Farmland heavy metal pollution monitoring system

Abstract

The invention discloses a farmland heavy metal pollution monitoring system, and relates to the field of heavy metal detection, and the system comprises a creation module which is used for uploading the structure information of an area where a farmland is located, and creating a farmland twinborn model based on the structure information of the area where the farmland is located; the logic module is used for generating heavy metal sensor deployment logic according to the farmland twinborn model; according to the method, the farmland twinborn model is constructed and calibrated according to the farmland and the structure information of the area, the high matching degree of the model and the actual farmland is guaranteed, an accurate foundation is laid for monitoring, the areas are equally divided according to the topographic features of the twinborn model, the heavy metal sensors are differentially deployed, the lower the position is, the higher the density is, and comprehensive and targeted monitoring of each area of the farmland is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of heavy metal detection, in particular to a farmland heavy metal pollution monitoring system. BACKGROUND

[0002] Farmland heavy metal pollution is an environmental problem caused by the invasion of heavy metals such as Hg, Cd, and Pb into farmland, which is mainly caused by industrial emissions, misuse of fertilizers and pesticides, and sewage irrigation. It can destroy soil structure, reduce crop yield and quality, and accumulate through the food chain, endangering human health.

[0003] The invention patent application with publication number CN118521897B discloses a farmland soil heavy metal pollution monitoring and early warning method, which includes: collecting farmland soil surface images, soil spectrum data and ground sensor data; preprocessing the farmland soil surface images, soil spectrum data and ground sensor data; inputting the preprocessed data into the recognition network, and training the recognition network; wherein, before training, the recognition network is compressed, including the following steps: step one: sorting the weight values of the recognition network, and setting a weight threshold; step two: deleting neurons below the weight threshold to obtain a pruned recognition network; step three: fine-tuning the pruned recognition network: freezing the routing layer and the up-sampling layer, setting the learning rate to 0.0001, and training the fine-tuned recognition network; inputting the collected farmland soil surface images, soil spectrum data and ground sensor data into the trained recognition network to identify whether there is heavy metal pollution area in the farmland soil. The application aims to solve the problems of "the existing technology lacks efficient monitoring means for farmland soil heavy metal pollution, cannot accurately identify and timely warn potential pollution areas, cannot realize comprehensive coverage monitoring of farmland soil, the spatio-temporal resolution of monitoring data is not high, and cannot meet the real-time monitoring requirements of traditional soil sampling analysis method, which has high time cost and complex operation, and cannot adapt to the monitoring needs of large-scale land".

[0004] However, at present, due to the influence of human construction and natural factors on farmland environment, the heavy metal content in farmland exceeds the standard, which affects the fertility of farmland and seriously affects the growth of crops.

[0005] Therefore, we propose a farmland heavy metal pollution monitoring system. SUMMARY

[0006] In view of the above shortcomings of the prior art, the present application provides a farmland heavy metal pollution monitoring system, which can effectively solve the problems of the prior art.

[0007] To achieve the above purpose, the present application realizes the following technical solutions: This invention discloses a farmland heavy metal pollution monitoring system, comprising: The system comprises the following modules: a creation module for uploading structural information of the farmland area and creating a farmland twin model based on this information; a logic module for generating deployment logic for heavy metal sensors based on the farmland twin model; a control and storage module for controlling the heavy metal sensors to operate according to a preset cycle, receiving and storing operational sensing data from each heavy metal sensor during operation; an evaluation module for acquiring the sensing data stored in the control and storage module and evaluating farmland heavy metal pollution trend factors based on this data; and an early warning module for continuously acquiring farmland heavy metal pollution trend factors from the evaluation module, determining whether farmland faces heavy metal pollution risk based on these factors, and triggering an early warning based on the determination result.

[0008] Furthermore, the structural information of the farmland area uploaded during the creation module's operation includes: farmland edge coordinates and farmland edge coordinates centered on the farmland area. The farmland edge coordinates correspond one-to-one with each other, and the distances between each corresponding farmland edge coordinate and the farmland edge coordinate are equal. Moreover, the straight-line distance between the corresponding farmland edge coordinates and the farmland edge coordinates is greater than the maximum radius of the farmland and less than the maximum diameter of the farmland. The farmland's edge coordinates and the farmland's edge coordinates are both three. The three farmland edge coordinates are connected to form a closed finite surface, denoted as A. The three farmland edge coordinates are connected to form a closed finite surface, denoted as B. Based on A, B is calibrated to complete the creation of the farmland twin model. Among them, from the top-down view, the edge contour of B is within the range of A, and the three uploaded farmland edge coordinates follow the principle that the difference between the area of ​​B determined based on the three farmland edge coordinates and the actual farmland area is the smallest.

[0009] Furthermore, the calibration operation based on A for B is as follows: Identify the angle between A and B. If the angle between A and B is less than a preset value, B will be directly used as the farmland twin model. When the angle between A and B is greater than a preset threshold, B is calibrated based on A, and the calibrated B is used as the farmland twin model: ; In the formula: The calibration angle for B; Let A be the angle between A and B; For calibration coefficients; The preset value used for comparing the angle between A and B is user-defined on the system side, and it follows the principle that the higher the accuracy requirement for monitoring heavy metal pollution in farmland, the smaller the preset value, and vice versa. After calculating the calibration angle of B based on the above formula, the direction in which B and A form an angle is used as the flipping direction, and the intersection line of the angle between B and A is used as the flipping axis. Perform a flip operation to complete the calibration.

[0010] Furthermore, the calibration coefficient The value is determined based on the degree of difference between the farmland's location and the soil properties of the farmland. The center point of B is selected. Connecting the corresponding edge coordinates of B and A and extending to the center point yields three line segments. These three line segments are used to divide the farmland area into regions corresponding to A and B, with the farmland edge as the boundary. These regions are denoted as a1, a2, a3 and b1, b2, b3, respectively. Soil composition sensors are uniformly deployed within a1, a2, a3 and b1, b2, b3 to detect the soil composition. The calibration coefficients are then calculated. ; In the formula: 3 represents the number of control groups, including a1 and b1, a2 and b2, and a3 and b3; The total number of soil component types detected in the i-th control group; These represent the proportion of the nth component in the i-th region of the farmland and the proportion of the nth component in the i-th region of the farmland, respectively. The allocation weights for the nth component; in, The number of soil component types corresponding to the region with the most soil component types in the i-th control group is selected. When the value is zero, the value of the fraction inside the parentheses is taken as zero. , To preset the minimum value, All are positive numbers, and their values ​​satisfy... Furthermore, the smaller the maximum difference in content between the nth component and the components sensed by each soil component sensor, the better. The larger the value, the lower the value. The smaller the value, All values ​​are the average content of the nth component sensed by each soil component sensor, with a simultaneously set constraint interval of [0.75, 1.25]. The values ​​are calculated based on the above formula. If it is within the constrained range, then upload based on the above formula. Conversely, take the one closer to... The boundary value of the constraint interval, the soil composition sensor detects that the soil composition does not contain heavy metal components.

[0011] Furthermore, the logic module is internally equipped with a retrieval unit, which is used to access the creation module and retrieve the farmland twin model from the creation module. The farmland twin model retrieved by the retrieval unit is a calibrated farmland twin model. The heavy metal sensor deployment logic generated in the logic module is as follows: Using A as the plane reference plane, the farmland twin model is divided into three local models from the highest position to the lowest position. Heavy metal sensors are uniformly deployed in a matrix in the three local models. The goal is to achieve the sparsest density of heavy metal sensors in the local model with the highest position, the densest density in the local model with the lowest position, and a density between sparse and dense in the local model with the middle position. The number of heavy metal sensors deployed in the local model with the highest position is a preset number x, which is defined by the system user and is a non-zero even number. The number of heavy metal sensors deployed in the local model with the lowest position is 2x; The number of heavy metal sensors deployed in a moderately located local model is .

[0012] Furthermore, the preset cycle for the control and storage module to control the operation of the heavy metal sensor is customized by the system user and is adapted to the insertion of events such as the sewage discharge cycle of factories around the farmland, rainfall, and fertilizer and pesticide irrigation. That is, it runs once after each factory sewage discharge, rainfall, and fertilizer and pesticide irrigation event occurs, and factories within the area where the farmland is located are identified as surrounding factories. When storing the sensing data from the heavy metal sensors, the data is stored separately based on the heavy metal sensors from which the data originates. After storage separately, the data is stored separately again based on the sensing cycle from which the data originates. When storing separately based on the heavy metal sensors from which the data originates, the deployment location of each heavy metal sensor is used as the name of the distinguishing interval. When storing separately based on the sensing cycle from which the data originates, the sequence number of each sensing cycle is used as the name of the distinguishing interval.

[0013] Furthermore, the heavy metal sensor data includes several sets of heavy metal content values, and when the assessment module evaluates the trend factor of heavy metal pollution in farmland, it follows the following: Retrieve the sensing data from the four most recent heavy metal sensor runs in the evaluation module, denoted as X1, X2, X3, and X4. Calculate the heavy metal pollution trend factor in farmland using the sensing data from each of the two consecutive heavy metal sensor runs, based on the following formula: ; In the formula: As a trend factor for heavy metal pollution in farmland; To perceive the total number of data types; , The content of the m-th heavy metal in the sensing data from two consecutive heavy metal sensor operations; To configure weights; The result of calculating X1 and X2 is denoted as The calculation results of X2 and X3 are denoted as The calculation results of X3 and X4 are denoted as ; Once the location for the heavy metal sensors is determined, the burial depth is set to half the typical length of the rootstock of the plants grown in the farmland. All are positive numbers and and obey The greater the harm of a particular type of heavy metal to farmland, the higher its value.

[0014] Furthermore, the aforementioned ; In the formula: This represents the total number of heavy metal sensors deployed for each of the three local models. Let m be the content of the m-th heavy metal sensed by the p-th heavy metal sensor in this cycle; This is the proportionality coefficient. , This indicates the number of heavy metal sensors deployed in the local model with the highest position. The , The calculation logic is the same; The , , After obtaining, based on , , The change rates of two trend factors of heavy metal pollution in farmland were obtained. , recorded as .

[0015] Furthermore, during the operation phase of the early warning module, the latest three outputs of the farmland heavy metal pollution trend factor from the evaluation module are obtained. When the farmland heavy metal pollution trend factor continues to rise or the rate of change of any farmland heavy metal pollution trend factor is greater than a preset threshold, it is determined that there is a risk of heavy metal pollution in the farmland. The early warning module integrates an audible and visual alarm device. When the determination result indicates that there is a risk of heavy metal pollution in the farmland, the audible and visual alarm device will issue a preset audible and visual alarm.

[0016] Furthermore, the creation module is interconnected with a logic module and a retrieval unit via a wireless network; the logic module is interconnected with a control and storage module via a wireless network; the control and storage module is interconnected with an evaluation module via a wireless network; and the evaluation module is interconnected with an early warning module via a wireless network.

[0017] Compared with the known prior art, the technical solution provided by this invention has the following beneficial effects: This invention provides a farmland heavy metal pollution monitoring system. During operation, the system constructs and calibrates a farmland twin model using structural information of the farmland and its surrounding area, ensuring a high degree of matching between the model and the actual farmland, thus laying a precise foundation for monitoring. Heavy metal sensors are deployed in differentiated areas based on the terrain features of the twin model, with higher density at lower elevations, enabling comprehensive and targeted monitoring of all farmland areas. The sensor operating cycle can be adapted to events such as factory discharges and rainfall, and a preset cycle ensures no data is missed at key nodes. Sensing data is categorized and stored according to sensor location and acquisition cycle for easy retrieval and analysis later. Based on multiple sets of historical data and combined with heavy metal hazard weights, a pollution trend factor is calculated, simultaneously monitoring the trend change rate to accurately assess the pollution trend. When the trend continues to rise or the change rate exceeds a threshold, an audible and visual warning is triggered, providing timely reminders for prevention and control. This effectively improves the accuracy, comprehensiveness, and timeliness of farmland heavy metal pollution monitoring, providing support for farmland pollution prevention and control. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0019] Figure 1 This is a schematic diagram of a farmland heavy metal pollution monitoring system. Figure 2 This is a schematic diagram illustrating the logic of dividing farmland into different areas and farmland in this invention. Figure 3 This is a schematic diagram demonstrating the local model obtained by segmenting the farmland twin model in this invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0021] The present invention will be further described below with reference to embodiments. Example

[0022] This embodiment provides a farmland heavy metal pollution monitoring system, such as... Figure 1 As shown, it includes: Create a module to upload the structural information of the farmland area and create a farmland twin model based on the structural information of the farmland area; The structural information of the farmland area uploaded during the creation module operation phase includes: farmland edge coordinates, farmland edge coordinates centered on the farmland area, farmland edge coordinates correspond one-to-one with each other, and the distance between each corresponding farmland edge coordinate is equal, and the straight-line distance between the corresponding farmland edge coordinates is greater than the maximum radius of the farmland and less than the maximum diameter of the farmland. There are three edge coordinates for the farmland area and three edge coordinates for the farmland. The three edge coordinates of the farmland area are connected to form a closed finite surface, denoted as A. The three edge coordinates of the farmland are connected to form a closed finite surface, denoted as B. Based on A, B is calibrated to complete the creation of the farmland twin model. Among them, from the top-down view, the edge contour of B is within the range of A, and the three uploaded farmland edge coordinates follow the principle that the difference between the area of ​​B determined based on the three farmland edge coordinates and the actual farmland area is the smallest. The calibration operation for B based on A is as follows: Identify the angle between A and B. If the angle between A and B is less than a preset value, B will be directly used as the farmland twin model. When the angle between A and B is greater than a preset threshold, B is calibrated based on A, and the calibrated B is used as the farmland twin model: ; In the formula: The calibration angle for B; Let A be the angle between A and B; For calibration coefficients; The preset value used for comparing the angle between A and B is user-defined on the system side, and it follows the principle that the higher the accuracy requirement for monitoring heavy metal pollution in farmland, the smaller the preset value, and vice versa. After calculating the calibration angle of B based on the above formula, the direction in which B and A form an angle is used as the flipping direction, and the intersection line of the angle between B and A is used as the flipping axis. Perform a flip operation to complete the calibration; Note: When B and A form an angle, if there is no intersecting line, then expand the angle between B and A until an intersecting line appears and is then applied. The above formula is based on half of By controlling the degree of calibration of B, B can be adjusted to align with the direction of A during the calibration process, taking into account the authenticity of its own tilt state. This makes the calibrated B and A more consistent in their tilt directions, enabling B to reflect the flow state of water in the farmland soil after receiving water. This serves as a preliminary basis for the deployment of heavy metal sensors, effectively supporting the rational deployment of heavy metal sensors and ensuring the authenticity of the data sensed by each subsequent heavy metal sensor and its compatibility with the farmland. Calibration coefficient The value is determined based on the degree of difference between the farmland's location and the soil properties of the farmland. The center point of B is selected. Connecting the corresponding edge coordinates of B and A and extending to the center point yields three line segments. These three line segments are used to divide the farmland area into regions corresponding to A and B, with the farmland edge as the boundary. These regions are denoted as a1, a2, a3 and b1, b2, b3, respectively. Soil composition sensors are uniformly deployed within a1, a2, a3 and b1, b2, b3 to detect the soil composition. The calibration coefficients are then calculated. ; In the formula: 3 represents the number of control groups, including a1 and b1, a2 and b2, and a3 and b3; The total number of soil component types detected in the i-th control group; These represent the proportion of the nth component in the i-th region of the farmland and the proportion of the nth component in the i-th region of the farmland, respectively. The allocation weights for the nth component; To ensure that the calibration coefficient accurately reflects the difference in soil properties between the farmland and the surrounding area, the two areas were first divided into three control areas. The weighted average of the component differences in each area was obtained by a soil composition sensor to calculate the contribution of each group to the difference. The average of the three groups was then used as the calibration coefficient. At the same time, a constraint interval was set to avoid anomalies, so that the B-side calibration could be based on the actual differences in soil properties. This ensured that the tilt state of the calibrated farmland twin model was more consistent with the tilt state of the farmland area, thereby ensuring that the farmland twin model conformed to the actual flow pattern of water after the farmland received water. in, The number of soil component types corresponding to the region with the most soil component types in the i-th control group is selected. When the value is zero, the value of the fraction inside the parentheses is taken as zero. , To preset the minimum value, All are positive numbers, and their values ​​satisfy... Furthermore, the smaller the maximum difference in content between the nth component and the components sensed by each soil component sensor, the better. The larger the value, the lower the value. The smaller the value, All values ​​are the average content of the nth component sensed by each soil component sensor, with a simultaneously set constraint interval of [0.75, 1.25]. The values ​​are calculated based on the above formula. If it is within the constrained range, then upload based on the above formula. Conversely, take the one closer to... The boundary value of the constraint interval, the soil composition sensor detects that the soil composition does not contain heavy metal components; The logic module is used to generate heavy metal sensor deployment logic based on the farmland twin model; The logic module has a retrieval unit inside, which is used to access the creation module and retrieve the farmland twin model from the creation module. The farmland twin model retrieved by the retrieval unit is the calibrated farmland twin model. The deployment logic for the heavy metal sensor generated in the logic module is as follows: Using A as the plane reference plane, the farmland twin model is divided into three local models from the highest position to the lowest position. Heavy metal sensors are uniformly deployed in a matrix in the three local models. The goal is to achieve the sparsest density of heavy metal sensors in the local model with the highest position, the densest density in the local model with the lowest position, and a density between sparse and dense in the local model with the middle position. The number of heavy metal sensors deployed in the local model with the highest position is a preset number x, which is defined by the system user and is a non-zero even number. The number of heavy metal sensors deployed in the local model with the lowest position is 2x; The number of heavy metal sensors deployed in a moderately located local model is ; The control and storage module is used to control the heavy metal sensors to operate according to a preset cycle, and to receive the operating sensing data of each heavy metal sensor during operation and store the operating sensing data of the heavy metal sensors. The preset cycle for the control and storage module to control the operation of the heavy metal sensor is defined by the system user and is adapted to the insertion of events such as sewage discharge cycle of factories around farmland, rainfall, and fertilizer and pesticide irrigation. That is, it runs once after each sewage discharge, rainfall, or fertilizer and pesticide irrigation event occurs. Factories within the farmland area are identified as surrounding factories. When storing the sensing data of heavy metal sensors, the data is stored separately based on the heavy metal sensors from which the sensing data originates. After the data is stored separately, it is stored separately again based on the sensing cycle from which the sensing data originates. When storing separately based on the heavy metal sensors from which the sensing data originates, the deployment location of each heavy metal sensor is used as the name of the distinguishing interval. When storing separately based on the sensing cycle from which the sensing data originates, the sequence number of each sensing cycle is used as the name of the distinguishing interval. The assessment module is used to acquire the sensing data stored in the control and storage module, and to assess the trend factors of heavy metal pollution in farmland based on the sensing data. The heavy metal sensor data includes several sets of heavy metal content values. When the assessment module evaluates the trend factors of heavy metal pollution in farmland, it follows the following rules: Retrieve the sensing data from the four most recent heavy metal sensor runs in the evaluation module, denoted as X1, X2, X3, and X4. Calculate the heavy metal pollution trend factor in farmland using the sensing data from each of the two consecutive heavy metal sensor runs, based on the following formula: ; In the formula: As a trend factor for heavy metal pollution in farmland; To perceive the total number of data types; , The content of the m-th heavy metal in the sensing data from two consecutive heavy metal sensor operations; To configure weights; The above formula comprehensively quantifies the changing trend of heavy metal content. It retrieves the latest four sensing data to calculate the change between two consecutive data points and uses... , The ratio of these factors, combined with the weights that are positively correlated with the degree of harm from heavy metals, highlights the contribution of key heavy metals and comprehensively reflects the changes in multiple heavy metals. This allows the trend factor to objectively and accurately reflect the pollution dynamics and provide a scientific basis for risk assessment. The result of calculating X1 and X2 is denoted as The calculation results of X2 and X3 are denoted as The calculation results of X3 and X4 are denoted as ; Once the location for the heavy metal sensors is determined, the burial depth is set to half the typical length of the rootstock of the plants grown in the farmland. All are positive numbers and and obey The greater the harm of a particular type of heavy metal to farmland, the higher its value. It should be noted that: The typical length of the rootstock of crops planted in farmland is determined by calculating the average value of the rootstock after the system user samples the crops planted in farmland. It is a broad parameter that allows for deviations and can also be customized by the user. ; In the formula: This represents the total number of heavy metal sensors deployed for each of the three local models. Let m be the content of the m-th heavy metal sensed by the p-th heavy metal sensor in this cycle; This is the proportionality coefficient. , This indicates the number of heavy metal sensors deployed in the local model with the highest position. The above formula is designed to address the data integration bias caused by differences in sensor density in different local models. It calculates the average content of the region using the total number of sensors in each local model, G1, G2, and G3, and then weights it using a proportional coefficient to ensure that regions with different densities have a reasonable proportion in the integration result, matching the actual spatial monitoring characteristics. Furthermore, it sets a specific proportional coefficient based on the differences in sensor density to provide accurate basic data for comparing adjacent sensing data and ensure the authenticity of the pollution trend factor calculation. , The calculation logic is the same; , , After obtaining, based on , , The change rates of two trend factors of heavy metal pollution in farmland were obtained. , recorded as ; The early warning module is used to continuously acquire the heavy metal pollution trend factors of farmland in the assessment module, determine whether there is a risk of heavy metal pollution in farmland based on the heavy metal pollution trend factors of farmland, and trigger an early warning based on the determination result. During the early warning module operation phase, the latest three outputs of the assessment module are obtained as farmland heavy metal pollution trend factors. When the farmland heavy metal pollution trend factor continues to rise or the change rate of any farmland heavy metal pollution trend factor is greater than the preset threshold, it is determined that there is a risk of heavy metal pollution in the farmland. The early warning module integrates an audible and visual alarm device. When the determination result is that there is a risk of heavy metal pollution in farmland, the audible and visual alarm device will issue a preset audible and visual alarm. The creation module interacts with the logic module and the retrieval unit via a wireless network. The logic module interacts with the control and storage module via a wireless network. The control and storage module interacts with the evaluation module via a wireless network. The evaluation module interacts with the early warning module via a wireless network.

[0023] In this embodiment, the creation module uploads the structural information of the farmland area and creates a farmland twin model based on the structural information. The logic module runs subsequently to generate heavy metal sensor deployment logic based on the farmland twin model. The retrieval unit synchronously accesses the creation module and retrieves the farmland twin model from the creation module. The control and storage module further controls the heavy metal sensors to operate according to a preset cycle and receives the operating perception data of each heavy metal sensor during operation. The operating perception data of the heavy metal sensors is stored. Then, the evaluation module obtains the perception data stored in the control and storage module, evaluates the farmland heavy metal pollution trend factor based on the perception data, and finally, the early warning module continuously obtains the farmland heavy metal pollution trend factor from the evaluation module, determines whether there is a risk of heavy metal pollution in the farmland based on the farmland heavy metal pollution trend factor, and triggers an early warning based on the determination result.

[0024] The following is an application example of the system described in the above embodiments: A provincial agricultural technology extension center deployed this system to monitor and provide risk warnings for heavy metal pollution (focusing on cadmium, lead, and mercury) in a rice paddy of approximately 15 mu (about 1 hectare). The system was designed to ensure the safety of rice cultivation by monitoring heavy metal pollution in the paddy field, with a focus on monitoring three common harmful heavy metals: cadmium, lead, and mercury. The system was designed for a 15-mu (about 1 hectare) rice paddy with a one-to-one correspondence between the paddy field's edge coordinates and the paddy field's edge coordinates, where a chemical company regularly discharges pollutants. The paddy field itself has a terrain that slopes from north to south, with a flat central area.

[0025] In the initial stage of system deployment, staff uploaded the structural information of the farmland area through a creation module. Three representative points were selected for the farmland edge coordinates: 30°12′05″N, 118°25′10″E; 30°12′10″N, 118°25′15″E; and 30°12′08″N, 118°25′08″E. The corresponding farmland edge coordinates corresponded one-to-one with these three farmland edge coordinates, and the straight-line distance between each pair of corresponding coordinates was 100 meters (calculated, the farmland's maximum radius is approximately 80 meters and its maximum diameter is approximately 160 meters; 100 meters satisfies the requirement of being "greater than the maximum radius and less than the maximum diameter"). Connecting the three sets of farmland edge coordinates together forms a closed finite surface A, and connecting them again forms a closed finite surface B. The area of ​​B has the smallest difference from the actual area of ​​the farmland, and from a top-down perspective, the edge outline of B is completely within the range of A.

[0026] After entering the calibration phase, the system user, considering the high-precision requirements for heavy metal monitoring in rice cultivation, set the preset value for comparing the angle between A and B to 5°. The system calculation showed that the actual angle between A and B was 3°, less than the preset value of 5°. Therefore, B was directly used as the final farmland twin model, completing the core operation of the creation module.

[0027] After the creation module is completed, the logic module initiates its internal retrieval unit, accesses the creation module, and retrieves the calibrated farmland twin model. Using surface A as the planar reference surface, the system divides the farmland twin model into three local models along the direction from the highest position (northern highlands) to the lowest position (southern depression): the northern highland local model, the central flat local model, and the southern lowland local model. The system-side user-defined preset deployment quantity of heavy metal sensors, x, is 4 (a non-zero even number). According to the deployment logic: the northern highland local model (highest position) deploys 4 heavy metal sensors, evenly distributed in a 2×2 matrix; the southern lowland local model (lowest position) deploys 8 heavy metal sensors (2x), evenly distributed in a 4×2 matrix; and the central flat local model (medium position) deploys 6 heavy metal sensors (between 4 and 8), evenly distributed in a 3×2 matrix. Simultaneously, based on the typical length of rice rootstock (approximately 40 cm), the staff uniformly sets the burial depth of all heavy metal sensors to 20 cm (half the typical rootstock length).

[0028] After the sensors are deployed, the control and storage module begins operation. System users, considering the actual conditions of the farmland, set the preset operating cycle of the heavy metal sensors to 7 days / cycle, and also set up an "event-triggered operation" mechanism—the sensors run an additional time each time an event occurs within the farmland area (area A) such as a chemical plant discharging wastewater (the plant discharges wastewater on the 15th of each month), farmland rainfall (after the local meteorological station detects effective rainfall), or farmland fertilizer and pesticide irrigation (after staff record irrigation operations). For example, the system starts sensor operation according to the preset cycle on May 1st (cycle 1), the sensor runs again after local rainfall on May 10th (cycle 2), the sensor runs a third time after the chemical plant discharges wastewater on May 15th (cycle 3), and the sensor runs a fourth time after the farmland completes irrigation on May 20th (cycle 4). After each sensor operation, the system receives the sensing data from each sensor (including the cadmium, lead, and mercury content values ​​at different locations) and stores it according to the rule of "sensor deployment location + sensing cycle number": first, the intervals are named according to the deployment location of each sensor, such as "North-1", "North-2", "Central-1"... "South-8"; then, the secondary intervals are named according to the sensing cycle number, such as "Cycle 1 (May 1)", "Cycle 2 (May 10)", "Cycle 3 (May 15)", "Cycle 4 (May 20)", to ensure that the data storage is clear and traceable.

[0029] The assessment module periodically retrieves sensing data from the control and storage module to calculate trend factors for heavy metal pollution in farmland. This assessment retrieved the latest four sets of sensing data (cycles 1 to 4) (denoted as X1, X2, X3, and X4), focusing on cadmium, lead, and mercury (total number of sensing data types M=3). After processing, the system obtained the weighted average content of each heavy metal in X1, X2, X3, and X4, which served as the basis for calculating the trend factors. Based on the degree of harm of the three heavy metals to farmland, the system set weights ωm: mercury is the most harmful, with a weight of 0.4; lead is second, with a weight of 0.3; and cadmium has a weight of 0.3 (the sum of the three is 1). Finally, three sets of trend factors for heavy metal pollution in farmland were obtained: the calculation result for X1 and X2 is 0.12, the calculation result for X2 and X3 is 0.15, and the calculation result for X3 and X4 is 0.18. Simultaneously, the system further calculated the change rates of the trend factors twice: the first change rate was 25%, and the second change rate was 20%.

[0030] The early warning module continuously acquires the trend factors and change rates output by the assessment module. Based on local heavy metal safety standards for rice cultivation, the system user sets the "preset threshold for the change rate of heavy metal pollution trend factors" to 15%. The early warning module detected three consecutive trend factors (0.12, 0.15, 0.18) showing a continuous upward trend, and two change rates (25% and 20%) exceeded the preset threshold of 15%, thus determining that the farmland was at risk of heavy metal pollution. Immediately, the integrated audible and visual alarm device activated, emitting a continuously flashing red light visual alarm and simultaneously playing a voice alarm stating, "Farmland heavy metal pollution risk warning; please immediately investigate the source of pollution and take control measures." Upon receiving the alarm, staff from the Agricultural Technology Extension Center quickly went to the farmland and, combining data from sensor deployment points and the discharge situation of chemical plants within area A, investigated the source of pollution, providing accurate data for subsequent pollution control and remediation.

[0031] In summary, during operation, the system in the above embodiments constructs and calibrates a farmland twin model using structural information of farmland and its surrounding area, ensuring a high degree of matching between the model and the actual farmland. This lays a precise foundation for monitoring. Heavy metal sensors are deployed in differentiated areas based on the terrain features of the twin model, with higher density at lower locations. This enables comprehensive and targeted monitoring of all areas of farmland. The sensor operating cycle can be adapted to events such as factory discharges and rainfall. Combined with preset cycles, this ensures no data is missed at key nodes. Sensing data is categorized and stored according to sensor location and acquisition cycle for easy retrieval and analysis later. Based on multiple sets of historical data and combined with heavy metal hazard weights, pollution trend factors are calculated, and the rate of trend change is monitored simultaneously to accurately assess pollution trends. When the trend continues to rise or the rate of change exceeds a threshold, an audible and visual warning is triggered to promptly remind and prevent pollution. This effectively improves the accuracy, comprehensiveness, and timeliness of farmland heavy metal pollution monitoring, providing support for farmland pollution prevention and control.

[0032] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A monitoring system for heavy metal pollution in farmland, characterized in that, include: Create a module to upload the structural information of the farmland area and create a farmland twin model based on the structural information of the farmland area; The logic module is used to generate heavy metal sensor deployment logic based on the farmland twin model; The control and storage module is used to control the heavy metal sensors to operate according to a preset cycle, and to receive the operating sensing data of each heavy metal sensor during operation and store the operating sensing data of the heavy metal sensors. The assessment module is used to acquire the sensing data stored in the control and storage module, and to assess the trend factors of heavy metal pollution in farmland based on the sensing data. The early warning module is used to continuously acquire the heavy metal pollution trend factors of farmland in the assessment module, determine whether there is a risk of heavy metal pollution in farmland based on the heavy metal pollution trend factors, and trigger an early warning based on the determination result.

2. The farmland heavy metal pollution monitoring system according to claim 1, characterized in that, The structural information of the farmland area uploaded during the creation module operation phase includes: farmland edge coordinates and farmland edge coordinates centered on the farmland area. The farmland edge coordinates correspond one-to-one with each other, and the distance between each corresponding farmland edge coordinate is equal. Furthermore, the straight-line distance between the corresponding farmland edge coordinates is greater than the maximum radius of the farmland and less than the maximum diameter of the farmland. The farmland's edge coordinates and the farmland's edge coordinates are both three. The three farmland edge coordinates are connected to form a closed finite surface, denoted as A. The three farmland edge coordinates are connected to form a closed finite surface, denoted as B. Based on A, B is calibrated to complete the creation of the farmland twin model. Among them, from the top-down view, the edge contour of B is within the range of A, and the three uploaded farmland edge coordinates follow the principle that the difference between the area of ​​B determined based on the three farmland edge coordinates and the actual farmland area is the smallest.

3. The farmland heavy metal pollution monitoring system according to claim 2, characterized in that, The calibration operation based on A for B is as follows: Identify the angle between A and B. If the angle between A and B is less than a preset value, B will be directly used as the farmland twin model. When the angle between A and B is greater than a preset threshold, B is calibrated based on A, and the calibrated B is used as the farmland twin model: ； In the formula: The calibration angle for B; Let A be the angle between A and B; For calibration coefficients; The preset value used for comparing the angle between A and B is user-defined on the system side, and it follows the principle that the higher the accuracy requirement for monitoring heavy metal pollution in farmland, the smaller the preset value, and vice versa. After calculating the calibration angle of B based on the above formula, the direction in which B and A form an angle is used as the flipping direction, and the intersection line of the angle between B and A is used as the flipping axis. Perform a flip operation to complete the calibration.

4. The farmland heavy metal pollution monitoring system according to claim 3, characterized in that, The calibration coefficient The value is determined based on the degree of difference between the farmland's location and the soil properties of the farmland. The center point of B is selected. Connecting the corresponding edge coordinates of B and A and extending to the center point yields three line segments. These three line segments are used to divide the farmland area into regions corresponding to A and B, with the farmland edge as the boundary. These regions are denoted as a1, a2, a3 and b1, b2, b3, respectively. Soil composition sensors are uniformly deployed within a1, a2, a3 and b1, b2, b3 to detect the soil composition. The calibration coefficients are then calculated. ； In the formula: 3 represents the number of control groups, including a1 and b1, a2 and b2, and a3 and b3; The total number of soil component types detected in the i-th control group; These represent the proportion of the nth component in the i-th region of the farmland and the proportion of the nth component in the i-th region of the farmland, respectively. The allocation weights for the nth component; in, The number of soil component types corresponding to the region with the most soil component types in the i-th control group is selected. When the value is zero, the value of the fraction inside the parentheses is taken as zero. , To preset the minimum value, All are positive numbers, and their values ​​satisfy... Furthermore, the smaller the maximum difference in content between the nth component and the components sensed by each soil component sensor, the better. The larger the value, the lower the value. The smaller the value, All values ​​are the average content of the nth component sensed by each soil component sensor, with a simultaneously set constraint interval of [0.75, 1.25]. The values ​​are calculated based on the above formula. If it is within the constrained range, then upload based on the above formula. Conversely, take the one closer to... The boundary value of the constraint interval, the soil composition sensor detects that the soil composition does not contain heavy metal components.

5. A farmland heavy metal pollution monitoring system according to claim 2, characterized in that, The logic module is equipped with a retrieval unit, which is used to access the creation module and retrieve the farmland twin model from the creation module. The farmland twin model retrieved by the retrieval unit is the calibrated farmland twin model. The heavy metal sensor deployment logic generated in the logic module is as follows: Using A as the plane reference plane, the farmland twin model is divided into three local models from the highest position to the lowest position. Heavy metal sensors are uniformly deployed in a matrix in the three local models. The goal is to achieve the sparsest density of heavy metal sensors in the local model with the highest position, the densest density in the local model with the lowest position, and a density between sparse and dense in the local model with the middle position. The number of heavy metal sensors deployed in the local model with the highest position is a preset number x, which is defined by the system user and is a non-zero even number. The number of heavy metal sensors deployed in the local model with the lowest position is 2x; The number of heavy metal sensors deployed in a moderately located local model is 6. The farmland heavy metal pollution monitoring system according to claim 1, characterized in that, The preset cycle for the control and storage module to control the operation of the heavy metal sensor is defined by the system user and is adapted to the insertion of events such as sewage discharge cycle of factories around farmland, rainfall, and fertilizer and pesticide irrigation. That is, it runs once after each factory sewage discharge, rainfall, and fertilizer and pesticide irrigation event occurs, and factories within the area where the farmland is located are identified as surrounding factories. When storing the sensing data from the heavy metal sensors, the data is stored separately based on the heavy metal sensors from which the data originates. After storage separately, the data is stored separately again based on the sensing cycle from which the data originates. When storing separately based on the heavy metal sensors from which the data originates, the deployment location of each heavy metal sensor is used as the name of the distinguishing interval. When storing separately based on the sensing cycle from which the data originates, the sequence number of each sensing cycle is used as the name of the distinguishing interval.

7. The farmland heavy metal pollution monitoring system according to claim 1, characterized in that, The heavy metal sensor data includes several sets of heavy metal content values. When the assessment module evaluates the trend factor of heavy metal pollution in farmland, it follows the following: Retrieve the sensing data from the four most recent heavy metal sensor runs in the evaluation module, denoted as X1, X2, X3, and X4. Calculate the heavy metal pollution trend factor in farmland using the sensing data from each of the two consecutive heavy metal sensor runs, based on the following formula: ； In the formula: As a trend factor for heavy metal pollution in farmland; To perceive the total number of data types; , The content of the m-th heavy metal in the sensing data from two consecutive heavy metal sensor operations; To configure weights; The result of calculating X1 and X2 is denoted as The calculation results of X2 and X3 are denoted as The calculation results of X3 and X4 are denoted as ; Once the location for the heavy metal sensors is determined, the burial depth is set to half the typical length of the rootstock of the plants grown in the farmland. All are positive numbers and and obey The greater the harm of a particular type of heavy metal to farmland, the higher its value.

8. A farmland heavy metal pollution monitoring system according to claim 7, characterized in that, The ; In the formula: This represents the total number of heavy metal sensors deployed for each of the three local models. Let m be the content of the m-th heavy metal sensed by the p-th heavy metal sensor in this cycle; This is the proportionality coefficient. , This indicates the number of heavy metal sensors deployed in the local model with the highest position. The , The calculation logic is the same; The , , After obtaining, based on , , The change rates of two trend factors of heavy metal pollution in farmland were obtained. , recorded as .

9. A farmland heavy metal pollution monitoring system according to claim 1, characterized in that, During the operation of the early warning module, the latest three outputs of the assessment module are obtained as farmland heavy metal pollution trend factors. When the farmland heavy metal pollution trend factor continues to rise or the change rate of any farmland heavy metal pollution trend factor is greater than a preset threshold, it is determined that there is a risk of heavy metal pollution in the farmland. The early warning module integrates an audible and visual alarm device. When the determination result indicates that there is a risk of heavy metal pollution in the farmland, the audible and visual alarm device will issue a preset audible and visual alarm.

10. A farmland heavy metal pollution monitoring system according to claim 1, characterized in that, The creation module is interconnected with the logic module and the retrieval unit via a wireless network. The logic module is interconnected with the control and storage module via a wireless network. The control and storage module is interconnected with the evaluation module via a wireless network. The evaluation module is interconnected with the early warning module via a wireless network.

Images