Agricultural ecological environment monitoring method and system based on digital twin

Abstract

The present invention relates to the field of ecological monitoring technology, specifically to an agricultural ecological environment monitoring method and system based on digital twins, comprising the following steps: acquiring monitoring data within farmland, constructing a mapping relationship, performing time series arrangement, merging multi-source data, calling a virtual grid to update a three-dimensional model, generating mapping results, analyzing crop growth images, tracking environmental changes, comparing ecological stability intervals, extracting risk trends, calling a three-dimensional model, and generating a risk warning interface layer. In the present invention, by acquiring continuous data of the agricultural ecological environment and constructing a mapping relationship to arrange the information in time series, the response speed and processing accuracy to changes in the agricultural ecological environment are improved, the monitoring and management of the crop growth environment are optimized, and the real-time updated three-dimensional environmental model is used to intuitively display environmental changes. By comparing humidity and temperature changes, spatial fragments are screened, and detailed monitoring and accurate warning of environmental changes are strengthened.

Description

Technical Field

[0001] The present invention relates to the field of ecological monitoring technology, and in particular to an agricultural ecological environment monitoring method and system based on digital twins. Background Art

[0002] The field of ecological monitoring technology encompasses the monitoring and management of various elements within the agricultural ecological environment, including the tracking and analysis of factors such as soil quality, climate change, crop growth, and irrigation conditions. The core of this field is the comprehensive, real-time monitoring and data collection of the agricultural ecological environment through sensors, remote sensing technology, and information and communication technologies, ensuring effective control of various environmental factors influencing agricultural production. Agricultural ecological and environmental monitoring technology encompasses not only the monitoring of natural environmental elements such as soil, air, and water quality, but also the potential environmental impacts of waste generated by agricultural production activities and pesticide use. The collection, processing, and analysis of monitoring data can provide a scientific basis for agricultural production management and help formulate sustainable agricultural development strategies.

[0003] Among them, the agricultural ecological environment monitoring method refers to a technical solution for dynamically monitoring and analyzing various environmental factors in the agricultural ecological environment through specific technical means. The technical matters targeted by this technical subject cover agricultural soil quality monitoring, crop growth status, changes in meteorological conditions, etc. By using sensors to collect environmental data and transmitting it through wireless networks, combined with data storage and processing technologies, the collected data can be monitored and analyzed in real time. This patented solution mainly obtains various environmental data through automated monitoring equipment and transmits the data to the monitoring platform for analysis and processing through data transmission, thereby achieving real-time supervision and early warning of the agricultural ecological environment.

[0004] Existing technologies are insufficient for dynamic and refined management of the agricultural ecosystem. While traditional technologies can collect data through sensors and remote sensing, they lack efficiency and accuracy in data integration, real-time updates, and the deep integration of multi-source information. This limitation results in inadequate and inaccurate responses to changes in the agricultural ecosystem, making it impossible to fully utilize collected data for efficient decision-making, which in turn impacts crop yield and quality. Without real-time updates and accurate analysis of environmental data, responses to sudden climate changes and abnormal environmental conditions are delayed or inadequate, resulting in losses in both agricultural production efficiency and environmental protection measures. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the existing technology and propose an agricultural ecological environment monitoring method and system based on digital twins.

[0006] In order to achieve the above objectives, the present invention adopts the following technical solution, which is an agricultural ecological environment monitoring method based on digital twins, including the following steps:

[0007] S1: Acquire continuous data from soil moisture monitors, temperature sensors, and meteorological data collection devices deployed in the farmland area, establish corresponding mapping relationships through sensor number identification, arrange the collected information in time sequence, and generate a synchronized scene image of the ecological environment;

[0008] S2: Based on the spatial grid coordinates corresponding to the humidity, temperature, and meteorological conditions recorded in the synchronized scene image of the ecological environment, the regional virtual grid built into the digital twin is called, the real-scene sensor synchronization information is connected, and the regional three-dimensional environmental model representation is updated according to the real-time environmental information to generate the farmland area mapping result;

[0009] S3: Based on the synchronous state trajectory of the grid entities in the farmland area mapping results, by comparing the change gradients of humidity and temperature between different time points, screening spatial segments with continuous fluctuation amplitudes, analyzing the crop growth images within the spatial segments, and generating an environmental response sensitivity distribution map;

[0010] S4: Call the block node number in the environmental response sensitive distribution map, track the synchronous evolution process of the humidity drop segment and the temperature rise node, compare the segment difference of the change amplitude curve with the set ecological stability interval, and generate the agricultural ecological environment risk trend identification result.

[0011] As a further solution of the present invention, the ecological environment synchronized scene image includes spatial feature grids, time-series alignment segments, and multi-source data fusion areas; the farmland area mapping results include grid number status sets, real-time environmental parameter groups, and three-dimensional structure representation sets; the environmental response sensitive distribution map includes parameter change highlighting areas, crop status response areas, and environmental factor interference zones; the agricultural ecological environment risk trend identification results include offset area numbers, trend curve changes, and ecological indicator change segments.

[0012] As a further solution of the present invention, the steps for acquiring the ecological environment synchronized scene image are specifically as follows:

[0013] S111: Acquire continuous data from soil moisture monitors, temperature sensors, and meteorological data collection devices deployed in the farmland area, add number identifiers, evaluate the mapping relationship between device numbers and data records, sort the data records by timestamp field, and generate a sensor time series number dataset;

[0014] S112: calling the data records corresponding to the device numbers in the sensor time series number data set, screening the soil moisture, temperature, and meteorological data entries with overlapping timestamps, performing a pairing operation on the temporally overlapping data according to the spatial coordinates of the data records, and combining the data records based on the spatial coordinate matching results as a judgment condition to generate a list of overlapping spatiotemporal joint values;

[0015] S113: According to the spatial coordinate field and time field in the overlapping spatiotemporal joint value list, the coordinates are used as image pixel numbers, the time field is mapped to the image frame sequence number, and each type of value is mapped to a color channel according to the pixel ratio to generate an ecological environment synchronization scene image.

[0016] As a further solution of the present invention, the steps for obtaining the farmland area mapping result are specifically as follows:

[0017] S211: Using the humidity, temperature, and meteorological condition data in the synchronized scene image of the ecological environment, calling the regional virtual grid coordinates built into the digital twin, evaluating the mapping relationship between the grid coordinates and the sensor data, and generating a gridded environmental parameter set;

[0018] S212: Based on the gridded environmental parameter set, soil hydraulic conductivity, crop coefficient, and evapotranspiration are extracted from the multi-region parameter mapping information using the formula:

[0019] ;

[0020] Calculate the environmental volatility, compare the environmental volatility with the material response threshold of the regional 3D model, and generate the regional dynamic response scalar;

[0021] in, is the environmental volatility, represents the temperature measurement value, represents the grid weight factor, represents the measured value of soil hydraulic conductivity, represents the measured mean evapotranspiration, Represents the quantitative value of meteorological conditions, represents the crop growth stage coefficient, Represents the humidity percentage value;

[0022] S213: Calling the regional dynamic response scalar, updating the texture resolution of the three-dimensional environment model according to the light intensity and wind speed in the real-time environmental information, and generating a farmland area mapping result.

[0023] As a further solution of the present invention, the step of obtaining the environmental response sensitive distribution map is specifically as follows:

[0024] S311: Based on the synchronization state trajectory of the grid entity in the farmland area mapping result, the humidity and temperature records of the grid unit in the time dimension are called, the fluctuation degree of the humidity change rate and the temperature change rate in the continuous time period are compared, and whether the continuous fluctuation amplitude exceeds the joint standard of the humidity fluctuation threshold and the temperature fluctuation threshold is determined, and the spatial segments that meet the joint standard conditions are selected to generate a continuous fluctuation segment set;

[0025] S312: Calling the spatial position pointed to by the continuous fluctuation segment set, performing image segmentation on the crop growth image, extracting the crop entity boundary in the image, and calculating the boundary morphological feature value and color distribution value, determining whether there is a spatial offset phenomenon of crop structure change or color migration within the segment, and obtaining the crop offset feature distribution value;

[0026] S313: Using the crop offset characteristic distribution value, calling the humidity change rate and temperature change rate corresponding to the continuous fluctuation segment set, performing difference calculation and combined judgment on the crop offset value and the environmental variable change rate, using the formula:

[0027] ;

[0028] Calculate the environmental response difference coefficient of the spatial segment, evaluate the correlation between the segment position and the environmental response difference coefficient, and obtain the environmental response sensitivity distribution map;

[0029] in, Representative The environmental response variance coefficient of the fragment, Representative The crop offset feature distribution value of the fragment, For the The humidity change rate of the segment, For the The temperature change rate of the segment, For the Clip crop image Continuity factor for segment boundaries, is the number of boundary segments in the segment image.

[0030] As a further solution of the present invention, the steps for obtaining the agricultural ecological environment risk trend identification results are specifically as follows:

[0031] S411: Calling the block node number in the environmental response sensitivity distribution map, extracting the humidity change data and temperature change data corresponding to the block, comparing the humidity decrease rate and the temperature increase rate node by node, analyzing the synchronous change trend of the humidity change rate and the temperature change rate, and obtaining a synchronous change trend comparison value;

[0032] S412: Based on the synchronous change trend control value, extract the continuous segments of the humidity drop rate and the temperature rise rate, divide the fluctuation range into sections according to the sections, and compare the section differences with the set ecological stability range using the formula:

[0033] ;

[0034] Calculate the regional difference rate, classify and judge the regional difference rate and the interval stability setting boundary, record the consecutive numbered segments that exceed the stable interval, and obtain the trend deviation area index sequence;

[0035] in, represents the regional difference rate, Indicates the The humidity change rate of the node, Indicates the The temperature change rate of the node, Indicates the Humidity fluctuation value of the node, Indicates the The temperature fluctuation value of the node, Indicates the The humidity of the node is extremely poor, Indicates the The temperature difference at the node is extremely low. Indicates the number of nodes in the segment;

[0036] S413: Based on the trend offset region index sequence, track the level number of the indicated region, detect the fluctuation trend distribution characteristics of humidity and temperature in the region, compare the fluctuation direction and intensity, and obtain the agricultural ecological environment risk trend identification result.

[0037] As a further solution of the present invention, the method further includes step S5:

[0038] S5: Based on the agricultural ecological environmental risk trend identification results, the three-dimensional environmental model in the digital twin interface is called, the meteorological condition animation layer and the soil status are superimposed, and it is determined whether the boundary change range of the warning block is continuously expanding. The highlighted area and the visible range of the interface are marked to obtain the agricultural ecological risk warning interface layer;

[0039] The agricultural ecological risk warning interface layer includes a dynamic boundary layer, a highlighted visible area, and a risk level overlay map.

[0040] As a further solution of the present invention, the steps for obtaining the agricultural ecological risk warning interface layer are specifically as follows:

[0041] S511: Based on the agricultural ecological environment risk trend identification results, the three-dimensional environmental model in the digital twin interface is called, and the time series data of wind speed, rainfall, and temperature are loaded into the interface. The soil moisture value, soil salinity value, and soil organic matter concentration layers are superimposed to obtain the three-dimensional layer fusion coordinate distribution value;

[0042] S512: Based on the fused coordinate distribution values ​​of the three-dimensional layer, the difference between the meteorological layer parameter and the soil state layer parameter of the marked area in the continuous frames of the real time and the previous time period is obtained, and the block boundary change trend is judged according to the time axis of the difference sequence. The continuous frame blocks with outward expansion boundaries are marked as expansion areas, and the continuous boundary expansion recognition result is obtained;

[0043] S513: Based on the continuous boundary expansion identification results, the spatial coordinate set of the changed area is extracted in the digital twin interface, and the coordinates are compared within the visible range. The width and color parameters of the highlighted block boundary line are set, and the areas that meet the expansion judgment conditions are highlighted in the interface. The layer name and display order are set to generate the agricultural ecological risk warning interface layer.

[0044] The agricultural ecological environment monitoring system based on digital twin is used to implement the above-mentioned agricultural ecological environment monitoring method based on digital twin, and the system includes:

[0045] The index recognition module acquires continuous data from soil moisture monitors, temperature sensors, and meteorological data collection devices deployed in the farmland area, establishes corresponding index relationships according to sensor numbers, filters data groups with overlapping time intervals, and generates synchronized scene images of the ecological environment.

[0046] The virtual mapping module calls the regional virtual grid point numbers in the digital twin environment based on the calibrated spatial grid coordinates in the synchronized scene image of the ecological environment, compares the humidity values, temperature values ​​and meteorological parameter values ​​between the grid points, updates the environmental performance of the corresponding nodes in the three-dimensional model, and obtains the farmland area mapping result;

[0047] The change recognition module calls the humidity change sequence and temperature change sequence of the virtual nodes in the farmland area mapping result, extracts the continuously changing spatial segments according to the change values ​​between adjacent time points, and generates an environmental response sensitivity distribution map;

[0048] The trend analysis module extracts the humidity change trajectory and temperature change trajectory of the node according to the spatial node number identified by the environmental response sensitivity distribution map, identifies the change curves of the two change trajectories in the node, judges the direction and magnitude of the change trend, and generates an agricultural ecological environment risk trend identification result;

[0049] The risk warning module calls the spatial level number recorded in the agricultural ecological environment risk trend identification result, matches the farmland area boundary node in the three-dimensional model, superimposes the meteorological parameter layer and soil moisture layer under the area, judges the boundary changes of the area corresponding to the warning number, and obtains the agricultural ecological risk warning range layer.

[0050] Compared with the prior art, the advantages and positive effects of the present invention are:

[0051] In the present invention, by acquiring continuous data of the agricultural ecological environment and constructing a mapping relationship to arrange the information in time series, and further merging and processing multi-source data with overlapping time intervals, the integration and application of data are made more efficient and accurate. Continuous data collection and composite analysis improve the response speed and processing accuracy to changes in the agricultural ecological environment, and optimize the monitoring and management of the crop growth environment. Through the real-time updated three-dimensional environmental model, environmental changes can be intuitively displayed, and the basis and accuracy of decision-making can be improved. By comparing humidity and temperature changes, screening spatial fragments, and generating environmental response sensitive distribution maps, detailed monitoring and accurate early warning of environmental changes are strengthened, providing a dynamic, multi-dimensional monitoring perspective with significant improvements in both spatial and temporal resolution, greatly enhancing the effectiveness of risk warnings and operational adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a schematic diagram of the workflow of the present invention;

[0053] Figure 2 This is a flow chart for obtaining an ecological environment synchronization scene image in the present invention;

[0054] Figure 3 This is a flow chart for obtaining farmland area mapping results in the present invention;

[0055] Figure 4 This is a flow chart for obtaining an environmental response sensitive distribution map in the present invention;

[0056] Figure 5 This is a flow chart for obtaining the agricultural ecological environment risk trend identification results in the present invention;

[0057] Figure 6 This is a flow chart for obtaining the agricultural ecological risk warning interface layer in the present invention. DETAILED DESCRIPTION

[0058] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0059] In the description of the present invention, it should be understood that the terms "length," "width," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0060] See also Figure 1 The present invention provides a technical solution, an agricultural ecological environment monitoring method based on digital twins, comprising the following steps:

[0061] S1: Acquire continuous data from soil moisture monitors, temperature sensors, and meteorological data collection devices deployed in the farmland area, establish corresponding mapping relationships through sensor number identification, arrange the collected information in time sequence, merge and process multi-source data with time overlap according to spatial coordinates, and generate a synchronized scene image of the ecological environment;

[0062] S2: Based on the spatial grid coordinates corresponding to the humidity, temperature, and meteorological conditions recorded in the synchronized scene image of the ecological environment, the regional virtual grid built into the digital twin is called, the synchronization information of the real-scene sensor is connected, the multi-region parameter mapping information is identified, and the regional three-dimensional environmental model representation is updated according to the real-time environmental information to generate the farmland area mapping result;

[0063] S3: Based on the synchronous state trajectory of the grid entities in the farmland area mapping results, by comparing the change gradients of humidity and temperature between different time points, screening spatial segments with continuous fluctuation amplitudes, analyzing the crop growth images within the spatial segments, locating the crop and environmental node blocks with synchronous differences, and generating an environmental response sensitivity distribution map;

[0064] S4: Call the block node number in the environmental response sensitivity distribution map, track the synchronous evolution process of the humidity drop segment and the temperature rise node, compare the segment difference of the change amplitude curve with the set ecological stability interval, extract the regional level number of the trend deviation, and generate the agricultural ecological environment risk trend identification result;

[0065] S5: Based on the results of agricultural ecological environmental risk trend identification, the three-dimensional environmental model in the digital twin interface is called, and the meteorological condition animation layer and soil status are superimposed to determine whether the boundary change range of the warning block is continuously expanding. The highlighted area and visible range of the interface are marked to obtain the agricultural ecological risk warning interface layer;

[0066] The synchronized scene images of the ecological environment include spatial feature grids, time-series alignment segments, and multi-source data fusion areas. The farmland area mapping results include grid number status sets, real-time environmental parameter groups, and three-dimensional structure representation sets. The environmental response sensitive distribution map includes parameter change highlight areas, crop status response areas, and environmental factor interference zones. The agricultural ecological environment risk trend identification results include offset area numbers, trend curve changes, and ecological indicator change segments. The agricultural ecological risk warning interface layers include dynamic boundary layers, highlighted visible areas, and risk level overlay maps.

[0067] See also Figure 2 ,The specific steps for acquiring the ecological environment synchronization scene image are:

[0068] S111: Acquire continuous data from soil moisture monitors, temperature sensors, and meteorological data collection devices deployed in the farmland area, add number identifiers, evaluate the mapping relationship between device numbers and data records, sort the data records by timestamp field, and generate a sensor time series number dataset;

[0069] When obtaining continuous data from soil moisture monitors, temperature sensors, and meteorological collection devices deployed in farmland areas, it is necessary to clarify the deployment coordinates and numbers of various sensors, set the soil moisture monitor setting number to SM003, the temperature sensor number to TP001, and the meteorological collection device number to WS001~WS002, read the continuous data content reported by each device, and add the device number identifier to each data item. Then, use the number as the primary key to establish a mapping structure between the number and the data record. During the reading process, first read the data content according to the sensor type, and synchronize the data according to the data reporting cycle set for each device. Set the humidity monitor to report one data piece at a period of 30 minutes, the temperature sensor to 15 minutes, and the meteorological collection device to 60 minutes. In the synchronized data structure, use a unified timestamp format as the alignment benchmark, standardize the upload time field of each device to the UTC+8 time zone time format "YYYY-MM-DDHH:mm:ss", and then synchronize the records according to the time format. This field is sorted in ascending time order to generate a time-series numbered dataset containing the mapping relationship between the number, data item, and timestamp, so that subsequent association operations can accurately index and match data. For example, for the record with number SM001 and collection time of 10:00:00 on April 16, 2025, its data value is soil moisture 22.5%, then the mapping entry is {device number: SM001, timestamp: 2025-04-16 10:00:00, soil moisture: 22.5} Similarly, for the temperature data collected by TP002 at the same time point of 18.7℃, the corresponding mapping entry is {device number: TP002, timestamp: 2025-04-1610:00:00, temperature: 18.7}. After processing, the data will be arranged in ascending order of time. If multiple devices upload data at 10:00:00, 10:15:00, and 10:30:00 respectively, all data rows will be arranged in the above time order to generate the sensor time series number dataset.

[0070] S112: Calling the data records corresponding to the device numbers in the sensor time series number dataset, screening the soil moisture, temperature, and meteorological data entries with overlapping timestamps, performing a pairing operation on the temporally overlapping data based on the spatial coordinates of the data records, and combining the data records based on the spatial coordinate matching results as a judgment condition to generate a list of overlapping spatiotemporal joint values;

[0071] Through the device number index corresponding to all time series data, and then the timestamp field intersection screening of different types of sensor data, set the reading device SM002, TP001 and WS001 to co-exist in the time series records. If the three have data entries at 10:00:00, 10:30:00, and 11:00:00 on April 16, 2025, the data with the intersection timestamp is selected as the candidate set, and compared based on its spatial coordinate information to determine whether the three sets of data come from For sampling points with a geographical distance of no more than 10 meters, if the conditions are met, the three types of data are combined into a joint record, and the coordinates of the three are set as (35.712, 120.991), (35.713, 120.990), and (35.712, 120.992). Then, the distance calculated by comparing their longitude and latitude is [(35.712-35.713)²+(120.991-120.990)²]≈0.0014 degrees≈156 meters. Since it exceeds the threshold, it is determined that their spatial positions are inconsistent. , this group of data does not participate in the combination operation; on the contrary, if the distance between the three is less than or equal to 10 meters, they will be combined into a spatiotemporal joint value record. Set the time point at 10:30:00, SM002=21.9%, TP001=17.6℃, WS001=wind speed 3.2m / s, rainfall 1.5mm, then the corresponding combined record is {timestamp: 2025-04-1610:30:00, coordinates: (35.712, 120.991), humidity: 21.9%, temperature: 17.6℃, Wind speed: 3.2m / s, rainfall: 1.5mm}. This type of combination records joint data with the same coordinate timestamp. The accuracy of the spatial coordinates is based on six decimal places to ensure that the spatial consistency during the matching judgment process can be controlled within an accuracy range of ≤5 meters. At the same time, the spatial threshold during the judgment process needs to be set based on the sensor deployment density. For example, in this example, one collection unit is deployed for every 30 square meters of farmland, so the spatial matching judgment threshold is preferably set to 6 meters to avoid mixing data from different plots and forming overlapping spatiotemporal joint value lists.

[0072] Table 1. Results of multi-sensor screening at overlapping time points:

[0073]

[0074] As shown in Table 1, the spatial intersection matching condition is only met between the data at 10:00:00 and 11:00:00, so only two joint records are generated for the next image conversion process.

[0075] S113: Based on the spatial coordinate field and the time field in the overlapping spatiotemporal joint value list, the coordinates are used as image pixel numbers, the time field is mapped to an image frame sequence number, and each type of value is mapped to a color channel according to the pixel ratio to generate an ecological environment synchronized scene image;

[0076] The latitude and longitude in the spatial coordinates are integerized and the pixel position is calculated. Set (35.712, 120.991) to the 35712th row and 120991th column position in the image as the image pixel number. The time field 2025-04-16 10:00:00 is uniformly formatted as "frame number 0001", and the frame numbers 0002 and 0003 are calculated backward every 30 minutes. And so on. The humidity, temperature, wind speed, and rainfall values ​​contained in the joint record need to be converted to RGB channels through pixel value mapping. Set the humidity (0%) to be mapped to the R channel, the temperature (-20℃) to be mapped to the G channel, and the wind speed (20 m / s) or rainfall (100mm) is mapped to the B channel, and the values ​​need to be linearly converted. The channel value corresponding to the temperature of 17.6℃ is set to G=(17.6+20) / 70×255≈138, the humidity of 21.9% corresponds to R=21.9 / 100×255≈56, and the rainfall of 1.5mm corresponds to B=1.5 / 100×255≈4. The pixel point RGB(56, 138, 4) is combined and filled into the pixel position of the 35712th row and the 120991th column in the image frame number 0001 to form an image sequence data set composed of time-advanced, multi-channel values. Each frame in this dataset represents an ecological image frame constructed from spatial monitoring data at a certain time point. The changes between frames reflect the fluctuation process of environmental factors. The image generation process adopts a two-dimensional pixel matrix structure. The dimension of each frame can be set to 2000 rows × 2000 columns, corresponding to a coverage area of ​​approximately 4 square kilometers. Pixels are sparsely filled according to the sensor deployment range. The image output format is set to PNG format. The frame sequences are uniformly named Frame-0001.png and Frame-0002.png to facilitate timeline sorting and processing, and generate synchronized scene images of the ecological environment.

[0077] See also Figure 3 ,The specific steps for obtaining the farmland area mapping results are:

[0078] S211: Using humidity, temperature, and meteorological condition data from the synchronized scene image of the ecological environment, the built-in regional virtual grid coordinates of the digital twin are called, the mapping relationship between the grid coordinates and the sensor data is evaluated, and a gridded environmental parameter set is generated;

[0079] Real-time data was collected by deploying three monitoring points (numbered M1-M3) in the farmland area. Monitoring point M1 recorded a humidity of 65%, a temperature of 28°C, and cloudy weather conditions (code C=2), M2 recorded a humidity of 72%, a temperature of 25°C, and sunny weather conditions (C=1), and M3 recorded a humidity of 58%, a temperature of 30°C, and overcast weather conditions (C=3). The coordinates of the monitoring points were converted into virtual grid coordinates G1 (boundary density 0.8), G2 (boundary density 1.2), and G3 (boundary density 0.5). The humidity, temperature, and weather codes were bound to the corresponding grid coordinates, generating a structured dataset containing grid coordinates, environmental parameters, and boundary density. The output was a gridded environmental parameter set.

[0080] Table 2: Sensor data and grid mapping table:

[0081]

[0082] As shown in Table 2, the boundary density is obtained by calculating the ratio of the grid side length to the number of adjacent grids. Assuming the side length of G1 is 10 meters and the number of adjacent grids is 8, the boundary density is 10 / (8×1.25)=0.8.

[0083] S212: Based on the gridded environmental parameter set, soil hydraulic conductivity, crop coefficient, and evapotranspiration are extracted from the multi-region parameter mapping information using the formula:

[0084] ;

[0085] Calculate the environmental volatility, compare the environmental volatility with the material response threshold of the regional 3D model, and generate the regional dynamic response scalar;

[0086] in, is the environmental volatility, represents the temperature measurement value, represents the grid weight factor, represents the measured value of soil hydraulic conductivity, represents the measured mean evapotranspiration, Represents the quantitative value of meteorological conditions, represents the crop growth stage coefficient, Represents the humidity percentage value;

[0087] The soil hydraulic conductivity, crop coefficient and evapotranspiration were extracted from the multi-region parameter mapping information. For the G1 grid in Table 1, the soil hydraulic conductivity AK was measured to be 0.15 cm / h by the ring knife method, the crop coefficient AF was set to 1.2 according to the corn heading period, the evapotranspiration AE was calculated to be 4.2 mm / d by the Penman formula, and the meteorological condition code AC was quantified as 0.9, 1.1 and 1.3 according to sunny (1), cloudy (2) and overcast (3), respectively. The grid weight factor AW was calculated by the formula AW = boundary density × 0.5 + 0.3. The AW of G1 was set to 0.8 × 0.5 + 0.3 = 0.7 and substituted into the formula.

[0088] Substitute the data and calculate the numerator:

[0089] AT=28, AW=0.7, AK=0.15→28×0.7×0.15=2.94;

[0090] Calculate the denominator:

[0091] AE=4.2, AC=1.1→ ;

[0092] Compute the absolute value term:

[0093] AF=1.2, AH=65→|1.2-65 / 100|=|1.2-0.65|=0.55;

[0094] Substitute into the formula for calculation:

[0095] ;

[0096] The environmental volatility Compared with the material response threshold (preset to 0.8), since 0.75<0.8, it is determined that the dynamics of the grid environment is within the limit, and the generated regional dynamic response scalar value is 0.75. The formula enhances the impact of spatial heterogeneity on the dynamic index by introducing the boundary density weight factor and meteorological quantization value, avoiding the error caused by traditional homogenization.

[0097] S213: calling the regional dynamic response scalar, updating the texture resolution of the three-dimensional environment model according to the light intensity and wind speed in the real-time environmental information, and generating the farmland area mapping result;

[0098] Based on the light intensity and wind speed in the real-time environmental information, taking the G1 grid as an example, the light intensity L measured by the photon sensor is 30,000 lux, and the wind speed V measured by the ultrasonic anemometer is 2.5 m / s. The texture resolution of L is increased by 1 level for every 10,000 lux, and the particle motion speed is increased by 0.2 units for every 1 m / s of V. Based on S=0.75 (lower than the threshold of 0.8), a conservative update strategy is adopted: the texture resolution is adjusted to (30,000 / 10,000)×0.75=2.25 levels, and the particle motion speed is set to 2.5×0.2×0.75=0.375 units. The output is a farmland area mapping result consistent with the physical environment.

[0099] See also Figure 4 ,The specific steps for obtaining the environmental response sensitivity distribution map are:

[0100] S311: Based on the synchronization state trajectory of the grid entities in the farmland area mapping result, the humidity and temperature records of the grid cells in the time dimension are called, and the fluctuation degree of the humidity change rate and the temperature change rate in the continuous time period is compared. It is determined whether the continuous fluctuation amplitude exceeds the joint standard of the humidity fluctuation threshold and the temperature fluctuation threshold. The spatial segments that meet the joint standard conditions are selected to generate a set of continuous fluctuation segments;

[0101] Based on the synchronous state trajectory of the grid entities in the farmland area mapping results, it is necessary to determine the corresponding number and spatial positioning of each grid unit in the time series. Taking the 10-meter resolution grid in the area as the unit, the 7-day continuous state trajectory of each grid number is recorded to obtain soil moisture data and temperature data from time points T1 to T7. The humidity value is collected from the soil moisture sensor buried 20 cm deep in the surface layer, and the temperature data is collected from the meteorological monitoring unit deployed at the height of the crop canopy. The sampling interval is once a day. The humidity values ​​from T1 to T7 are set to 29%, 31%, 28%, 27%, 25%, 26%, and 30%, and the temperatures are 21.8℃, 22.1℃, 21.4℃, 20.9℃, 21.2℃, 21.7℃, and 22.3℃, respectively. The humidity change rate and temperature change rate are calculated for this type of data. The formula change rate = (next day - previous day) / number of days is used to obtain the difference between adjacent time points. The maximum difference amplitude within three consecutive days is calculated. For example, the humidity value from T3 to T5 drops from 28% to 25%, with an amplitude of 3% and a change rate of 1%. If the humidity fluctuation threshold is set to 0.8% and the temperature fluctuation threshold is 0.5°C, then the grid has exceeded the joint fluctuation judgment standard during the period T3-T5 and enters the screening range. All grid entities in the entire area are counted, and the change rate amplitude interval within the continuous time period is calculated respectively. The grid ID and corresponding time window of the segment where the fluctuation amplitude exceeds the joint standard are recorded. The spatial segment area is delineated and encoded according to its coordinates to generate a segment index list. Each segment is identified by (segment number-start and end date-spatial location) to establish a unified numbering structure. For example, segment P104-T3T5-X35Y72 represents the 104th fluctuation segment, with the time period from T3 to T5 and the spatial location at the X35Y72 grid. The set of spatial segments that meet the conditions is screened to obtain the continuous fluctuation segment set.

[0102] S312: Calling the spatial position pointed to by the continuous fluctuation segment set, performing image segmentation on the crop growth image, extracting the crop entity boundary in the image, and calculating the boundary morphological feature value and color distribution value, to determine whether there is a spatial offset phenomenon of crop structure change or color migration within the segment, and obtain the crop offset feature distribution value;

[0103] Call each fragment index in the continuous fluctuation fragment set, locate its corresponding spatial grid, and call the crop growth image sequence consistent with its time range, slice and align the images to ensure that the analysis area corresponds to the spatial fragment. Take fragment P104 as an example, its coordinates are located at X35Y72, and the corresponding time period is T3 to T5, that is, the crop image with this spatial position between May 12 and May 14, 2023 is called, and the image boundary extraction tool is used to calculate the edge contour of the crop entity. The green vegetation area is extracted by threshold segmentation method, and the contour tracking algorithm is used to record its boundary coordinates. The boundary coordinate difference between consecutive image frames is calculated. For example, the main stem boundaries in image 1 and image 2 are located at pixel coordinates (120, 150) and (125, 152) respectively. The difference is the Euclidean distance ((125-120)²+(152-150)²)≈5.38 pixels. This boundary offset is the basic indicator of morphological characteristics. At the same time, the mean change in the RGB value of the crop's main color is obtained. Image 1 is (45, 120, 30), and Image 2 is (50, 130, 35). The difference is +5 for the red channel, +10 for the green channel, and +5 for the blue channel. After transformation, the color change intensity is approximately 8.16 units. These two values ​​are combined with the corresponding weights of the crop type to construct a composite index. For example, if the weights are 0.6 and 0.4, respectively, the crop offset value is 0.6 × 5.38 + 0.4 × 8.16 ≈ 6.69 units. This process is repeated for each image segment to form a set of crop offset values ​​corresponding to the spatial segment. The combined change trend of crop outline morphology and color distribution within different segments is further analyzed to determine whether there is spatial offset phenomenon of structural change or color migration. Structural change is defined as a boundary displacement exceeding 5 pixels, and color migration is defined as an RGB change exceeding 10 units. Through this calculation process, a comprehensive offset value is obtained for each segment, and the crop offset characteristic distribution value is obtained.

[0104] S313: Using the crop offset characteristic distribution value, call the humidity change rate and temperature change rate corresponding to the continuous fluctuation segment set, perform difference calculation and combination judgment on the crop offset value and the environmental variable change rate, using the formula:

[0105] ;

[0106] Calculate the environmental response difference coefficient of the spatial segment, evaluate the correlation between the segment position and the environmental response difference coefficient, and obtain the environmental response sensitivity distribution map;

[0107] in, Representative The environmental response variance coefficient of the fragment, Representative The crop offset feature distribution value of the fragment, For the The humidity change rate of the segment, For the The temperature change rate of the segment, For the Clip crop image Continuity factor for segment boundaries, is the number of boundary segments in the image segment;

[0108] The crop offset feature distribution value is jointly analyzed with the humidity change rate and temperature change rate corresponding to the segment. In this process, a unified data structure needs to be constructed to accommodate the parameter values. Taking segment P104 as an example, its crop offset value is 6.69, the humidity change rate is 1.0% / day, and the temperature change rate is 0.7℃ / day. The image continuity factor comes from the boundary overlap ratio between image frames. Assuming that the boundary overlap ratios in three consecutive image frames are 85%, 90%, and 88%, the total factor is 2.63. Substituting each parameter into the numerical calculation:

[0109] Substituting the values ​​into:

[0110] ;

[0111] The environmental response difference coefficient of segment P104 is 9.11. Comparing this value with the set threshold of 4.0, it can be determined that it belongs to the response abnormal area. The threshold of 4.0 comes from the 95% confidence interval limit set in data statistics. Segments exceeding this value are considered environmentally sensitive areas. The response difference values ​​of the segments are spatially located, the coordinate positions are recorded, and the one-to-one mapping between the spatial position information and the response values ​​is evaluated, which is the environmental response sensitivity distribution map.

[0112] See also Figure 5 The specific steps for obtaining the results of agricultural ecological environment risk trend identification are as follows:

[0113] S411: Calling the block node number in the environmental response sensitivity distribution map, extracting the humidity change data and temperature change data corresponding to the block, comparing the humidity decrease rate and the temperature increase rate node by node, analyzing the synchronous change trend of the humidity change rate and the temperature change rate, and obtaining the synchronous change trend comparison value;

[0114] Call the block node number in the environmental response sensitivity distribution map, extract the number by identifying the geographic unit number represented by each node in the map, and retrieve the corresponding humidity and temperature sensor observation data. Set the node numbers 101, 102, 103, etc. to correspond to different observation points, and extract the time series data of humidity and temperature under each number. The humidity change rate can be obtained by the ratio of the humidity difference between adjacent moments to the time interval. The temperature change rate is calculated in the same way. If the time interval in node 101 is 2 hours and the humidity drops from 60% to 55%, the humidity change rate is:

[0115] ;

[0116] If the temperature rises from 22°C to 24°C, the rate of temperature change is:

[0117] ;

[0118] Repeat the above calculation process to obtain the humidity and temperature change rate sequence of each node, perform synchronization judgment on each node, mark the nodes where the humidity decreases and the temperature increases simultaneously, and if there are at least three consecutive nodes with the same synchronous change trend in a certain segment, it is regarded as a synchronous change segment. Count the node combinations that meet the synchronization conditions to form a humidity and temperature synchronization trend data structure for subsequent comparison, establish and identify trend change curves for all synchronous segments, and generate synchronous change trend comparison values.

[0119] S412: Based on the synchronous change trend control value, extract the continuous segments of humidity decrease rate and temperature increase rate, divide the fluctuation range by segment, and compare the segment differences with the set ecological stability range using the formula:

[0120] ;

[0121] Calculate the regional difference rate, classify and judge the regional difference rate and the interval stability setting boundary, record the consecutive numbered segments that exceed the stable interval, and obtain the trend deviation area index sequence;

[0122] in, represents the regional difference rate, Indicates the The humidity change rate of the node, Indicates the The temperature change rate of the node, Indicates the Humidity fluctuation value of the node, Indicates the The temperature fluctuation value of the node, Indicates the The humidity of the node is extremely poor, Indicates the The temperature difference at the node is extremely low. Indicates the number of nodes in the segment;

[0123] Parameter meaning and formula calculation derivation process:

[0124] Assume that the total number of monitoring points is five, namely , numbered respectively All parameters are derived from hourly meteorological monitoring records at the monitoring point for three consecutive days. Humidity and temperature are recorded at 10-minute intervals by a digital temperature and humidity acquisition terminal. The rate of change is calculated by taking data from adjacent time periods. The fluctuation value is measured by the standard deviation, and the range is calculated as the difference between the maximum and minimum values.

[0125] The first monitoring point: humidity change rate % / h, derived from the humidity in the monitoring record day 2 period from 64% to 61%, with a time interval of 2 hours, the rate of change is:

[0126] ;

[0127] The unit conversion is unified to 0.015, the percentage system corresponds to 0.15, and the normalized value is 0.14 temperature change rate ℃ / h, the collection time is the same 2-hour period, the temperature rises from 23.4℃ to 27.8℃, the rate of change is:

[0128] ;

[0129] The normalized value after unit conversion is 0.22 humidity fluctuation value , derived from the standard deviation of temperature fluctuation values ​​of 5 consecutive sets of humidity data , the same treatment method has extremely poor humidity , take the difference between the maximum value 68% and the minimum value 48% and normalize it to process the temperature extreme , take the difference between the maximum value of 33.1℃ and the minimum value of 25.6℃ and normalize it;

[0130] In this way, set the following parameters as follows:

[0131] The second monitoring point:

[0132] , , , , , ;

[0133] The third monitoring point:

[0134] , , , , , ;

[0135] The fourth monitoring point:

[0136] , , , , , ;

[0137] The fifth monitoring point:

[0138] , , , , , ;

[0139] Calculate the molecular part:

[0140] ;

[0141] ;

[0142] ;

[0143] ;

[0144] ;

[0145] Numerator total:

[0146] ;

[0147] Calculate the denominator:

[0148] ;

[0149] ;

[0150] Calculate the denominator:

[0151] ;

[0152] Substitute into the formula for calculation:

[0153] ;

[0154] The results show that the regional difference rate between humidity and temperature changes in the current continuous segment is 0.00864, ​​which is less than the upper limit of the ecological stability boundary set value of 0.0100. Therefore, it can be determined that there is no trend deviation fluctuation in this segment during the current observation period and it is not included in the trend deviation regional index sequence. This value reflects the intensity measurement of the deviation of humidity and temperature changes in this segment and serves as an important numerical criterion for determining whether to include it in deviation identification.

[0155] S413: Based on the trend shift region index sequence, track the level number of the indicated region, detect the fluctuation trend distribution characteristics of humidity and temperature in the region, compare the fluctuation direction and intensity, and obtain the agricultural ecological environment risk trend identification result;

[0156] Identify and track the regional hierarchical numbers of all nodes marked as trend shifts, map the numbers to the hierarchical structure in the spatial region, set nodes 103 and 104 to belong to region A2, and nodes 107 and 108 to belong to region B1, collect the humidity change rate and temperature change rate of the nodes in the region group by group, establish a climate change dataset within the regional hierarchy, and perform directional consistency judgment on each set of data. The humidity shows a downward trend and the temperature shows an upward trend, and the change amplitude must exceed the average change rate within the regional segment. The judgment basis can set the humidity change critical value to 1.5% / h and the temperature rise critical value to 0.8℃ / h. If the proportion of nodes in the region greater than this threshold exceeds 60%, the region is determined to be a trend shift area, and the nodes that meet the conditions in the region are marked as shift trend points. The spatial aggregation of all trend shift points is counted. The aggregation degree can be defined as the average distance between adjacent shift points is less than 100 meters, or a combination of more than three consecutive shift points is formed in space, which will be marked as a risk area. The shift point combination and the area where it is located are recorded uniformly to generate the agricultural ecological environment risk trend identification results.

[0157] See also Figure 6 , the specific steps for obtaining the agricultural ecological risk warning interface layer are as follows:

[0158] S511: Based on the results of agricultural ecological environmental risk trend identification, the 3D environmental model in the digital twin interface is called, and the time series data of wind speed, rainfall, and temperature are loaded into the interface. The soil moisture value, soil salinity value, and soil organic matter concentration layers are superimposed to obtain the fused coordinate distribution value of the 3D layer.

[0159] It is necessary to extract the key parameters in the risk identification layer, including the location of spatial coordinate points, the time series arrangement structure and the corresponding risk level information. The extraction process is carried out based on the preset zoning coding rules, and the attributes of each risk point are identified by combining the grid data of soil moisture, salinity, and organic matter concentration in each area. A certain area is set to be identified as a mild risk between 9:00 and 21:00 on May 5, 2024. The corresponding coordinate points, time numbers and risk tags will be synchronously extracted and organized into a unified data set. The three-dimensional environmental model built in the digital twin platform is called, and a layer mapping table is established in the model coordinate system, corresponding to each coordinate point in the above data set and the three-dimensional model space. The mapping relationship between the points is established, and the coordinate correction and structure access operations are completed. On this basis, the meteorological data layer is introduced, and the temperature, wind speed and rainfall values ​​in the area are loaded hourly. These are mapped to the corresponding location points in chronological order to form a meteorological layer structure with spatial and temporal attributes. The soil state layer is superimposed on the model interface, and the soil moisture value, salinity value and organic matter content value are extracted. The soil information and meteorological information of each location point are associated and combined one-to-one, and marked as different time series attribute combinations under the same spatial node. The fused data set uniformly mapped in the 3D interface is obtained. This data set serves as the basis for subsequent layer comparison, boundary judgment and visual highlighting to obtain the fused coordinate distribution value of the 3D layer.

[0160] S512: Based on the fused coordinate distribution values ​​of the 3D layer, the difference between the meteorological layer parameter and the soil state layer parameter of the marked area in the continuous frames of the real time and the previous period is obtained, and the block boundary change trend is determined according to the time axis of the difference sequence. The continuous frame blocks with outward expansion boundaries are marked as expansion areas, and the continuous boundary expansion recognition result is obtained;

[0161] Perform a difference comparison operation on the layer contents in any two consecutive time periods. The comparison focuses on the changes in soil and meteorological conditions at each spatial location at different times. Set a numerical comparison of the rainfall value, temperature and wind speed at a certain location between 9:00 and 10:00 on May 5. At the same time, judge the numerical changes of its soil moisture, salinity and organic matter concentration in the corresponding period. For numerical changes, centralized comparison can be performed by setting a static time window, classifying the change values ​​by region and screening whether they exceed the preset change threshold. Set a threshold when more than 70% of the nodes in a certain grid area are in the weather. If the numerical changes in the image or soil exceed the standard change range, the area is marked as a potential boundary change block, and then such blocks are numbered and tracked. The status of the numbered blocks in multiple time periods is continuously tracked. If the expansion conditions are met in three consecutive time periods, it is considered that the boundary has actually expanded. The boundary coordinate set is extracted and stored in the spatial dimension, and the adjacent areas that meet the conditions are merged to form a new round of boundary identification range. On this basis, the boundary change judgment structure of the current stage is constructed, and the parameters such as the area, directionality and duration of the boundary expansion are further counted to obtain the continuous boundary expansion identification result.

[0162] S513: Based on the continuous boundary expansion recognition results, the spatial coordinate set of the changed area is extracted in the digital twin interface, and the coordinates are compared within the visible range. The width and color parameters of the highlighted block boundary line are set, and the areas that meet the expansion judgment conditions are highlighted in the interface. The layer name and display order are set to generate the agricultural ecological risk warning interface layer;

[0163] In the digital twin rendering interface, a spatial coordinate set is extracted and mapped. This set comes from all regional nodes marked as boundary expansion in the previous stage. Their spatial locations are uniformly called and combined into a boundary segment set through the provided 3D coordinate interface. Rendering parameters are set for the boundary segments, including line color, width, layer transparency, and other attributes. The expansion area border is set to red, 2 pixels wide, and 0.7 transparency. A layer name such as "Risk-Border-Layer" is bound to it and the display priority parameter is set to 1, giving it priority during interface loading. A layer legend and control buttons are added to the layer control panel on the right, allowing users to turn the layer on and off. Based on this, the current interface's visible window range is extracted, and only the area boundaries within the view range are rendered to improve layer loading efficiency. The number of each expansion area is displayed on the boundary line in the interface, along with real-time expansion status time information, such as "Area A expanded at 12:00 on May 5th", to help users identify the timing and location of the change. The boundary line information is bound to the current 3D scene layer and embedded in the environmental model structure to generate the agricultural ecological risk warning interface layer.

[0164] The agricultural ecological environment monitoring system based on digital twin is used to implement the above-mentioned agricultural ecological environment monitoring method based on digital twin. The system includes:

[0165] The index recognition module acquires continuous data from soil moisture monitors, temperature sensors, and meteorological data collection devices deployed in the farmland area, establishes corresponding index relationships according to sensor numbers, filters data groups with overlapping time intervals, and generates synchronized scene images of the ecological environment.

[0166] The virtual mapping module uses the calibrated spatial grid coordinates in the synchronized scene image of the ecological environment to call the regional virtual grid point numbers in the digital twin environment, compare the humidity, temperature and meteorological parameter values ​​between the grid points, and update the environmental performance of the corresponding nodes in the 3D model to obtain the farmland area mapping results.

[0167] The change recognition module calls the humidity change sequence and temperature change sequence of the virtual nodes in the farmland area mapping results, extracts the continuously changing spatial segments based on the change values ​​between adjacent time points, and generates an environmental response sensitivity distribution map;

[0168] The trend analysis module extracts the humidity and temperature change trajectories of the nodes according to the spatial node numbers identified by the environmental response sensitivity distribution map, identifies the change curves of the two change trajectories in the nodes, judges the direction and magnitude of the change trend, and generates the agricultural ecological environment risk trend identification results;

[0169] The risk warning module calls the spatial level number recorded in the agricultural ecological environment risk trend identification results, matches the farmland area boundary nodes in the three-dimensional model, superimposes the meteorological parameter layer and soil moisture layer under the area, judges the boundary changes of the area corresponding to the warning number, and obtains the agricultural ecological risk warning range layer.

[0170] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. The agricultural ecological environment monitoring method based on digital twins is characterized by: The following steps are involved: S1: Acquire continuous data from soil moisture monitors, temperature sensors, and meteorological data collection devices deployed in the farmland area, establish corresponding mapping relationships through sensor number identification, arrange the collected information in time sequence, and generate a synchronized scene image of the ecological environment; S2: Based on the spatial grid coordinates corresponding to the humidity, temperature, and meteorological conditions recorded in the synchronized scene image of the ecological environment, the regional virtual grid built into the digital twin is called, the real-scene sensor synchronization information is connected, and the regional three-dimensional environmental model representation is updated according to the real-time environmental information to generate the farmland area mapping result; S3: Based on the synchronous state trajectory of the grid entities in the farmland area mapping result, by comparing the change gradients of humidity and temperature between different time points, screening spatial segments with continuous fluctuation amplitudes, analyzing the crop growth images within the spatial segments, and generating an environmental response sensitivity distribution map; S4: calling the block node number in the environmental response sensitivity distribution map, tracking the synchronous evolution process of the humidity drop section and the temperature rise node, comparing the change amplitude curve with the set ecological stability range for segment differences, and generating the agricultural ecological environment risk trend identification result; S5: Based on the agricultural ecological environmental risk trend identification results, the three-dimensional environmental model in the digital twin interface is called, the meteorological condition animation layer and the soil status are superimposed, and it is determined whether the boundary change range of the warning block is continuously expanding. The highlighted area and the visible range of the interface are marked to obtain the agricultural ecological risk warning interface layer; The agricultural ecological risk warning interface layer includes a dynamic boundary layer, a highlighted visible area, and a risk level overlay map; The steps for obtaining the agricultural ecological risk warning interface layer are as follows: S511: Based on the agricultural ecological environment risk trend identification results, the three-dimensional environmental model in the digital twin interface is called, and the time series data of wind speed, rainfall, and temperature are loaded into the interface. The soil moisture value, soil salinity value, and soil organic matter concentration layers are superimposed to obtain the three-dimensional layer fusion coordinate distribution value; S512: Based on the fused coordinate distribution values ​​of the three-dimensional layer, the difference between the meteorological layer parameter and the soil state layer parameter of the marked area in the continuous frames of the real time and the previous time period is obtained, and the block boundary change trend is judged according to the time axis of the difference sequence. The continuous frame blocks with outward expansion boundaries are marked as expansion areas, and the continuous boundary expansion recognition result is obtained; S513: Based on the continuous boundary expansion identification results, the spatial coordinate set of the changed area is extracted in the digital twin interface, and the coordinates are compared within the visible range. The width and color parameters of the highlighted block boundary line are set, and the areas that meet the expansion judgment conditions are highlighted in the interface. The layer name and display order are set to generate the agricultural ecological risk warning interface layer.

2. The agricultural ecological environment monitoring method based on digital twin according to claim 1 is characterized in that: The ecological environment synchronized scene image includes spatial feature grids, time-series alignment segments, and multi-source data fusion areas. The farmland area mapping results include grid number status sets, real-time environmental parameter groups, and three-dimensional structure representation sets. The environmental response sensitive distribution map includes parameter change highlighting areas, crop status response areas, and environmental factor interference zones. The agricultural ecological environment risk trend identification results include offset area numbers, trend curve changes, and ecological indicator change segments.

3. The agricultural ecological environment monitoring method based on digital twin according to claim 1 is characterized in that: The steps for acquiring the ecological environment synchronization scene image are specifically as follows: S111: Acquire continuous data from soil moisture monitors, temperature sensors, and meteorological data collection devices deployed in the farmland area, add number identifiers, evaluate the mapping relationship between device numbers and data records, sort the data records by timestamp field, and generate a sensor time series number dataset; S112: calling the data records corresponding to the device numbers in the sensor time series number data set, screening the soil moisture, temperature, and meteorological data entries with overlapping timestamps, performing a pairing operation on the temporally overlapping data according to the spatial coordinates of the data records, and combining the data records based on the spatial coordinate matching results as a judgment condition to generate a list of overlapping spatiotemporal joint values; S113: According to the spatial coordinate field and time field in the overlapping spatiotemporal joint value list, the coordinates are used as image pixel numbers, the time field is mapped to the image frame sequence number, and each type of value is mapped to a color channel according to the pixel ratio to generate an ecological environment synchronization scene image.

4. The agricultural ecological environment monitoring method based on digital twins according to claim 3 is characterized in that: The steps for obtaining the farmland area mapping result are specifically as follows: S211: Using the humidity, temperature, and meteorological condition data in the synchronized scene image of the ecological environment, calling the regional virtual grid coordinates built into the digital twin, evaluating the mapping relationship between the grid coordinates and the sensor data, and generating a gridded environmental parameter set; S212: Based on the gridded environmental parameter set, soil hydraulic conductivity, crop coefficient, and evapotranspiration are extracted from the multi-region parameter mapping information using the formula: ； Calculate the environmental volatility, compare the environmental volatility with the material response threshold of the regional 3D model, and generate the regional dynamic response scalar; in, is the environmental volatility, represents the temperature measurement value, represents the grid weight factor, represents the measured value of soil hydraulic conductivity, represents the measured mean evapotranspiration, Represents the quantitative value of meteorological conditions, represents the crop growth stage coefficient, Represents the humidity percentage value; S213: Calling the regional dynamic response scalar, updating the texture resolution of the three-dimensional environment model according to the light intensity and wind speed in the real-time environmental information, and generating a farmland area mapping result.

5. The agricultural ecological environment monitoring method based on digital twin according to claim 4 is characterized in that: The steps for obtaining the environmental response sensitivity distribution map are specifically as follows: S311: Based on the synchronization state trajectory of the grid entity in the farmland area mapping result, the humidity and temperature records of the grid unit in the time dimension are called, the fluctuation degree of the humidity change rate and the temperature change rate in the continuous time period are compared, and whether the continuous fluctuation amplitude exceeds the joint standard of the humidity fluctuation threshold and the temperature fluctuation threshold is determined, and the spatial segments that meet the joint standard conditions are selected to generate a continuous fluctuation segment set; S312: Calling the spatial position pointed to by the continuous fluctuation segment set, performing image segmentation on the crop growth image, extracting the crop entity boundary in the image, and calculating the boundary morphological feature value and color distribution value, determining whether there is a spatial offset phenomenon of crop structure change or color migration within the segment, and obtaining the crop offset feature distribution value; S313: Using the crop offset characteristic distribution value, calling the humidity change rate and temperature change rate corresponding to the continuous fluctuation segment set, performing difference calculation and combined judgment on the crop offset value and the environmental variable change rate, using the formula: ； Calculate the environmental response difference coefficient of the spatial segment, evaluate the correlation between the segment position and the environmental response difference coefficient, and obtain the environmental response sensitivity distribution map; in, Representative The environmental response variance coefficient of the fragment, Representative The crop offset feature distribution value of the fragment, For the The humidity change rate of the segment, For the The temperature change rate of the segment, For the Clip crop image Continuity factor for segment boundaries, is the number of boundary segments in the segment image.

6. The agricultural ecological environment monitoring method based on digital twin according to claim 5 is characterized in that: The specific steps for obtaining the agricultural ecological environment risk trend identification results are as follows: S411: Calling the block node number in the environmental response sensitivity distribution map, extracting the humidity change data and temperature change data corresponding to the block, comparing the humidity decrease rate and the temperature increase rate node by node, analyzing the synchronous change trend of the humidity change rate and the temperature change rate, and obtaining a synchronous change trend comparison value; S412: Based on the synchronous change trend control value, extract the continuous segments of the humidity drop rate and the temperature rise rate, divide the fluctuation range into sections according to the sections, and compare the section differences with the set ecological stability range using the formula: ； Calculate the regional difference rate, classify and judge the regional difference rate and the interval stability setting boundary, record the consecutive numbered segments that exceed the stable interval, and obtain the trend deviation area index sequence; in, represents the regional difference rate, Indicates the The humidity change rate of the node, Indicates the The temperature change rate of the node, Indicates the Humidity fluctuation value of the node, Indicates the The temperature fluctuation value of the node, Indicates the The humidity of the node is extremely poor, Indicates the The temperature difference at the node is extremely large. Indicates the number of nodes in the segment; S413: Based on the trend offset region index sequence, track the level number of the indicated region, detect the fluctuation trend distribution characteristics of humidity and temperature in the region, compare the fluctuation direction and intensity, and obtain the agricultural ecological environment risk trend identification result.

7. The agricultural ecological environment monitoring system based on digital twins is characterized by: The system is used to implement the agricultural ecological environment monitoring method based on digital twins according to any one of claims 1 to 6, and the system includes: The index recognition module acquires continuous data from soil moisture monitors, temperature sensors, and meteorological data collection devices deployed in the farmland area, establishes corresponding index relationships according to sensor numbers, filters data groups with overlapping time intervals, and generates synchronized scene images of the ecological environment. The virtual mapping module calls the regional virtual grid point numbers in the digital twin environment based on the calibrated spatial grid coordinates in the synchronized scene image of the ecological environment, compares the humidity values, temperature values ​​and meteorological parameter values ​​between the grid points, updates the environmental performance of the corresponding nodes in the three-dimensional model, and obtains the farmland area mapping result; The change recognition module calls the humidity change sequence and temperature change sequence of the virtual nodes in the farmland area mapping result, extracts the continuously changing spatial segments according to the change values ​​between adjacent time points, and generates an environmental response sensitivity distribution map; The trend analysis module extracts the humidity change trajectory and temperature change trajectory of the node according to the spatial node number identified by the environmental response sensitivity distribution map, identifies the change curves of the two change trajectories in the node, judges the direction and magnitude of the change trend, and generates an agricultural ecological environment risk trend identification result; The risk warning module calls the spatial level number recorded in the agricultural ecological environment risk trend identification result, matches the farmland area boundary node in the three-dimensional model, superimposes the meteorological parameter layer and soil moisture layer under the area, judges the boundary changes of the area corresponding to the warning number, and obtains the agricultural ecological risk warning range layer.

Images