneous optimization of irrigation and fertilization decisions can be achieved. ization to ensure the precise input of water and fertilizer resources; the control parameters of each link are generated by real-time data, avoiding the rigid response caused by artificially set rules, and giving the system the ability to adjust itself; the control information of the entire process is recorded and integrated in real time, further supporting subsequent operation traceability and strategy evaluation; at the same time, by building a farm visualization dashboard in a three-dimensional scene, environmental information, crop status, operation path and historical operation data are integrated and displayed, so that managers can intuitively grasp the overall operation status of the farmland and conduct interactive queries, thereby improving the system operation and maintenance efficiency and management transparency, and overall promoting the upgrade of the water-fertilizer integrated system from rule-triggered to data-driven, significantly improving the intelligence level and resource utilization efficiency of agricultural irrigation and fertilization.

[0013] Preferably, the present invention further provides an Internet of Things-based integrated water and fertilizer control system for executing the above-mentioned Internet of Things-based integrated water and fertilizer control method, wherein the Internet of Things-based integrated water and fertilizer control system comprises:

[0014] The twin modeling module is used to obtain measured environmental data, historical environmental data, and regional geographic information of farmland areas, and conduct spatial modeling and time series learning of farmland areas to build a digital twin model of farmland. The digital twin model of farmland is used to simulate the state of areas without sensor deployment to obtain virtual sensor node data.

[0015] The crop assessment module is used to collect multispectral image data of crops and extract crop physiological characteristic data; it uses environmental measurement data, virtual sensor node data and crop physiological characteristic data to build a crop physiological status model and generate crop physiological status assessment data;

[0016] The water and fertilizer control module is used to collect water quality monitoring data in real time; dynamically adjust the fertilizer-liquid ratio based on crop physiological status assessment data and water quality monitoring data, and control the operation of multi-channel fertilizer pumps to generate fertilizer control parameters;

[0017] Intelligent decision-making module, used to determine the control decision data of irrigation duration, flow rate and fertilizer amount based on virtual sensor node data and fertilizer control parameters;

[0018] An execution control module is used to drive the solenoid valve, irrigation pump, and fertilizer pump to perform irrigation and fertilization operations based on the control decision data, and generate fertilization and irrigation execution log data;

[0019] The visual twin module is used to build a three-dimensional visual smart farm twin dashboard based on virtual sensor node data sets, crop physiological status assessment data, water quality monitoring data, and fertilization and irrigation execution log data, and generate a twin visual data view.

[0020] Through the collaborative work of the above modules, the present invention forms a complete closed-loop integrated water and fertilizer intelligent control system, improves the system's adaptability and control accuracy to the complex dynamic environment of farmland, and realizes full-process automated management from data acquisition, status assessment, decision analysis to execution control and visual display; through the spatial and temporal correlation analysis capabilities of the twin modeling module, it effectively makes up for the blind spots of sensor coverage and ensures the integrity and continuity of the overall environmental status data of the farmland; the multispectral information extraction and physiological characteristic analysis of the crop assessment module enable the dynamic changes of crop nutritional needs to be perceived in real time, providing a scientific basis for precise fertilization; the cooperation between the water and fertilizer regulation module and the intelligent decision-making module ensures that the water and fertilizer supply dynamically adapts to the real-time needs of crops, significantly improving nutrient utilization efficiency and crop growth balance; the execution control module ensures the stable and precise operation of each control unit, and realizes the synchronous and efficient execution of the irrigation and fertilization processes; the visual twin module provides a comprehensive and intuitive farm status display, which facilitates the rapid grasp of the overall production picture, the tracing of historical operation processes and the diagnosis of abnormal conditions. During the operation of the overall system, relying on the data flow and logical connection between modules, the automation, intelligence and efficient management level of smart farms are continuously improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments thereof made with reference to the following drawings:

[0022] Figure 1This is a schematic flow chart of the steps of a water-fertilizer integrated control system and control method based on the Internet of Things of the present invention;

[0023] Figure 2 for Figure 1 Detailed step flow diagram of step S1;

[0024] Figure 3 for Figure 1 Detailed step flow chart of step S2 in FIG. DETAILED DESCRIPTION

[0025] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. It is obvious that the embodiments described are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts are within the scope of protection of the present invention.

[0026] In addition, the accompanying drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor and / or microcontroller approaches.

[0027] It should be understood that although the terms "first," "second," and the like may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used solely to distinguish one element from another. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element, without departing from the scope of the exemplary embodiments. The term "and / or" as used herein includes any and all combinations of one or more of the listed associated items.

[0028] To achieve this, please refer to Figures 1 to 3 The present invention provides a water-fertilizer integrated control method based on the Internet of Things, the method comprising the following steps:

[0029] Step S1: Obtaining measured environmental data, historical environmental data, and regional geographic information of the farmland area, and performing spatial modeling and time series learning on the farmland area to construct a farmland digital twin model; using the farmland digital twin model to simulate the state of the area without sensor deployment to obtain virtual sensor node data;

[0030] Step S2: Collect multispectral image data of crops and extract crop physiological characteristic data; use environmental measurement data, virtual sensor node data and crop physiological characteristic data to build a crop physiological state model and generate crop physiological state assessment data;

[0031] Step S3: real-time collection of water quality monitoring data; dynamic adjustment of the fertilizer-liquid ratio based on crop physiological status assessment data and water quality monitoring data, and control of the operation of the multi-channel fertilizer pump to generate fertilizer control parameters;

[0032] Step S4: Determine control decision data for irrigation duration, flow rate, and fertilizer amount based on virtual sensor node data and fertilizer control parameters;

[0033] Step S5: driving the solenoid valve, irrigation pump, and fertilizer pump to perform irrigation and fertilization operations based on the control decision data, and generating fertilization and irrigation execution log data;

[0034] Step S6: Build a three-dimensional visual smart farm twin dashboard based on the virtual sensor node data set, crop physiological status assessment data, water quality monitoring data, and fertilization and irrigation execution log data to generate a twin visual data view.

[0035] In the embodiment of the present invention, reference Figure 1 FIG. 1 is a flow chart showing the steps of a water-fertilizer integrated control method based on the Internet of Things according to the present invention. In this example, the water-fertilizer integrated control method based on the Internet of Things includes the following steps:

[0036] Step S1: Obtaining measured environmental data, historical environmental data, and regional geographic information of the farmland area, and performing spatial modeling and time series learning on the farmland area to construct a farmland digital twin model; using the farmland digital twin model to simulate the state of the area without sensor deployment to obtain virtual sensor node data;

[0037] In the embodiment of the present invention, firstly, soil sensors, air environment monitoring devices and light collection modules are deployed in the farmland area to obtain six types of environmental measured data, including soil temperature, soil moisture, soil conductivity, air temperature, air humidity and light intensity. The collection frequency of each type of data is fixed at once every 30 minutes, and the data format is a structured numerical record corresponding to the timestamp; then, the same six types of environmental monitoring data in the past five years in the established farmland area historical database are called, and data are extracted according to the time granularity of days, and the period is segmented according to months to form a historical environmental data set; then, a drone equipped with a multispectral camera system and an RTK high-precision positioning device is used to conduct a full-coverage low-altitude flight scan of the farmland area to obtain geographical elements such as regional terrain, boundaries and irrigation ditches, and surface elevation changes, and through rasterization processing and spatial resampling, The geographic information is unified to a spatial resolution of 0.5 meters. After completing the above data collection, a three-dimensional data interpolation method and time smoothing weighted processing are used to superimpose and fuse the measured data and historical data according to spatial coordinates and time axes, forming a space-time joint numerical matrix at each sampling node to constitute the basic data of farmland spatial structure and dynamic changes. Based on the above fused data, the soil temperature, humidity, conductivity and air temperature, humidity and light estimation values ​​of the area without sensors are calculated according to a fixed grid grid lattice (such as one node every 10 meters). The estimation adopts the distance-weighted average method based on known nodes, in which the weighting factor is set as a function inversely proportional to the square of the distance, and the threshold radius is set to 50 meters. If it exceeds the threshold, the weight is zero, and finally a data set of equally spaced virtual nodes is formed in the entire area. The data of each virtual node is composed of the same six indicators as the measured nodes.

[0038] Step S2: Collect multispectral image data of crops and extract crop physiological characteristic data; use environmental measurement data, virtual sensor node data and crop physiological characteristic data to build a crop physiological state model and generate crop physiological state assessment data;

[0039] In the embodiment of the present invention, the crop canopy in the farmland area is first photographed in parallel strips by a multispectral imaging device installed on an agricultural automatic rail car. The imaging device used includes a visible light band (wavelength range of 400-700nm), a red edge band (wavelength range of 700-740nm) and a near infrared band (wavelength range of 740-900nm). The imaging resolution is 0.1 meters and the shooting frequency is once every 7 days. After acquisition, all images are uniformly subjected to radiation correction and geometric registration processing; then, the collected multispectral images are subjected to band-by-band processing, based on the red light and The normalized difference vegetation index (NDVI) is calculated using the near-infrared band as follows: (near-infrared band value - red light band value) / (near-infrared band value + red light band value), and the calculation result is bound to the geographic coordinates to generate a raster map. At the same time, the mean and standard deviation of the RGB color values ​​are extracted from the visible light image to determine the distribution of leaf color depth. The curvature variation of the crop leaf edge is identified based on the image edge detection method, and the leaf curling degree is determined by calculating the second-order derivative of the edge segment. The curling degree is set from 0 to 3, corresponding to no curl, slight curl, moderate curl, and Severe curling; in addition, for the distribution of reflectance of each band at each pixel point, the difference between the reflectance of the five bands is used to statistically analyze its variation range, extract the reflectance response gradient, and construct a crop physiological characteristic data set consisting of NDVI index value, color statistical value, curling degree level and reflectance gradient; then, the data set is uniformly time-calibrated and spatially matched with the environmental measured data and virtual sensor node data obtained in step S1, and the data are aligned according to the plot number and acquisition time to form a unified data structure for each plot; then, the multi-temporal crop growth in each plot is analyzed. The moving average processing (with a window width of 3 days) is performed on the numerical sequence of physiological characteristics to weaken the impact of short-term fluctuations. The difference between the current NDVI value, color mean, and curl level and the historical mean is compared at fixed time intervals (every 7 days). If the NDVI value decreases by more than 10%, the yellowing ratio of the color mean exceeds 15%, and the curl level increases by more than 1 level, and any two of the above conditions are met at the same time, the current crop status of the area is marked as "abnormal", otherwise it is marked as "normal". Finally, the crop physiological status assessment data containing plot number, timestamp and status mark fields are generated.

[0040] Step S3: real-time collection of water quality monitoring data; dynamic adjustment of the fertilizer-liquid ratio based on crop physiological status assessment data and water quality monitoring data, and control of the operation of the multi-channel fertilizer pump to generate fertilizer control parameters;

[0041] In the embodiment of the present invention, an integrated water quality online monitoring module including an electrode pH probe, a conductivity sensor and a TDS sensor is first installed at the water inlet of the farmland water source. The pH value, conductivity (in μS / cm) and total dissolved solids concentration TDS (in mg / L) are collected every 10 minutes. The collection range is limited to pH 5.5-8.5, conductivity 0-2000 μS / cm, and TDS 0–1200 mg / L, the collected water quality data is transmitted to the data control terminal via the RS485 bus; then, the status of each plot is classified according to the crop physiological status assessment data generated in step S2, and marked as "normal" or "abnormal" according to the status field, and then the water quality data and crop status data are associated with the plot number; then, the required fertilizer solution concentration adjustment direction is determined according to the plot crop status and water quality indicators. If the plot status is "abnormal" and the pH is less than 6.0 or greater than 7.5, the phosphorus solution ratio is increased by 10%; if the conductivity is less than 400 μS / cm, the nitrogen solution ratio is increased by 15% based on the original ratio; if the TDS is greater than 900 mg / L, the potassium solution ratio is reduced by 10%; each type of adjustment is recorded as a percentage, and the adjustment result is used to generate the single plot fertilizer setting parameters. The system uses a parameter group, including the ratio of nitrogen, phosphorus, and potassium solutions (the total is limited to 100%) and the target concentration of each type of fertilizer solution (in mg / L). The target concentration range is set to 100–300 mg / L for nitrogen solution, 50–150 mg / L for phosphorus solution, and 80–250 mg / L for potassium solution. The operation control unit of the electronically controlled multi-channel fertilizer pump is driven according to this parameter group. The opening time ratio of the solenoid valves in each fertilizer channel is controlled by PWM signals to ensure that the actual output fertilizer concentration is consistent with the set value. The fertilization liquid mixing process is accurately measured by a three-channel flowmeter. The duration of a single fertilization task does not exceed 20 minutes, and the output pressure is stably controlled between 0.3 and 0.5 MPa. Finally, the original data of the plot number, fertilizer solution ratio, flow rate, duration, and water quality corresponding to each operation are packaged to generate the fertilizer control parameters.

[0042] Step S4: Determine control decision data for irrigation duration, flow rate, and fertilizer amount based on virtual sensor node data and fertilizer control parameters;

[0043] In the embodiment of the present invention, first, the virtual sensor node data generated in step S1 is spatially interpolated at intervals of 0.5 meters to obtain the current values ​​and 72-hour change trends of three types of indicators, namely, soil moisture (in %), soil conductivity (in mS / cm), and soil temperature (in ℃) at each location of the farmland. The soil moisture threshold is set between 30% and 70%, the conductivity threshold is set between 0.2 and 2.5 mS / cm, and the temperature threshold is set between 10 and 35 ℃. Secondly, combined with the nitrogen, phosphorus, and potassium solution concentrations and the total fertilizer solution ratio set for each plot in the fertilizer control parameters in step S3, the target fertilizer amount for each plot is uniformly converted into the volume of fertilizer solution per square meter (in L / m 2 ), the conversion method is the total fertilizer requirement (unit is L) divided by the plot area (unit is m 2 ); Then, for each plot, the irrigation flow rate (in L / min) is determined proportionally according to the proportion of pixels in the virtual nodes within the plot with soil moisture below 35% or electrical conductivity below 0.3mS / cm. If the proportion of low-humidity areas is greater than 50%, the irrigation flow rate is set to 120% of the standard value; if the proportion is between 30% and 50%, it is set to 110% of the standard value; otherwise, it is set to the standard value; the standard flow rate is fixed at 50L / min based on the pipe network diameter and the water output capacity of the irrigation pump; then, based on the total amount of fertilizer applied and the set irrigation flow rate, the irrigation time of each plot is calculated. The calculation method is the total amount of fertilizer liquid applied (in The calculation is based on the total irrigation time (in L) divided by the irrigation flow rate (in L / min). If the calculated value exceeds 30 minutes, the total amount is forced to be applied in two batches with an interval of 3 hours to avoid excessive water pressure on the crop roots in a short period of time. At the same time, the fertilizer amount is adjusted based on the proportional relationship between the pixel area with soil conductivity below 0.3mS / cm and the fertilizer concentration. If this area accounts for more than 60%, the total fertilizer solution amount is increased by 10%. All control parameters are recorded in units of plots, including plot number, set irrigation time (in minutes), set irrigation flow rate (in L / min), and total fertilizer solution amount (in L), and finally generate control decision data.

[0044] Step S5: driving the solenoid valve, irrigation pump, and fertilizer pump to perform irrigation and fertilization operations based on the control decision data, and generating fertilization and irrigation execution log data;

[0045] In the embodiment of the present invention, first, according to the control decision data generated in step S4, the three instruction parameters of irrigation time (in minutes), irrigation flow (in L / min), and total amount of fertilizer solution (in L) corresponding to each block number are received by the PLC controller in the irrigation control center, and the solenoid valve address mapping table of the corresponding number is called out. The PLC controller issues an opening instruction to the solenoid valve of the specified address using the Modbus communication protocol. The solenoid valve response time is controlled within 1 second. The on-off action of the solenoid valve is fed back to the controller in real time to confirm that the state is correct and then start the irrigation pump; the irrigation pump is a variable frequency constant pressure pump with a rated power of 1.5kW. After starting, the pump speed is adjusted by the inverter according to the flow control parameters to stably control the output pressure at 0.4MPa. The water outlet is connected to a flow meter to monitor the flow change in real time and sample once every 10 seconds. If the sampling results deviate from the target flow value by more than ±5% for three consecutive times, the speed is automatically adjusted to compensate for the error. After the irrigation pump has been running for 5 minutes, the three-channel electric fertilizer pump is started through the relay control device. The opening of the flow valve of each channel is controlled according to the fertilizer-liquid ratio set in step S3. The flow valve response time is 500 milliseconds. After the fertilizer liquid is injected into the main pipe, it is completely mixed with the irrigation water flow and the total output is detected by the mixing flow meter. If the mixed concentration is lower than the set value, the channel flow rate is increased by 10% through the pressure pump. During the fertilization process, the data of all solenoid valves, electric pumps, flow meters and pressure sensors are recorded by the real-time monitoring module of the control center, including on-off time, operating current, flow value, pressure curve and fertilizer ratio. After the task is completed, a fertilization and irrigation execution log data file is automatically generated. Each record contains the plot number, irrigation start time, end time, total fertilizer amount (unit: L), irrigation duration (unit: minute), fertilizer ratio of each channel, solenoid valve opening time, electric pump operating status and flow meter reading sequence.

[0046] Step S6: Build a three-dimensional visual smart farm twin dashboard based on the virtual sensor node data set, crop physiological status assessment data, water quality monitoring data, and fertilization and irrigation execution log data to generate a twin visual data view.

[0047] In the embodiment of the present invention, the virtual sensor node dataset generated in step S1 is first imported into the GIS three-dimensional geographic rendering engine, and the soil temperature (in ℃), humidity (in %), conductivity (in mS / cm) and air temperature, humidity and light intensity of each node are coordinate matched and spatially mapped. A basic environmental layer is constructed with a 0.5-meter grid as the unit, and each parameter is assigned an RGB value range using a color gradient encoding method, where soil moisture is encoded using the blue channel and light intensity is encoded using the yellow channel. All parameter values ​​are updated every 10 minutes; then, the crop physiological status assessment data generated in step S2 is associated with the plot number and the timestamp, and the canopy color distribution is rendered in the crop layer using the NDVI index range (0-1), and the leaf curling degree level (0-3) corresponds to the three-dimensional undulation of the canopy, thereby realizing a three-dimensional visual expression of the crop growth status, where curling degree level 0 corresponds to a smooth surface and level 3 corresponds to a maximum undulation height of 0.5 meters; then, the pH value, conductivity and TDS water quality monitoring data collected in step S3 are introduced, and a three-dimensional NDVI index is constructed at each water inlet node. The water quality layer performs segmented averaging of data in a time series format, calculating the average value every hour and encoding changes in water quality indicators using a color scale. pH values ​​deviating from the neutral range (6.5–7.5) are marked in red, and high TDS areas are enhanced with gray shading to indicate areas with excessively high concentrations. The fertilization and irrigation execution log data generated in step S5 is then parsed into a task trajectory layer. The water flow path is drawn according to the solenoid valve control time period, and the water and fertilizer delivery direction is dynamically displayed in the form of a linear animation. Annotation text of the fertilizer ratio and fertilizer amount is superimposed. The trajectory animation is played frame by frame per plot, and the time axis granularity is set to 1 minute. The above four layers are uniformly superimposed into the same 3D coordinate system using geographic coordinates and timestamps. A 3D visual smart farm scene is generated using the WebGL rendering engine. Layer check buttons, sliding time axes, and plot filtering conditions are set in the operation interface. Finally, a 3D twin dashboard integrating environmental status, crop status, water quality indicators, and fertilization execution is constructed. Twin visual data views are generated in the form of panoramic bird's-eye views, local magnification views, and time series views, allowing continuous monitoring and backtracking of the water and fertilizer control system.

[0048] The present invention overcomes the key problems of insufficient data coverage and delayed response of traditional water and fertilizer systems by introducing multi-source data fusion, digital twin reconstruction, multispectral perception, dynamic water quality control and three-dimensional visualization integration, and realizes continuous perception and dynamic completion of farmland environmental status. Even when sensors cannot be fully deployed, the regional environmental distribution can be accurately reconstructed, effectively improving the spatial accuracy of regulation. Combined with crop canopy multispectral image analysis, it can obtain intuitive performance parameters of crop growth status, avoid relying solely on soil data to judge crop fertilizer requirements, and thus improve the pertinence and scientific nature of fertilizer-liquid ratio. By adjusting the fertilization plan through the linkage between crop health indicators and water quality status, the simultaneous optimization of irrigation and fertilization decisions can be achieved. ization to ensure the precise input of water and fertilizer resources; the control parameters of each link are generated by real-time data, avoiding the rigid response caused by artificially set rules, and giving the system the ability to adjust itself; the control information of the entire process is recorded and integrated in real time, further supporting subsequent operation traceability and strategy evaluation; at the same time, by building a farm visualization dashboard in a three-dimensional scene, environmental information, crop status, operation path and historical operation data are integrated and displayed, so that managers can intuitively grasp the overall operation status of the farmland and conduct interactive queries, thereby improving the system operation and maintenance efficiency and management transparency, and overall promoting the upgrade of the water-fertilizer integrated system from rule-triggered to data-driven, significantly improving the intelligence level and resource utilization efficiency of agricultural irrigation and fertilization.

[0049] Preferably, step S1 includes the following steps:

[0050] Step S11: collecting soil temperature, soil moisture, soil conductivity, air temperature and humidity, and light intensity data of the farmland area to obtain actual environmental measurement data;

[0051] Step S12: Acquire historical environmental data of the farmland area;

[0052] Step S13: collecting topography, boundaries and geographical element information of the farmland area and performing digital processing to obtain regional geographical information;

[0053] Step S14: constructing a farmland regional spatial model based on the measured environmental data, historical environmental data and regional geographic information to obtain a farmland spatial structure model;

[0054] Step S15: constructing a farmland environment time series based on the measured environmental data and historical environmental data to obtain a farmland time dynamic model;

[0055] Step S16: constructing a farmland digital twin model based on the farmland spatial structure model and the farmland temporal dynamic model;

[0056] Step S17: Simulate the state of the area without sensor deployment based on the farmland digital twin model to obtain virtual sensor node data.

[0057] In the embodiment of the present invention, wireless soil monitoring terminals and meteorological collection modules are first deployed at intervals of 20 meters in the farmland area. The soil monitoring terminals are used to collect soil temperature (in degrees Celsius, ranging from 0 to 40), soil moisture (in percent, ranging from 10 to 90), and soil conductivity (in mS / cm, ranging from 0.1 to 2.5). The meteorological collection module collects air temperature (in degrees Celsius, ranging from -10 to 45), air humidity (in percent, ranging from 10 to 95), and light intensity (in lux, ranging from 1000 to 100000). The data collection cycle is set to 10 minutes, and the data is collected through L The oRa wireless transmission gateway aggregates and stores the data in the data server as the measured environmental data; then the historical meteorological data and soil moisture monitoring records stored in the area for the past five years are called up in the agricultural data archive system, the recording time is aligned in hourly units, and the format of the measured data is unified to form a historical environmental data set; then the multispectral photography device and lidar equipment mounted on the drone are used to perform low-altitude scanning of the entire area, extracting geographic features such as elevation, slope, boundary, ditch, canal and road, and converting the original point cloud data into DEM (digital elevation map) and vector layer, with the spatial resolution limited to 0.5 meters, and all geographic information is collected. The information is unified into the UTM coordinate system through projection coordinate conversion to form regional geographic information data; after completing the data collection, the measured environmental data, historical environmental data and regional geographic information are rasterized according to the spatial position in accordance with the coordinate superposition method, and the average soil and meteorological values ​​of each 0.5-meter grid node are calculated using bilinear interpolation to construct a farmland spatial structure grid composed of six environmental elements and form a farmland spatial structure model; then the measured values ​​of each sensor node within 30 days are trend fitted, and the fixed time sliding average method (window width is 6 hours) is used to extract the change rate and periodic change amplitude of soil and meteorological elements, and Dynamically changing data sequences are generated by time index, and a temporal dynamic model of farmland is constructed. Based on the spatial structure grid and the time series change curve, each grid node is defined as a six-dimensional vector. The six dimensions are the current soil temperature, soil moisture, electrical conductivity, air temperature, air humidity, and light intensity. In areas where sensors are not deployed, the inverse distance weighted method is used for data estimation. The weight function is defined as inversely proportional to the square of the distance, and the weight cutoff radius is limited to 30 meters. Six-dimensional virtual data points are generated in each undeployed area. Finally, the spatial environment values ​​and estimated values ​​of all grid nodes are integrated to form a complete data atlas as virtual sensor node data.

[0058] The present invention integrates measured farmland environment data, historical meteorological and soil moisture data, and high-precision geographic information to construct a farmland structure foundation covering both spatial and temporal dimensions. This can accurately reflect the distribution characteristics and changing patterns of the internal microenvironment of the farmland, and provide a high-resolution, high-continuity environmental input basis for subsequent irrigation and fertilization control. By establishing a coupling system of spatial structure and temporal changes, the integrity and dynamics of farmland state expression are effectively improved, so that when the system faces complex terrain or data-missing areas, it can still rely on surrounding multi-dimensional information to reconstruct the environmental state, thereby enhancing the system's environmental cognition ability. The digitally processed geographic element information can support the precise correspondence between environmental data and spatial position, and realize refined control granularity. By generating virtual sensor node data, dynamic compensation for sensor blind spots is achieved, overcoming the problem of incomplete data coverage caused by cost or terrain restrictions in actual deployment, ensuring the continuity of environmental parameters within the entire field, providing high-quality underlying data support for accurate decision-making, and significantly improving the spatial perception ability and decision-making scientificity of the entire water and fertilizer control system.

[0059] Preferably, step S2 includes the following steps:

[0060] Step S21: Collect visible light, red edge, and near-infrared multispectral images of crop canopies to obtain crop multispectral image data;

[0061] Step S22: extracting NDVI index, crop leaf color, curling degree, and spectral reflectance characteristics based on the crop multispectral image data to obtain crop physiological characteristic data;

[0062] Step S23: Based on the measured environmental data, the virtual sensor node data and the crop physiological characteristic data, a fusion model of the crop growth environment and the crop state is performed to obtain a crop physiological state model;

[0063] Step S24: Evaluate the current crop health and nutritional status based on the crop physiological status model to obtain crop physiological status evaluation data.

[0064] In the embodiment of the present invention, first, in the middle and late stages of crop growth, a track-type automatic inspection device equipped with a multispectral camera collects crop canopy images at intervals of 10 meters along the planting row. Each image contains a visible light band (400-700nm), a red edge band (700-740nm), and a near-infrared band (740-900nm). The resolution is set to 1024×1024 pixels, and the image acquisition time is fixed between 10:00 and 11:00 a.m. every day to avoid the influence of changes in illumination angle on spectral reflectance. The pixel-level normalized vegetation index NDVI is calculated for the red and near-infrared channels of the collected images using the formula NDVI=(NI R-Red) / (NIR+Red), the obtained value is mapped to a grayscale image for storage, and the mean and standard deviation of the NDVI index of each image are recorded in a structured table; then the color channel extraction tool is used to extract the main color tone of the leaf surface from the RGB channel data of the visible light image, and the color of each pixel is converted to the HSV color space. The proportion of pixels with a hue value between 90°–150° is counted to determine the green health level, and this proportion is set as the leaf color index; then image edge detection is performed, and the Canny operator is used to extract the leaf edge curve contour and calculate its curvature. If the mean edge curvature is greater than 0.25, it is marked as curl level 3. If it is between The values ​​between 0.1 and 0.25 are marked as level 2, otherwise they are marked as level 1 or 0, which is used to indicate the degree of leaf curling. In each band channel, the five-dimensional spectral reflectance index is extracted according to the change trend of pixel reflectance: red light reflection intensity, near-infrared reflection intensity, red edge jump amplitude, reflection difference mean and maximum reflection gradient. Finally, the NDVI index, leaf color index, curling level and five spectral reflectance values ​​are merged into complete crop physiological characteristic data. The crop physiological characteristic data of the same period in the same plot are selected, and the environmental measured data (including soil moisture, temperature, conductivity, light, air temperature and humidity) and virtual sensor node data of the corresponding time are synchronized. The spatial positions are matched by plot numbers and paired by timestamps. The environmental values ​​of the adjacent sensor nodes are matched at the center point of each image, and the distance-weighted average method (the radius is set to 10 meters) is used to fuse them into a single-point growth environment characteristic value. All image records of each plot are analyzed separately. If the NDVI index is lower than 0.6 and the green pixel ratio is lower than 70%, and the curling level is higher than 2, or any two of the five spectral reflectance values ​​deviate from the historical mean by more than 20%, the crop status during this period is evaluated as "abnormal", otherwise it is evaluated as "normal". Finally, the status assessment results, timestamps and key characteristic values ​​of each plot are summarized to form the crop physiological status assessment data.

[0065] By acquiring multispectral images encompassing visible light, red edge, and near-infrared bands, the present invention can comprehensively perceive the reflectance characteristics of crop canopies, accurately extract crop growth, color changes, and leaf structural status, and provide a rich foundation for the construction of physiological characteristic data. By extracting image features, the present invention obtains indicators such as the NDVI index, leaf color, and curling degree, which can intuitively reflect crop photosynthesis intensity, nutrient absorption, and stress response performance, addressing the defect that traditional water and fertilizer systems that rely solely on soil data cannot identify the true needs of crops. By integrating crop physiological characteristics with environmental monitoring data, a more targeted crop growth status expression system can be constructed, enabling water and fertilizer regulation to no longer be limited to soil-side parameter levels, but to extend to crop health status judgment, significantly improving the scientific nature of crop nutrition diagnosis and the accuracy of regulatory decisions. The final output crop physiological status assessment data has both timeliness and regionality, reflecting the physiological dynamics of individual crops while supporting differentiated fertilization strategies at the field scale, providing a key decision-making basis for tailored fertilization strategies based on local conditions and seedlings in precision agriculture.

[0066] Preferably, step S22 includes the following steps:

[0067] Step S221: Calculating the normalized pixel values ​​of the red light band and the near infrared band based on the crop multispectral image data to obtain NDVI index image data;

[0068] Step S222: extracting color features from the visible light band in the crop multispectral image data to obtain crop leaf color data;

[0069] Step S223: analyzing the edge texture and deformation curvature of the crop leaf surface area based on the crop multispectral image data, extracting leaf curling degree characteristics, and obtaining leaf curling degree data;

[0070] Step S224: performing statistical analysis on the reflectance changes of each band based on the crop multispectral image data, extracting multidimensional spectral response indicators, and obtaining crop spectral reflectance characteristic data;

[0071] Step S225: performing feature fusion processing based on the NDVI index image data, the crop leaf color data, the leaf curl degree data and the crop spectral reflectance feature data to obtain crop physiological feature data.

[0072] In the embodiment of the present invention, for the collected crop multispectral image data, the pixel values ​​of the red band (center wavelength 660nm) and the near infrared band (center wavelength 850nm) are extracted, and the normalized vegetation index formula NDVI = (NIR-Red) / (NIR+Red) is calculated point by point for each pixel. The result is output in the form of a grayscale image. The grayscale value range is limited to 0 to 255, and the corresponding NDVI value range is 0 to 1. All NDVI images are named with the plot number and stored in the data set; the visible light band (R, G, B) of the same image is calculated point by point. Perform color channel extraction, convert each pixel into the HSV color space, extract the mean, mode and distribution skewness of the H (hue) component, and use it to characterize the main color tone and color consistency of the crop leaf surface. Set the proportion of pixels with H values ​​in the range of 90-150 degrees as the green coverage index, and the proportion is accurate to two decimal places. First, enhance the image with an edge enhancement filter, and then use the Canny edge detection method to extract the leaf contour. The detected edge curve is recorded in coordinate form and the second-order derivative is calculated to obtain the local curvature of each edge. If the average curvature is greater than If the curling degree is 0.25, it is defined as severe curling, if it is between 0.15–0.25, it is moderate curling, and below 0.15 is light curling or flat state. The curling degree is divided into levels 0 to 3. The final leaf curling degree data contains the image number, area ID and corresponding curling level; the pixel reflectance of the five bands of red light, green light, blue light, red edge and near infrared are extracted respectively, and the mean, standard deviation and adjacent band reflectance difference of the reflection intensity of each band are calculated for each pixel point. The maximum reflection difference, average reflection gradient and red edge jump amplitude of the entire image are statistically analyzed to generate a spectrum. There are five response indicators in total, and the numerical precision is uniformly retained to two decimal places. The above data are integrated into a crop spectral reflectance characteristic data table; the above NDVI index image data, crop leaf color data, leaf curl degree data and crop spectral reflectance characteristic data are merged with the image number as the primary key, and a 9-dimensional crop physiological characteristic vector is uniformly constructed for each record. The 9 dimensions are NDVI mean, green coverage, hue skewness, curl level, red light reflection, red edge reflection, near-infrared reflection, maximum reflection difference and average gradient. All vector data are exported as a structured physiological characteristic dataset.

[0073] By systematically processing pixel information in different bands of crop multispectral images, the present invention can comprehensively extract crop physiological status characteristics from multiple dimensions, thereby improving the accuracy and dimensionality of crop growth identification; constructing NDVI index images by red and near-infrared bands, it can realize quantitative analysis of crop greenness and photosynthetic activity; color feature extraction reflects leaf pigment changes, which helps to judge nitrogen or chlorophyll deficiency; edge texture and curvature analysis reveals leaf curling trends, effectively reflecting the morphological response of crops to stresses such as drought and high temperature; statistical processing of the reflectivity of each band generates multidimensional spectral indicators, which can identify potential diseases, nutrient imbalances and other problems; and finally, the above-mentioned image-derived information is integrated into a unified crop physiological feature vector, which not only enhances the expression ability of crop health status, but also provides a decision-making basis based on the actual state of crops for precise water and fertilizer regulation, avoids misjudgment caused by a single parameter, and improves the data reliability and adaptability of the diagnosis process.

[0074] Preferably, step S3 includes the following steps:

[0075] Step S31: collecting pH value, conductivity and total dissolved solids of water sources in the farmland area in real time to obtain water quality monitoring data;

[0076] Step S32: dynamically adjusting the concentrations of nitrogen, phosphorus, and potassium solutions in the fertilizer solution based on the crop physiological status assessment data and the water quality monitoring data to obtain fertilizer solution ratio parameters;

[0077] Step S33: Control the operation of the multi-channel fertilizer pump based on the fertilizer-liquid ratio parameter to obtain the fertilizer control parameter.

[0078] In the embodiment of the present invention, an integrated water quality online monitoring module is first installed at the front end of the main water source inlet pipe of the farmland. The module includes a pH electrode sensor, a conductivity measurement probe and a TDS infrared detection element. The sensor sampling frequency is set to once every 5 minutes, and the measurement accuracy is ±0.1pH unit, ±2% conductivity reading (unit μS / cm) and ±5% respectively. TDS value (in mg / L). All sensors are connected to the central monitoring controller via the RS485 bus and upload sampling data in real time to the water and fertilizer control server as water quality monitoring data. Based on the status category marked in the crop physiological status assessment data corresponding to each farmland area, the area is divided into "normal" or "abnormal". If the status is "abnormal" and the pH value in the water quality monitoring data is less than 6.0 or greater than 7.5, the target concentration of the phosphorus solution is increased by 10% to 20%, and the increase value is determined by the degree of deviation. If the conductivity is lower than 400μS / cm, the target concentration of the nitrogen solution is increased by 15% accordingly. If the TDS value is higher than 900mg / L, the target concentration of the potassium solution is reduced by 10%. The original target concentrations are limited to 200mg / L for nitrogen solution, 100mg / L for phosphorus solution, and 150mg / L for potassium solution. The adjusted concentration variation range does not exceed ±30%. The obtained adjusted values ​​are output as fertilizer solution ratio parameter tables for each crop plot. The parameter table is indexed by plot number and contains the target concentrations of each fertilizer solution. The fertilizer control unit controls the operation of a multi-channel fertilizer pump based on the fertilizer-liquid ratio parameters. Each channel corresponds to a nitrogen, phosphorus, and potassium solution. Channel control uses PWM signals to precisely adjust the opening of the proportional solenoid valve to ensure real-time concentration accuracy within ±5% during the liquid mixing process. A three-channel flow sensor is installed in the main mixing pipe to detect the flow rates of the three solutions. Error correction is performed every 30 seconds between the actual and set ratios. The fertilizer pump operates by first introducing clean water for 3 minutes to establish a base flow, then opening each channel one by one. The injection time of each solution is controlled according to the flow rate converted from the concentration. The controller records the channel opening time, valve opening, and actual flow rate in real time. Fertilizer control parameters include the plot number, set concentration per channel (in mg / L), flow rate (in L / min), duration (in minutes), and the total amount of mixed solution. All parameters are recorded and directly output to the irrigation decision-making process for subsequent control of the irrigation pump and solenoid valve to perform specific operations.

[0079] By introducing real-time monitoring of the pH value, electrical conductivity and total dissolved solids content of irrigation water sources, the present invention can dynamically grasp the acidity and alkalinity of the water body, the salinity level and the concentration of dissolved matter, thereby providing accurate water quality background information for fertilizer liquid regulation, avoiding the decline in nutrient utilization or the increase in crop stress risk due to abnormal water quality; combined with the crop physiological status assessment data for joint analysis, targeted adjustment of the nitrogen, phosphorus and potassium solution concentrations is achieved, so that the fertilizer liquid ratio can accurately match the current crop nutritional needs and the environmental carrying capacity, and improve the timeliness and effectiveness of fertilizer input; the multi-channel fertilizer pump is driven by the fertilizer liquid ratio parameters to achieve automatic control and real-time output of the fertilizer liquid concentration and ratio, ensuring the continuity, accuracy and response speed of the fertilization operation, eliminating the problem of insufficient adaptability of the traditional fixed ratio method to environmental changes and crop heterogeneity, and fundamentally improving the intelligence level and resource allocation efficiency of the water and fertilizer regulation process.

[0080] Preferably, step S4 includes the following steps:

[0081] Step S41: Analyzing the environmental status of the area without sensors deployed in the farmland area based on the virtual sensor node data to obtain virtual environmental status data;

[0082] Step S42: Calculating the water and fertilizer requirements of the farmland area based on the fertilizer control parameters and the virtual environment state data to obtain water and fertilizer regulation demand data;

[0083] Step S43: Based on the water and fertilizer control demand data, intelligent decision analysis is performed on the irrigation duration, irrigation flow rate and fertilizer application amount to obtain control decision data.

[0084] In the embodiment of the present invention, the virtual sensor node data generated in step S17 is first called, and an environmental analysis is performed on each 0.5m×0.5m grid unit in all areas where no sensors are deployed. Six types of data are extracted from the area, namely soil temperature (in °C), soil moisture (in %), soil conductivity (in mS / cm), air temperature (in °C), air humidity (in %) and light intensity (in lux). Cluster statistics are performed according to spatial position, and the average and maximum values ​​of the six indicators are calculated within each 1-acre plot to generate the virtual environmental status data of the plot. According to the concentration values ​​of nitrogen, phosphorus and potassium solutions recorded in the fertilizer control parameters and the target volume of fertilizer solution for each plot, the current nutrient input demand of the crop is standardized and converted. At the same time, the soil moisture below 35% and the electrical conductivity below 0.3mS / cm are used as the thresholds for judging water shortage and fertilizer shortage. If any plot in the virtual environment meets any of the above two conditions and the duration exceeds 6 hours, it will be marked as requiring supplementary irrigation or fertilization. For areas that meet the conditions, the water requirement is calculated according to the degree of water shortage. The calculation method is to subtract the current soil moisture from the standard target soil moisture content of 40%, and multiply the result by the soil volume (0.3m per square meter).3 The required water volume (calculated) is the required water volume in liters. The fertilizer requirement is calculated by subtracting the current conductivity from the standard conductivity of 0.8 mS / cm, multiplied by the soil unit volume and the fertilizer solution concentration coefficient (15 mg / L for every 0.1 mS / cm). This ultimately yields the water and fertilizer control demand data for each acre of land, including the required water volume, fertilizer volume, and target concentrations of the three fertilizer solutions. Based on this water and fertilizer control demand data, the corresponding irrigation duration and irrigation flow rate are determined. The irrigation flow rate is fixed at 50 L / min based on the pipeline specifications. Irrigation duration = required water volume / 50. The fertilizer amount is proportionally converted to the fertilizer injection volume based on the required concentrations of the three liquids. The fertilizer amount is controlled to be gradually injected within 5 minutes before irrigation. Each fertilizer channel is injected in the order of phosphorus solution, potassium solution, and nitrogen solution, with a 1-minute interval. The controller outputs the irrigation duration (in minutes), irrigation flow (in L / min), and injection volume (in L) of the three fertilizer solutions for each plot as the final control decision data.

[0085] The present invention realizes comprehensive perception of the farmland environmental status by utilizing virtual sensor node data, fills the information gap caused by insufficient actual sensor deployment, and improves the accuracy and integrity of environmental monitoring; combines fertilizer control parameters to accurately calculate the water and fertilizer requirements of the farmland, effectively avoids the waste of water and fertilizer resources, and improves resource utilization efficiency; realizes scientific regulation of irrigation duration, flow rate and fertilizer amount through intelligent decision-making analysis, which not only ensures the optimal growth environment for crops, but also enhances the automation and intelligence level of irrigation and fertilization management, thereby promoting energy conservation, emission reduction, and increased production and efficiency in agricultural production.

[0086] Preferably, step S41 includes the following steps:

[0087] Step S411: spatially reconstructing soil temperature, soil moisture, and soil conductivity at each location based on the virtual sensor node data to obtain soil environmental state simulation data;

[0088] Step S412: performing time series analysis on the air temperature, air humidity, and light intensity at each location in the virtual sensor node data set to obtain meteorological environment state simulation data;

[0089] Step S413: Based on the soil environment state simulation data and the meteorological environment state simulation data, a soil-meteorological coupling analysis is performed on the environmental state of the farmland area where no sensors are deployed to obtain virtual environmental state data.

[0090] In the embodiment of the present invention, the virtual sensor node data generated in step S17 is first called to extract the soil temperature (in °C, ranging from 0-40), soil moisture (in %, ranging from 10-90), and soil conductivity (in mS / cm, ranging from 0.1-2.5) in all areas where no sensors are deployed. A regular grid of 0.5 m × 0.5 m is established for all data according to spatial position. The known nodes are used as interpolation cores, and spatial reconstruction is performed using the distance weighted average method, where the weighting coefficient is 1 / the square of the distance, the calculation cutoff radius is set to 30 meters, and the points beyond the range are not included in the calculation. Each grid cell outputs the estimated values ​​of the three types of soil parameters and is marked with a timestamp to form a complete soil environmental state simulation data set; air temperature (in °C, ranging from -10-45), air humidity (in %, ranging from 10-95), and light intensity (in lux, ranging from 1000-100000) are extracted from the same virtual sensor node data, and a 10-minute time interval is constructed for each grid cell. The time series of each node was analyzed. For each parameter, sliding statistical processing within 48 hours was performed, and four indicators (mean, maximum, minimum, and variation range) were calculated. For each node, 12 data items (three types of parameters × four indicators) were output, forming a meteorological environmental state simulation dataset. The spatially reconstructed soil environmental state data were matched one-to-one with the meteorological environmental state data after time series analysis according to the node location. A coupling analysis was performed on each grid cell. The analysis method was as follows: if the soil moisture was less than 35% and the light intensity was greater than 60,000 lux for 6 consecutive hours, it was marked as a high evaporation area; if the soil electrical conductivity was less than 0.3 mS / cm and the air humidity was less than 40%, it was marked as a nutrient loss area; and if the soil temperature was greater than 35°C and the air temperature was greater than 40°C, it was marked as a high temperature stress area. Each area could be assigned multiple labels, and all coupling analysis results were recorded as attribute fields attached to the corresponding spatial unit, ultimately forming virtual environmental state data containing environmental factor values ​​and environmental state labels.

[0091] By combining spatial reconstruction with time-series analysis, the present invention achieves multi-dimensional, dynamic simulation of the soil and meteorological environment in farmland areas, thereby improving the ability to accurately perceive environmental conditions. Through the coupled analysis of soil and meteorological data, it can more comprehensively and accurately reflect environmental changes in areas where sensors are not deployed, making up for monitoring blind spots and enhancing the integrity and reliability of environmental information. This high-precision environmental simulation provides a scientific basis for subsequent agricultural management decisions, helps to achieve precision agriculture, and improves the optimization of the crop growth environment and resource utilization efficiency.

[0092] Preferably, step S5 includes the following steps:

[0093] Step S51: performing on-off control on the solenoid valve based on the control decision data, and driving the irrigation channel to open and close, thereby obtaining solenoid valve control execution data;

[0094] Step S52: Based on the control decision data, the irrigation pump is started and stopped, and the irrigation flow rate and duration are adjusted to obtain the irrigation pump operation data;

[0095] Step S53: performing flow control and channel selection on the fertilizer pump based on the control decision data, and executing a fertilizing operation with a set fertilizer-liquid ratio to obtain operation data of the fertilizer pump;

[0096] Step S54: Record the entire fertigation process based on the solenoid valve control execution data, the irrigation pump operation data, and the fertilizer pump operation data to obtain fertigation and irrigation execution log data.

[0097] In an embodiment of the present invention, after receiving the control decision data generated in step S43, the central control unit calls the corresponding solenoid valve address according to the number of each plot, sends a switch signal through the RS485 bus, controls the on / off state of the solenoid valve, and the response time of the solenoid valve does not exceed 1 second. The control signal is triggered by relay logic and executed. Each on / off operation records the actual action time, target state, voltage and current data, and execution feedback state to generate solenoid valve control execution data; according to the irrigation time (in minutes) and irrigation flow (in L / min) corresponding to each plot, the variable frequency irrigation pump is automatically called, and the start sequence is to delay the trigger by 5 seconds after the solenoid valve is opened. The rated power of the irrigation pump is 1.5kW, and the frequency converter adjusts the pump operating frequency to be consistent with the set flow. The control mode is PID closed-loop regulation. When the pump is running, the flow meter provides feedback data every 10 seconds. If the deviation between the measured value and the set value for three consecutive times exceeds 5%, the frequency is automatically corrected. The time, current, voltage, operating frequency and cumulative flow of each start and stop are recorded. The amount is recorded as irrigation pump operation data; according to the ratio of nitrogen, phosphorus and potassium solutions set in the fertilizer control parameters, the control unit sends signals to control the solenoid valve opening and flow threshold of the three-channel electric fertilizer pump. Each channel is equipped with an independent flow sensor. When the fertilizer pump is started, the clean water channel is opened to pre-flush the pipeline for 30 seconds, and then the fertilizer solution is mixed and injected in proportion. The target concentration is determined by the pump flow and injection time. The error of each channel is controlled within ±5%. During the entire fertilization process, the operation time, flow value, actual opening and mixed liquid output of each channel of the fertilizer pump are recorded to form the fertilizer pump operation data; the solenoid valve control execution data, irrigation pump operation data and fertilizer pump operation data are merged and archived according to the timestamp to generate a complete fertilization and irrigation execution log data. The log includes the plot number, irrigation start and end time, fertilizer solution three-channel injection ratio, total irrigation flow, total fertilizer amount, total solenoid valve opening time, pump operation status curve and control instruction receipt status of each stage. The log data is stored in real time in the central control database.

[0098] The present invention realizes the precise control of the irrigation and fertilization process of each zone of the farmland, ensures the distribution of water and fertilizer resources on demand, effectively avoids water resource waste and nutrient loss, and improves the utilization rate of water and fertilizer; it collects and records the operating status data of the solenoid valve, irrigation pump and fertilizer pump in real time to form a complete and traceable fertilization and irrigation operation record, providing detailed data support for subsequent planting management, troubleshooting and production traceability; it links the control execution units according to the control decision data to achieve synchronous coordination of the fertilization and irrigation processes, and avoids the risk of water accumulation at the crop roots, nutrient leaching or fertilizer damage caused by unreasonable control timing; by precisely adjusting the flow rate and running time of the irrigation pump, it ensures the uniformity of irrigation and the suitability of moisture in the crop root zone, and improves the consistency and quality stability of crop growth; the real-time adjustment function of the multi-channel flow of the fertilizer pump ensures the accuracy of the fertilizer liquid ratio, so that the concentration of each nutrient component is always within the target setting range, meeting the nutritional needs of crops at different growth stages; the complete fertilization and irrigation execution log data can be linked with the farmland digital twin system to continuously optimize subsequent control decisions and improve the intelligence level and adaptive regulation capability of the entire water-fertilizer integrated system.

[0099] Preferably, step S6 includes the following steps:

[0100] Step S61: performing spatial mapping and parameter visualization modeling on the virtual sensor node data to obtain a three-dimensional environmental perception data layer;

[0101] Step S62: performing crop health status encoding and image layer rendering on the crop physiological status assessment data to obtain a crop status visualization layer;

[0102] Step S63: performing time series integration and color coding on the water quality monitoring data to obtain a water quality visualization layer;

[0103] Step S64: performing operation path tracking and operation record graphical processing on the fertilization and irrigation execution log data to obtain a fertilization and irrigation process layer;

[0104] Step S65: Based on the three-dimensional environmental perception data layer, the crop status visualization layer, the water quality visualization layer, and the fertilization and irrigation process layer, the smart farm spatial scene is three-dimensionally integrated to obtain a three-dimensional visualization smart farm twin dashboard;

[0105] Step S66: Based on the three-dimensional visualization of the smart farm twin dashboard, the spatial distribution status, crop health status and fertilization and irrigation behavior are integrated and presented, and visual interaction is designed to obtain a twin visual data view.

[0106] In the embodiment of the present invention, the virtual sensor node data generated in step S17 is spatially positioned at a resolution of 0.5 m × 0.5 m to construct a multidimensional grid dataset containing six parameters, namely soil temperature (°C), soil moisture (%), soil conductivity (mS / cm), air temperature and humidity (°C, %), and light intensity (lux). The OpenGL rendering engine is used to assign independent color channels to the six parameters, with soil moisture mapped to the blue channel, soil conductivity mapped to the green channel, and light intensity mapped to the yellow channel. The three-dimensional environmental perception data layer is superimposed and generated; the crop physiological status output in step S24 is loaded. Evaluation data, according to the NDVI index, curling level and green coverage rate to encode the health status of the crops in the plots. Areas with NDVI less than 0.5 or curling level 3 are marked in red, and areas with high green coverage are marked in green. The three types of status are rendered as crop status visualization layers in 0.5-meter grid units; the pH, conductivity and TDS values ​​every 10 minutes in the water quality monitoring data are formed into a time series. pH less than 6.0 or greater than 8.0 is marked as acidic or alkaline, with the corresponding colors being red or blue. Conductivity higher than 1500μS / cm is marked as high salt area, with the color being purple. TDS higher than 1000 mg / L is marked as a gray high solid area, and the dynamic change of water quality is rendered by scrolling the time axis to form a water quality visualization layer; based on the fertilization and irrigation execution log data generated in step S54, the solenoid valve control time, the irrigation pump start period and the three-channel fertilizer pump operation record are time-aligned, and the irrigation path line is drawn according to the plot number. The line width is adjusted according to the flow ratio. The fertilization process is displayed with a flow arrow animation, and floating annotation texts such as the fertilization start and end time, the ratio and the total amount of liquid applied are added to form a fertilization and irrigation process layer; the four layers are imported into a unified coordinate system using the CesiumJS three-dimensional geographic rendering engine for three-dimensional visualization. Overlay synthesis, set layer switching function and timeline linkage mechanism, build a three-dimensional smart farm space scene that integrates environment, crop, water quality and operation data, and generate a three-dimensional visual smart farm twin dashboard; set up a space status display panel, crop health status switch button, water quality abnormality warning pop-up window and fertilization animation playback control bar for the twin dashboard interface, and control the interaction logic through the layer mutual exclusion mechanism and layer linkage. When the user clicks on any plot, the interface displays the current environmental value, crop status and the most recent fertilization parameters. The final output of the twin dashboard is a twin visual data view that supports click query, time backtracking, status highlighting and spatial positioning.

[0107] The present invention realizes the three-dimensional visualization integration of farmland environmental status, crop health status, water quality changes and fertilization and irrigation processes, which makes it convenient for managers to fully grasp the real-time production status and historical change trends of various areas of farmland on the same interface, thereby improving the intuitiveness of planting management and the accuracy of scientific decision-making; through the mapping of spatial distribution data and visual interactive design, it can quickly identify abnormal areas, timely locate water-fertilizer imbalance, pest and disease risks or equipment failure locations, and shorten the problem response time; the multi-layer overlay display method helps to analyze the correlation between various environmental factors and crop growth status, and assist in optimizing water and fertilizer regulation plans and farming strategies; the graphical processing of fertilization and irrigation paths and operation records can be used for operation standardization evaluation and mechanical operation path optimization, reducing energy consumption and labor intensity; the time series coding display of water quality data provides an efficient tool for long-term water source safety monitoring and water quality trend analysis, which is convenient for formulating water resource protection and utilization plans; the overall visualization dashboard system realizes dynamic data update and multi-terminal remote access through real-time linkage with the Internet of Things platform, which significantly improves the management efficiency and intelligence level of the smart farm system.

[0108] Preferably, the present invention further provides an Internet of Things-based integrated water and fertilizer control system for executing the above-mentioned Internet of Things-based integrated water and fertilizer control method, wherein the Internet of Things-based integrated water and fertilizer control system comprises:

[0109] The twin modeling module is used to obtain measured environmental data, historical environmental data, and regional geographic information of farmland areas, and conduct spatial modeling and time series learning of farmland areas to build a digital twin model of farmland. The digital twin model of farmland is used to simulate the state of areas without sensor deployment to obtain virtual sensor node data.

[0110] The crop assessment module is used to collect multispectral image data of crops and extract crop physiological characteristic data; it uses environmental measurement data, virtual sensor node data and crop physiological characteristic data to build a crop physiological status model and generate crop physiological status assessment data;

[0111] The water and fertilizer control module is used to collect water quality monitoring data in real time; dynamically adjust the fertilizer-liquid ratio based on crop physiological status assessment data and water quality monitoring data, and control the operation of multi-channel fertilizer pumps to generate fertilizer control parameters;

[0112] Intelligent decision-making module, used to determine the control decision data of irrigation duration, flow rate and fertilizer amount based on virtual sensor node data and fertilizer control parameters;

[0113] An execution control module is used to drive the solenoid valve, irrigation pump, and fertilizer pump to perform irrigation and fertilization operations based on the control decision data, and generate fertilization and irrigation execution log data;

[0114] The visual twin module is used to build a three-dimensional visual smart farm twin dashboard based on virtual sensor node data sets, crop physiological status assessment data, water quality monitoring data, and fertilization and irrigation execution log data, and generate a twin visual data view.

[0115] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is not limited by the above description. Therefore, it is intended that all changes that fall within the meaning and scope of the equivalent elements of the application documents are included in the present invention.

[0116] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A water-fertilizer integrated control method based on the Internet of Things, characterized in that: The following steps are involved: Step S1: Obtaining measured environmental data, historical environmental data, and regional geographic information of the farmland area, and performing spatial modeling and time series learning on the farmland area to construct a farmland digital twin model; using the farmland digital twin model to simulate the state of the area without sensor deployment to obtain virtual sensor node data; Step S2: collecting multispectral image data of crops and extracting crop physiological characteristic data; Use environmental measured data, virtual sensor node data and crop physiological characteristic data to build a crop physiological state model and generate crop physiological state assessment data; Step S3: real-time collection of water quality monitoring data; Dynamically adjust the fertilizer-liquid ratio based on crop physiological status assessment data and water quality monitoring data, control the operation of multi-channel fertilizer pumps, and generate fertilizer control parameters; Step S4: Determine control decision data for irrigation duration, flow rate, and fertilizer amount based on virtual sensor node data and fertilizer control parameters; Step S5: driving the solenoid valve, irrigation pump, and fertilizer pump to perform irrigation and fertilization operations based on the control decision data, and generating fertilization and irrigation execution log data; Step S6: Build a three-dimensional visual smart farm twin dashboard based on the virtual sensor node data set, crop physiological status assessment data, water quality monitoring data, and fertilization and irrigation execution log data to generate a twin visual data view.

2. The water-fertilizer integrated control method based on the Internet of Things according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: collecting soil temperature, soil moisture, soil conductivity, air temperature and humidity, and light intensity data of the farmland area to obtain actual environmental measurement data; Step S12: Acquire historical environmental data of the farmland area; Step S13: collecting topography, boundaries and geographical element information of the farmland area and performing digital processing to obtain regional geographical information; Step S14: constructing a farmland regional spatial model based on the measured environmental data, historical environmental data and regional geographic information to obtain a farmland spatial structure model; Step S15: constructing a farmland environment time series based on the measured environmental data and historical environmental data to obtain a farmland time dynamic model; Step S16: constructing a farmland digital twin model based on the farmland spatial structure model and the farmland temporal dynamic model; Step S17: Simulate the state of the area without sensor deployment based on the farmland digital twin model to obtain virtual sensor node data.

3. The water-fertilizer integrated control method based on the Internet of Things according to claim 1, characterized in that: Step S2 includes the following steps: Step S21: Collect visible light, red edge, and near-infrared multispectral images of crop canopies to obtain crop multispectral image data; Step S22: extracting NDVI index, crop leaf color, curling degree, and spectral reflectance characteristics based on the crop multispectral image data to obtain crop physiological characteristic data; Step S23: Based on the measured environmental data, the virtual sensor node data and the crop physiological characteristic data, a fusion model of the crop growth environment and the crop state is performed to obtain a crop physiological state model; Step S24: Evaluate the current crop health and nutritional status based on the crop physiological status model to obtain crop physiological status evaluation data.

4. The water-fertilizer integrated control method based on the Internet of Things according to claim 3, characterized in that: Step S22 includes the following steps: Step S221: Calculating the normalized pixel values ​​of the red light band and the near infrared band based on the crop multispectral image data to obtain NDVI index image data; Step S222: extracting color features from the visible light band in the crop multispectral image data to obtain crop leaf color data; Step S223: analyzing the edge texture and deformation curvature of the crop leaf surface area based on the crop multispectral image data, extracting leaf curling degree characteristics, and obtaining leaf curling degree data; Step S224: performing statistical analysis on the reflectance changes of each band based on the crop multispectral image data, extracting multidimensional spectral response indicators, and obtaining crop spectral reflectance characteristic data; Step S225: performing feature fusion processing based on the NDVI index image data, the crop leaf color data, the leaf curl degree data and the crop spectral reflectance feature data to obtain crop physiological feature data.

5. The water-fertilizer integrated control method based on the Internet of Things according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: collecting pH value, conductivity and total dissolved solids of water sources in the farmland area in real time to obtain water quality monitoring data; Step S32: dynamically adjusting the concentrations of nitrogen, phosphorus, and potassium solutions in the fertilizer solution based on the crop physiological status assessment data and the water quality monitoring data to obtain fertilizer solution ratio parameters; Step S33: Control the operation of the multi-channel fertilizer pump based on the fertilizer-liquid ratio parameter to obtain the fertilizer control parameter.

6. The water-fertilizer integrated control method based on the Internet of Things according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: Analyzing the environmental status of the area without sensors deployed in the farmland area based on the virtual sensor node data to obtain virtual environmental status data; Step S42: Calculating the water and fertilizer requirements of the farmland area based on the fertilizer control parameters and the virtual environment state data to obtain water and fertilizer regulation demand data; Step S43: Based on the water and fertilizer control demand data, intelligent decision analysis is performed on the irrigation duration, irrigation flow rate and fertilizer application amount to obtain control decision data.

7. The water-fertilizer integrated control method based on the Internet of Things according to claim 6, characterized in that: Step S41 includes the following steps: Step S411: spatially reconstructing soil temperature, soil moisture, and soil conductivity at each location based on the virtual sensor node data to obtain soil environmental state simulation data; Step S412: performing time series analysis on the air temperature, air humidity, and light intensity at each location in the virtual sensor node data set to obtain meteorological environment state simulation data; Step S413: Based on the soil environment state simulation data and the meteorological environment state simulation data, a soil-meteorological coupling analysis is performed on the environmental state of the farmland area where no sensors are deployed to obtain virtual environmental state data.

8. The water-fertilizer integrated control method based on the Internet of Things according to claim 1, characterized in that: Step S5 includes the following steps: Step S51: performing on-off control on the solenoid valve based on the control decision data, and driving the irrigation channel to open and close, thereby obtaining solenoid valve control execution data; Step S52: Based on the control decision data, the irrigation pump is started and stopped, and the irrigation flow rate and duration are adjusted to obtain the irrigation pump operation data; Step S53: performing flow control and channel selection on the fertilizer pump based on the control decision data, and executing a fertilizing operation with a set fertilizer-liquid ratio to obtain operation data of the fertilizer pump; Step S54: Record the entire fertigation process based on the solenoid valve control execution data, the irrigation pump operation data, and the fertilizer pump operation data to obtain fertigation and irrigation execution log data.

9. The water-fertilizer integrated control method based on the Internet of Things according to claim 1, characterized in that: Step S6 includes the following steps: Step S61: performing spatial mapping and parameter visualization modeling on the virtual sensor node data to obtain a three-dimensional environmental perception data layer; Step S62: performing crop health status encoding and image layer rendering on the crop physiological status assessment data to obtain a crop status visualization layer; Step S63: performing time series integration and color coding on the water quality monitoring data to obtain a water quality visualization layer; Step S64: performing operation path tracking and operation record graphical processing on the fertilization and irrigation execution log data to obtain a fertilization and irrigation process layer; Step S65: Based on the three-dimensional environmental perception data layer, the crop status visualization layer, the water quality visualization layer, and the fertilization and irrigation process layer, the smart farm spatial scene is three-dimensionally integrated to obtain a three-dimensional visualization smart farm twin dashboard; Step S66: Based on the three-dimensional visualization of the smart farm twin dashboard, the spatial distribution status, crop health status and fertilization and irrigation behavior are integrated and presented, and visual interaction is designed to obtain a twin visual data view.

10. A water-fertilizer integrated control system based on the Internet of Things, characterized in that: Used to execute the water-fertilizer integrated control method based on the Internet of Things according to claim 1, the water-fertilizer integrated control system based on the Internet of Things includes: The twin modeling module is used to obtain measured environmental data, historical environmental data, and regional geographic information of farmland areas, and conduct spatial modeling and time series learning of farmland areas to build a digital twin model of farmland. The digital twin model of farmland is used to simulate the state of areas without sensor deployment to obtain virtual sensor node data. The crop assessment module is used to collect multispectral image data of crops and extract crop physiological characteristic data; it uses environmental measurement data, virtual sensor node data and crop physiological characteristic data to build a crop physiological status model and generate crop physiological status assessment data; The water and fertilizer control module is used to collect water quality monitoring data in real time; dynamically adjust the fertilizer-liquid ratio based on crop physiological status assessment data and water quality monitoring data, and control the operation of multi-channel fertilizer pumps to generate fertilizer control parameters; Intelligent decision-making module, used to determine the control decision data of irrigation duration, flow rate and fertilizer amount based on virtual sensor node data and fertilizer control parameters; An execution control module is used to drive the solenoid valve, irrigation pump, and fertilizer pump to perform irrigation and fertilization operations based on the control decision data, and generate fertilization and irrigation execution log data; The visual twin module is used to build a three-dimensional visual smart farm twin dashboard based on virtual sensor node data sets, crop physiological status assessment data, water quality monitoring data, and fertilization and irrigation execution log data, and generate a twin visual data view.

Images