Cultivated land quality monitoring system based on cultivated land protection

Abstract

The invention provides a cultivated land quality monitoring system based on cultivated land protection. The cultivated land quality monitoring system comprises a multi-source sensing module, a heterogeneous transmission module, a space-time database module, a quality degradation prediction module, a dynamic early warning module, a repair decision module and a quality traceability block chain module. The multi-source sensing module collects data such as soil pH value, organic matter content, heavy metal concentration, crop canopy temperature and vegetation index; the heterogeneous transmission module carries out encryption packaging and format standardization on the data; the space-time database module stores the processed data and space-time associated information; the quality degradation prediction module generates multi-dimensional monitoring data and predicts a degradation risk; the dynamic early warning module generates a multi-stage early warning signal according to the risk value; the restoration decision module outputs soil improvement engineering technical parameters and ecological compensation schemes; and the quality traceability block chain module realizes data verifiable evidence storage. The intelligent level of cultivated land protection can be improved, and comprehensive technical support is provided for sustainable utilization of cultivated land quality.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural informatization and ecological environment monitoring, and more specifically, to a farmland quality monitoring system based on farmland protection. Background Art

[0002] In modern agricultural production, the protection and monitoring of arable land quality are crucial. With the acceleration of industrialization and urbanization, arable land faces numerous challenges, including soil pollution, declining fertility, and ecological degradation. Traditional methods for monitoring arable land quality rely primarily on manual sampling and laboratory analysis, which is not only time-consuming and labor-intensive but also hinders real-time, dynamic monitoring of arable land quality. In recent years, with the advancement of sensor technology, remote sensing technology, and geographic information system (GIS) technology, arable land quality monitoring has gradually moved towards automation and intelligentization. However, existing technologies still lack comprehensiveness in data collection, efficient data processing, and the applicability of monitoring results. For example, data from a single source cannot fully reflect the multidimensional characteristics of arable land quality; data processing lacks standardization and security, which can easily lead to information loss or leakage; and the application of monitoring results is often superficial, lacking effective integration with remediation decisions. Furthermore, the traceability and verification mechanisms for arable land quality monitoring data are imperfect, making it difficult to meet public demand for transparent arable land quality information.

[0003] In the process of implementing the embodiments of the present invention, the inventors found that there are at least the following problems or defects in the existing technology: the existing monitoring system cannot effectively integrate multi-source data, making it difficult to achieve comprehensive and dynamic monitoring of arable land quality; there is a lack of standardization and security mechanisms in the data processing and transmission process, resulting in insufficient data reliability; there is a lack of effective correlation between monitoring results and restoration decisions, making it difficult to directly guide the precise restoration of arable land quality; at the same time, there is a lack of effective data traceability and verification mechanisms, which cannot meet the transparency and traceability requirements of arable land quality information. Summary of the Invention

[0004] The present invention provides a farmland quality monitoring system based on farmland protection, comprising:

[0005] A multi-source sensing module is used to collect farmland monitoring data, including soil pH, organic matter content, heavy metal concentration, crop canopy temperature, and vegetation index data;

[0006] A heterogeneous transmission module, used for encrypting, encapsulating and formatting the farmland monitoring data output by the multi-source perception module;

[0007] A spatiotemporal database module, configured to store the farmland monitoring data processed by the heterogeneous transmission module and its spatiotemporal correlation information, wherein the spatiotemporal correlation information includes Beidou coordinates of monitoring points, data acquisition timestamps, and multi-source data mapping relationships;

[0008] a quality degradation prediction module, configured to fuse the processed farmland monitoring data and its spatiotemporal correlation information to generate multi-dimensional farmland monitoring data, and to generate a farmland quality degradation risk prediction value based on the multi-dimensional farmland monitoring data;

[0009] Dynamic early warning module, used to generate multi-level early warning signals based on the predicted value of cultivated land quality degradation risk and preset cultivated land quality thresholds;

[0010] Restoration decision-making module, used to output soil improvement engineering technical parameters and ecological compensation plans based on the predicted value of cultivated land quality degradation risk;

[0011] The quality traceability blockchain module is used to achieve verifiable storage of cultivated land monitoring data.

[0012] Furthermore, the multi-source perception module includes:

[0013] In-situ sensor array units, deployed at farmland monitoring points, include:

[0014] Ring electrode conductivity sensor, measuring the salt concentration in the 0-50cm soil layer;

[0015] Laser-induced breakdown spectroscopy sensor to detect the content of arsenic, cadmium, lead, mercury, and chromium;

[0016] Near-infrared spectral reflectance probe to continuously monitor the dynamic changes of soil organic matter;

[0017] UAV remote sensing unit equipped with a hyperspectral imager to obtain crop stress characteristic data;

[0018] Satellite analysis unit receives and processes NDVI index, surface temperature and soil moisture inversion data.

[0019] Furthermore, the heterogeneous transmission module performs:

[0020] The cultivated land monitoring data output by the in-situ sensor array unit is subjected to wavelet threshold denoising processing, and the calculation formula is:

[0021]

[0022] in, is the denoised signal, ψ j,k is the Daubechies wavelet basis function, j is the number of decomposition layers, k is the translation coefficient, is the adaptive threshold, σ is the noise standard deviation, N is the signal length, and I(·) is the indicator function;

[0023] The cultivated land monitoring data output by the UAV remote sensing unit is geo-referenced to generate an encrypted data packet with Beidou coordinates.

[0024] Furthermore, the quality degradation prediction module includes:

[0025] The spatiotemporal correlation analysis submodule calculates the coupling relationship between soil parameters and crop growth indicators:

[0026]

[0027] Among them, R sc is the soil-crop correlation, S i is the i-th soil parameter, is the mean value of soil parameters, C i is the corresponding crop growth index, is the mean value of crop indicators;

[0028] The degradation rate prediction submodule uses a multi-scale convolutional gated network to process the multi-dimensional cultivated land monitoring data and output the cultivated land quality change rate in the next 6 months.

[0029] Furthermore, the calculation steps of the multi-scale convolutional gated network include:

[0030] (a) Input the soil conductivity time series data in the spatiotemporal database module into the temporal convolution layer to extract the interannual variation feature vector H temp ;

[0031] (b) The UAV hyperspectral data in the spatiotemporal database module is input into the spatial convolution layer to extract the field-level spatial heterogeneity feature vector H spat ;

[0032] (c) Generate degradation risk index through gated fusion unit:

[0033] G t =σ(W g ·[H temp ,H spat ])

[0034] Among them, G t is the time gating weight vector, W g is the trainable parameter matrix, and σ is the sigmoid function.

[0035] Furthermore, the dynamic warning module performs:

[0036] When the heavy metal concentration in the predicted value of farmland quality degradation risk exceeds 80% of the screening value of the "Agricultural Land Soil Pollution Risk Control Standard", a first-level warning signal is generated;

[0037] When the predicted value of soil organic matter content decreases at a rate of more than 5% / month for three consecutive months and the crop NDVI decreases at a rate of more than 15%, a second-level warning signal will be generated;

[0038] The early warning signal triggers the cross-region compensation calculation of the repair decision module.

[0039] Furthermore, the quality traceability blockchain module includes:

[0040] The farmland quality fingerprint unit generates verifiable credentials including soil parameter hash values, remote sensing image feature values, and early warning records;

[0041] Smart contract unit, implements the following farmland quality constraint rules:

[0042] When the quality grade of newly added cultivated land is lower than that of occupied cultivated land, the land change approval process will be frozen.

[0043] When the progress of the soil remediation project does not reach 90% of the planned value, the ecological compensation deposit will be automatically allocated;

[0044] Distributed storage unit, synchronously stores the original data of quality assessment at the provincial farmland protection node.

[0045] Furthermore, the method for generating the farmland quality fingerprint unit includes:

[0046] The farmland monitoring data of the in-situ sensor array unit for 30 consecutive days is compressed into a feature matrix M sensor , where M sensor It is an m×n dimensional matrix, where m represents the number of monitoring days and n represents the number of sensor types;

[0047] Calculate the vegetation anomaly index from UAV multispectral data:

[0048]

[0049] Among them, NDVI current is the current normalized vegetation index, NDVI baseline is the normalized difference vegetation index of the historical base period;

[0050] Generate a unique fingerprint:

[0051] F print =SHA256(M sensor ||V anomaly ||t block )

[0052] Among them, || represents the data splicing operation, t block It is the blockchain timestamp and SHA256 is the secure hash algorithm.

[0053] Furthermore, the repair decision module includes:

[0054] The salt-alkali treatment submodule generates parameters for the underground pipe salt drainage project based on the predicted value of the farmland quality degradation risk:

[0055]

[0056] Where, L is the distance between concealed pipes (meters), K is the soil permeability coefficient (m / d), t is the design drainage cycle (days), H0 is the initial groundwater level (meters), H t is the target groundwater level (m), ΔS is the salt concentration gradient correction factor;

[0057] The heavy metal remediation submodule calculates the ratio between the amount of biochar applied and the chelating agent spraying concentration based on the predicted value of the farmland quality degradation risk.

[0058] Furthermore, the calculation of the amount of biochar applied includes:

[0059] Establishment of heavy metal adsorption kinetic model:

[0060]

[0061] Among them, Q e is the equilibrium adsorption capacity (mg / g), C e is the equilibrium concentration of heavy metals (mg / L), k f is the Freundlich adsorption coefficient, n is the nonlinear exponent;

[0062] According to the target repair concentration C target Reverse the minimum application rate:

[0063]

[0064] Among them, V soil is the volume of treated soil (m 3 ), ρ is soil bulk density (g / cm 3 ), C0 is the initial pollution concentration (mg / kg).

[0065] The above-described embodiments of the present invention have at least the following beneficial effects: First, through the deployment of a multi-source perception module, the present invention enables real-time collection of multi-dimensional data on cultivated land quality, including soil parameters, crop growth indicators, and remote sensing imagery, thereby comprehensively reflecting the dynamic changes in cultivated land quality. Second, the heterogeneous transmission module encrypts and encapsulates the collected data and standardizes its format, ensuring data security and consistency during transmission and providing a reliable data foundation for subsequent data storage and analysis. Third, the spatiotemporal database module stores processed monitoring data and its spatiotemporal correlation information, providing powerful data support for spatiotemporal analysis of cultivated land quality. Furthermore, the quality degradation prediction module generates degradation risk prediction values based on multi-dimensional cultivated land monitoring data, providing early warning of changing trends in cultivated land quality and providing a basis for scientific decision-making. Finally, the restoration decision-making module outputs soil improvement engineering technical parameters and ecological compensation plans based on the degradation risk prediction values, achieving precise restoration and ecological compensation for cultivated land quality, and enhancing the scientific nature and effectiveness of cultivated land protection.

[0066] At the same time, the introduction of a quality traceability blockchain module provides technical support for the verifiable storage of farmland monitoring data. By generating farmland quality fingerprints and binding rules for smart contract units, the authenticity, integrity, and immutability of farmland quality data can be ensured, strengthening public trust in farmland quality information. Furthermore, distributed storage units synchronously store raw quality assessment data at provincial farmland protection nodes, further enhancing data security and transparency, providing strong support for the long-term monitoring and sustainable utilization of farmland quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] The above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily apparent by reading the following detailed description with reference to the accompanying drawings, in which several embodiments of the present invention are shown by way of example and not limitation, in which:

[0068] Figure 1 This is a structural diagram of a farmland quality monitoring system based on farmland protection provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0069] The principles and spirit of the present invention will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided solely to enable those skilled in the art to better understand and implement the present invention, and are not intended to limit the scope of the present invention in any way. Rather, these embodiments are provided to make the present invention more thorough and complete, and to fully convey the scope of the present invention to those skilled in the art.

[0070] Those skilled in the art will appreciate that the embodiments of the present invention may be implemented as a system, apparatus, device, method, or computer program product. Therefore, the present invention may be implemented in the following forms: entirely in hardware, entirely in software (including firmware, resident software, microcode, etc.), or in a combination of hardware and software.

[0071] It should be noted that any number of elements in the drawings is for illustration only and not for limitation, and any naming is only for distinction and does not have any limiting meaning.

[0072] Reference below Figure 1 , Figure 1 This is a structural diagram of a farmland quality monitoring system based on farmland protection provided by an embodiment of the present invention. Figure 1 As shown, a farmland quality monitoring system 100 based on farmland protection includes:

[0073] The multi-source sensing module 101 is used to collect farmland monitoring data, including soil pH value, organic matter content, heavy metal concentration, crop canopy temperature and vegetation index data;

[0074] The heterogeneous transmission module 102 is used to encrypt, encapsulate and standardize the cultivated land monitoring data output by the multi-source perception module;

[0075] The spatiotemporal database module 103 is used to store the farmland monitoring data processed by the heterogeneous transmission module and its spatiotemporal correlation information, wherein the spatiotemporal correlation information includes the Beidou coordinates of the monitoring points, the data collection timestamps and the multi-source data mapping relationship;

[0076] The quality degradation prediction module 104 is configured to fuse the processed farmland monitoring data and its spatiotemporal correlation information to generate multi-dimensional farmland monitoring data, and generate a farmland quality degradation risk prediction value based on the multi-dimensional farmland monitoring data;

[0077] Dynamic early warning module 105, for generating multi-level early warning signals according to the predicted value of farmland quality degradation risk and the preset farmland quality threshold;

[0078] The restoration decision module 106 is used to output soil improvement engineering technical parameters and ecological compensation plans based on the predicted value of cultivated land quality degradation risk;

[0079] The quality traceability blockchain module 107 is used to realize the verifiable storage of cultivated land monitoring data.

[0080] It should be noted that the core of this system is the collection of farmland monitoring data through a multi-source sensing module. This data includes soil pH, organic matter content, heavy metal concentration, crop canopy temperature, and vegetation index. Soil pH is a key indicator of soil acidity and alkalinity, commonly used to assess soil chemical properties and suitability; organic matter content reflects soil fertility and is a key parameter of soil quality; and heavy metal concentration is used to monitor soil contamination and ensure farmland environmental safety. Crop canopy temperature and vegetation index (such as NDVI) are crop growth indicators obtained through remote sensing technology, reflecting crop growth status and stress conditions. The heterogeneous transmission module encrypts and encapsulates the collected farmland monitoring data and standardizes its format to ensure data security and consistency during transmission, facilitating subsequent processing and storage. The spatiotemporal database module stores processed farmland monitoring data and its spatiotemporal association information. This information includes the Beidou coordinates of the monitoring points, data acquisition timestamps, and multi-source data mapping relationships, providing a foundation for spatiotemporal data analysis. The quality degradation prediction module integrates processed farmland monitoring data and its spatiotemporal correlation information to generate multi-dimensional farmland monitoring data. Based on this data, it generates a predicted farmland quality degradation risk value, providing a basis for subsequent early warning and remediation efforts. The dynamic early warning module generates multi-level warning signals based on the predicted farmland quality degradation risk value and preset farmland quality thresholds, enabling timely identification of potential problems and the implementation of appropriate measures. The remediation decision module outputs soil improvement engineering technical parameters and ecological compensation plans based on the predicted farmland quality degradation risk value, providing specific guidance for remediation of farmland quality. The quality traceability blockchain module verifies and stores farmland monitoring data, ensuring its authenticity and immutability, and providing technical support for the long-term monitoring of farmland quality.

[0081] Specifically, the in-situ sensor array units in the multi-source perception module are deployed at farmland monitoring sites and include a variety of sensors. For example, a ring electrode conductivity sensor is used to measure salinity in the 0-50cm soil layer. Salt concentration is a key indicator of soil quality; both high and low salt concentrations can affect crop growth. Laser-induced breakdown spectroscopy sensors can detect the levels of heavy metals such as arsenic, cadmium, lead, mercury, and chromium. Excessive accumulation of these elements can pollute soil and crops, impacting food safety and the ecological environment. Near-infrared spectral reflectance probes are used to continuously monitor the dynamic changes in soil organic matter. Organic matter is a core indicator of soil fertility, and its changes directly reflect the evolving trends in soil quality. The drone remote sensing unit, equipped with a hyperspectral imager, can capture crop stress signature data, such as changes in crop spectral reflectance. This data can indicate whether crops are experiencing drought, pests and diseases, or other environmental stresses. The satellite analysis unit receives and processes NDVI, surface temperature, and soil moisture inversion data. The NDVI is a key indicator of vegetation growth, while surface temperature and soil moisture reflect the thermal and moisture conditions of the soil. Together, these data constitute a multi-dimensional data system for monitoring arable land quality. In the heterogeneous transmission module, wavelet threshold denoising is applied to the data output by the in-situ sensor array units. This process uses specific mathematical formulas to remove noise interference and ensure data accuracy and reliability. Data output by the drone remote sensing unit is georeferenced, generating encrypted data packets with Beidou coordinates for precise positioning and secure transmission.

[0082] The spatiotemporal correlation analysis submodule within the quality degradation prediction module can be further refined. For example, when calculating the coupled relationship between soil parameters and crop growth indicators, additional soil parameters and crop growth indicators, such as soil moisture, soil temperature, and crop chlorophyll content, can be incorporated to more comprehensively reflect the interaction between soil and crops. The multi-scale convolutional gating network employed in the degradation rate prediction submodule can be optimized for the characteristics of cultivated land in different regions. For example, in saline-alkali areas, sensitivity analysis of soil salinity changes can be added; in areas with heavy metal pollution, the analysis focuses on time-series changes in heavy metal concentrations. Furthermore, the warning signal generation conditions within the dynamic early warning module can be adjusted based on regional cultivated land quality standards and management needs. For example, in ecologically fragile areas, the warning threshold for heavy metal concentrations can be set lower to identify potential risks earlier and enable action to be taken. The salinization and alkali management submodule within the remediation decision module can adjust the calculation formula for the parameters of the underground salt drainage system based on different soil types and climatic conditions to accommodate different management needs. For example, in arid regions, the drainage cycle can be appropriately increased to reduce water waste.

[0083] In some embodiments, the multi-source perception module includes:

[0084] In-situ sensor array units, deployed at farmland monitoring points, include:

[0085] Ring electrode conductivity sensor, measuring the salt concentration in the 0-50cm soil layer

[0086] Laser-induced breakdown spectroscopy sensor, which detects the content of arsenic, cadmium, lead, mercury, and chromium; near-infrared spectroscopy reflectance probe, which continuously monitors the dynamic changes of soil organic matter;

[0087] UAV remote sensing unit equipped with a hyperspectral imager to obtain crop stress characteristic data;

[0088] Satellite analysis unit receives and processes NDVI index, surface temperature and soil moisture inversion data.

[0089] It should be noted that the multi-source sensing module in the present invention includes an in-situ sensor array unit, an unmanned aerial vehicle (UAV) remote sensing unit, and a satellite analysis unit. Together, these units form a comprehensive farmland monitoring data acquisition system. The in-situ sensor array unit is deployed at farmland monitoring sites, enabling real-time monitoring directly within the soil environment. The ring electrode conductivity sensor is used to measure salinity in the 0-50 cm soil layer. Salt concentration is a key indicator of soil quality; excessively high or low salt concentrations can affect crop growth. The laser-induced breakdown spectroscopy sensor can detect the levels of heavy metals such as arsenic, cadmium, lead, mercury, and chromium. Excessive accumulation of these heavy metals can pollute soil and crops, further impacting food safety and the ecological environment. The near-infrared spectral reflectance probe is used to continuously monitor the dynamic changes in soil organic matter. Organic matter is a core indicator of soil fertility, and its changes directly reflect the evolving trends in soil quality. The UAV remote sensing unit is equipped with a hyperspectral imager, which can capture crop stress characteristic data, such as changes in crop spectral reflectance. This data can indicate whether the crop is experiencing drought, pests and diseases, or other environmental stresses. The satellite analysis unit receives and processes NDVI index, surface temperature and soil moisture inversion data. The NDVI index is an important indicator for measuring vegetation growth conditions, while surface temperature and soil moisture reflect the thermal and moisture conditions of the soil. These data together constitute a multi-dimensional data system for arable land quality monitoring.

[0090] Specifically, the ring-electrode conductivity sensor in the in-situ sensor array unit can be set to collect data every 10 minutes to monitor dynamic changes in soil salinity. Its measurement range is typically 0-10 decimeters per meter, covering the normal range of soil salinity concentrations in most soils. Laser-induced breakdown spectroscopy (LIBS) uses high-energy laser pulses to excite elements in the soil, causing them to emit a characteristic spectrum, enabling rapid detection of heavy metals. This sensor's detection accuracy can reach microgram levels, meeting the needs of high-precision soil pollution monitoring. The near-infrared reflectance probe continuously monitors changes in soil organic matter content by measuring the soil's reflectivity to near-infrared light. This monitoring frequency can be set to once an hour to capture short-term dynamic changes in organic matter. The hyperspectral imager onboard the drone remote sensing unit captures spectral information about crops. By analyzing changes in reflectivity in specific wavelengths, it can identify crop stress. For example, when crops are experiencing drought stress, their reflectivity in the near-infrared wavelengths decreases significantly. The satellite analysis unit receives satellite remote sensing data and inverts information such as the NDVI index, surface temperature, and soil moisture. The calculation formula of the NDVI index is (NIR-R) / (NIR+R), where NIR is the reflectivity of the near-infrared band and R is the reflectivity of the red light band. This index can effectively reflect the growth status of vegetation.

[0091] To improve the adaptability and flexibility of the monitoring system, the in-situ sensor array can be optimized. For example, soil moisture and temperature sensors can be added to more comprehensively monitor soil physical properties. Soil moisture sensors can be capacitive or tensile sensors with a measurement range of 0-100% soil moisture content, providing real-time information on soil moisture conditions. Soil temperature sensors can be thermistors or platinum resistance sensors with a measurement range of -20°C to 50°C, monitoring soil temperature changes. Furthermore, the hyperspectral imager in the drone's remote sensing unit can adjust its flight altitude and frequency to suit different seasons and crop growth stages. For example, during peak crop growth, the flight altitude can be lowered to 100 meters, and the frequency increased to twice a week, to obtain more detailed information on crop stress. The satellite analysis unit can combine multiple satellite data sources, such as Landsat and Sentinel, using data fusion techniques to improve inversion accuracy. For example, combining high-resolution Landsat data with high-frequency Sentinel data can more accurately infer soil moisture and surface temperature.

[0092] In some embodiments, the heterogeneous transport module performs:

[0093] The cultivated land monitoring data output by the in-situ sensor array unit is subjected to wavelet threshold denoising processing, and the calculation formula is:

[0094]

[0095] in, is the denoised signal, ψ j,k is the Daubechies wavelet basis function, j is the number of decomposition layers, k is the translation coefficient, is the adaptive threshold, σ is the noise standard deviation, N is the signal length, and I(·) is the indicator function;

[0096] The cultivated land monitoring data output by the UAV remote sensing unit is geo-referenced to generate an encrypted data packet with Beidou coordinates.

[0097] It should be noted that the heterogeneous transmission module in this system undertakes the important task of processing and transmitting the cultivated land monitoring data output by the multi-source perception module. Among them, wavelet threshold denoising is performed on the data output by the in-situ sensor array unit, with the aim of removing noise interference in the data and improving the accuracy and reliability of the data. Wavelet threshold denoising is a signal processing method based on wavelet transform. By selecting appropriate thresholds and wavelet basis functions, it can effectively separate the useful components and noise components in the signal. The Daubechies wavelet basis function involved in the specific formula is a commonly used orthogonal wavelet basis with good time-frequency localization characteristics, which is suitable for processing complex monitoring data. The calculation formula for the adaptive threshold is: Where σ is the standard deviation of the noise, and N is the signal length. This threshold setting method dynamically adjusts the denoising strength based on the characteristics of the data. Furthermore, georeferencing is performed on the data output by the drone's remote sensing unit to combine the hyperspectral imaging data with geographic coordinates, generating encrypted data packets with Beidou coordinates, thereby achieving precise positioning and secure transmission of the data.

[0098] Specifically, in wavelet threshold denoising, the selection of Daubechies wavelet basis functions and the number of decomposition layers are crucial for denoising effectiveness. For example, the Daubechies wavelet db4 can be selected as the basis function, and the number of decomposition layers can be set to three to balance denoising effectiveness and computational complexity. The noise standard deviation σ can be obtained by estimating the local variance of the original signal, while the signal length N is determined based on the actual amount of data collected. Regarding georeferencing, hyperspectral imaging data acquired by the UAV remote sensing unit needs to be matched to Beidou coordinates using image processing algorithms. Specifically, pixel coordinates of the hyperspectral image can be converted to actual geographic coordinates through feature point matching or affine transformation-based methods. The generated encrypted data packet uses a symmetric encryption algorithm, such as AES, to ensure data security during transmission. The encrypted data packet includes not only the hyperspectral data but also the data acquisition timestamp and geographic location information for subsequent spatiotemporal analysis.

[0099] Preferably, in order to further optimize the performance of the heterogeneous transmission module, the wavelet threshold denoising algorithm can be improved. For example, an adaptive threshold adjustment mechanism can be introduced to dynamically adjust the threshold according to the local characteristics of the signal, so as to better adapt to different types of monitoring data. In addition, for the geographic registration of UAV remote sensing data, a multi-source data fusion method can be used to combine Beidou positioning data and ground control point information to improve the registration accuracy. For example, by setting a number of ground control points with known coordinates within the monitoring area, these points are used to correct the hyperspectral image, so that the registration accuracy reaches the sub-meter level. In the process of generating encrypted data packets, digital signature technology can also be introduced to ensure the integrity and non-repudiation of the data. For example, the encrypted data packet is signed using the RSA digital signature algorithm, and the recipient can confirm the source and integrity of the data by verifying the signature.

[0100] In some embodiments, the quality degradation prediction module includes:

[0101] The spatiotemporal correlation analysis submodule calculates the coupling relationship between soil parameters and crop growth indicators:

[0102]

[0103] Among them, R sc is the soil-crop correlation, S i is the i-th soil parameter, is the mean value of soil parameters, C i is the corresponding crop growth index, is the mean value of crop indicators;

[0104] The degradation rate prediction submodule uses a multi-scale convolutional gated network to process the multi-dimensional cultivated land monitoring data and output the cultivated land quality change rate in the next 6 months.

[0105] It should be noted that the quality degradation prediction module is a core component of this system. Its function is to generate multidimensional farmland monitoring data by analyzing processed farmland monitoring data and its spatiotemporal correlation information. Based on this data, it further generates farmland quality degradation risk predictions. This module includes a spatiotemporal correlation analysis submodule and a degradation rate prediction submodule. The spatiotemporal correlation analysis submodule assesses the impact of soil quality on crop growth by calculating the coupling relationship between soil parameters and crop growth indicators. The degradation rate prediction submodule uses a multi-scale convolutional gated network to process multidimensional farmland monitoring data and output the rate of change in farmland quality over a period of time. This prediction method comprehensively considers the spatiotemporal information of soil and crops, providing a scientific basis for dynamic monitoring and early warning of farmland quality.

[0106] Specifically, the coupling relationship calculation formula in the spatiotemporal correlation analysis submodule is based on the statistical correlation between soil parameters and crop growth indicators. Soil parameters include soil pH, organic matter content, and heavy metal concentrations, while crop growth indicators include crop canopy temperature and vegetation indices (such as NDVI). The mean values of soil parameters and crop growth indicators in the formula are obtained through statistical analysis of long-term monitoring data and are used to measure the average state of soil and crop. The calculated coupling relationship reflects the actual impact of soil quality changes on crop growth. For example, when soil organic matter content decreases, crop NDVI values also decrease, and changes in the coupling relationship value can quantify this impact. The multi-scale convolutional gated network in the degradation rate prediction submodule is a deep learning model capable of processing time series and spatially distributed data. The network extracts interannual variations in soil parameters through a temporal convolutional layer and field-level spatial heterogeneity in hyperspectral data through a spatial convolutional layer. Finally, a gated fusion unit generates a degradation risk index. The network training data can include historical monitoring data and known cases of cultivated land degradation to improve the accuracy and reliability of predictions.

[0107] Preferably, to improve the accuracy of the spatiotemporal correlation analysis submodule, more soil parameters and crop growth indicators, such as soil moisture, soil temperature, and crop chlorophyll content, can be introduced to more comprehensively reflect the interaction between soil and crops. At the same time, the calculation of the coupling relationship can adopt more complex statistical models, such as partial least squares regression (PLSR), to consider the potential correlation between multiple variables. For the degradation rate prediction submodule, the structure of the multi-scale convolutional gated network can be optimized according to the characteristics of cultivated land in different regions. For example, in areas with severe heavy metal pollution, sensitivity analysis of changes in heavy metal concentrations can be added; in saline-alkali areas, the time series changes in soil salt concentration can be analyzed in detail. In addition, the network training process can introduce transfer learning technology, using existing similar datasets for pre-training and then fine-tuning on local data to improve the adaptability and generalization ability of the model.

[0108] In some embodiments, the calculation steps of the multi-scale convolutional gating network include:

[0109] (a) Input the soil conductivity time series data in the spatiotemporal database module into the temporal convolution layer to extract the interannual variation feature vector H temp ;

[0110] (b) The UAV hyperspectral data in the spatiotemporal database module is input into the spatial convolution layer to extract the field-level spatial heterogeneity feature vector H spat ;

[0111] (c) Generate degradation risk index through gated fusion unit:

[0112] G t =σ(W g ·[H temp ,H spat ])

[0113] Among them, G t is the time gating weight vector, W g is the trainable parameter matrix, and σ is the sigmoid function.

[0114] It should be noted that the multi-scale convolutional gated network is the core technology of the degradation rate prediction submodule. Its function is to process multi-dimensional farmland monitoring data and output the rate of change of farmland quality over a period of time. This network uses temporal and spatial convolutional layers to extract the interannual variation characteristics of soil parameters and the field-level spatial heterogeneity characteristics of hyperspectral data, respectively. A gated fusion unit is then used to generate a degradation risk index. The temporal convolutional layer captures the temporal variation of soil parameters, such as seasonal fluctuations in soil electrical conductivity, while the spatial convolutional layer analyzes the differences in soil and crop characteristics across different regions within a field. The gated fusion unit fuses the extracted features using a weight vector and a trainable parameter matrix. Finally, a sigmoid function is used to output a degradation risk index, which is used to assess the risk of farmland quality degradation.

[0115] Specifically, the input to the temporal convolution layer is the soil electrical conductivity time series data from the spatiotemporal database module, which reflects the temporal evolution of soil salt concentration. Soil electrical conductivity is an important indicator of soil salinization, and its time series data can be regularly collected by an in-situ sensor array. The temporal convolution layer extracts interannual variation feature vectors through convolution operations. The convolution kernel size can be set based on the sampling frequency and temporal resolution of the soil conductivity data. For example, if the data is collected on a monthly basis, the convolution kernel size can be set to 12 to capture annual cyclical variations. The input to the spatial convolution layer is drone hyperspectral data, which reflects the crop growth status and spectral characteristics of the soil within a field. The spatial convolution layer extracts field-level spatial heterogeneity feature vectors through convolution operations. The convolution kernel size can be adjusted based on the actual field size and resolution. The temporal gating weight vector and trainable parameter matrix in the gated fusion unit are network parameters, which are optimized during training to ensure that the model accurately integrates temporal and spatial features to generate a reliable degradation risk index.

[0116] Preferably, in order to improve the prediction accuracy of the multi-scale convolutional gating network, the network structure can be optimized. For example, a multi-scale convolution kernel is introduced in the temporal convolution layer to simultaneously capture short-term and long-term temporal variation characteristics. For the spatial convolution layer, a multi-scale spatial pyramid pooling technique can be used to better handle the spatial heterogeneity of different scales within the field. In addition, the sigmoid function in the gated fusion unit can be replaced with other activation functions, such as ReLU or Leaky ReLU, to improve the nonlinear fitting ability of the network. During the training process, regularization techniques such as Dropout or L2 regularization can be introduced to prevent the model from overfitting. At the same time, in order to improve the generalization ability of the model, a transfer learning method can be used to apply the pre-trained model to farmland monitoring data from different regions and fine-tune it according to local data.

[0117] In some embodiments, the dynamic warning module performs:

[0118] When the heavy metal concentration in the predicted value of farmland quality degradation risk exceeds 80% of the screening value of the "Agricultural Land Soil Pollution Risk Control Standard", a first-level warning signal is generated;

[0119] When the predicted value of soil organic matter content decreases at a rate of more than 5% / month for three consecutive months and the crop NDVI decreases at a rate of more than 15%, a second-level warning signal will be generated;

[0120] The early warning signal triggers the cross-region compensation calculation of the repair decision module.

[0121] It should be noted that the dynamic early warning module is a key component of this system, generating multi-level early warning signals based on the predicted farmland quality degradation risk value and preset farmland quality thresholds. The predicted farmland quality degradation risk value, calculated by the quality degradation prediction module, is used to assess the potential degree of farmland quality degradation over a period of time. The preset farmland quality threshold is set based on national or local farmland quality standards and specific management requirements to determine whether farmland quality is within a safe range. When the predicted farmland quality degradation risk value exceeds the set threshold, the dynamic early warning module triggers a corresponding early warning signal, enabling timely action. For example, a level 1 early warning signal is generated when heavy metal concentrations approach or exceed 80% of the screening value in the "Agricultural Land Soil Pollution Risk Management Standards." A level 2 early warning signal is generated when soil organic matter content continuously decreases and crop growth indicators (such as NDVI) significantly degrade. The triggering of early warning signals not only alerts managers to potential farmland quality issues but also triggers responses from the remediation decision module, providing a basis for subsequent remediation measures.

[0122] Specifically, the dynamic early warning module's warning signal generation mechanism is based on key indicators of arable land quality. For example, the "Standards for the Risk Control of Soil Pollution in Agricultural Land" sets clear screening thresholds for heavy metal concentrations. When the predicted monitored heavy metal concentration exceeds 80% of this screening threshold, the system issues a Level 1 warning, indicating that the arable land may be at serious risk of heavy metal contamination. The system monitors trends in soil organic matter content and crop growth indicators (such as NDVI). If the soil organic matter content decreases by more than 5% per month for three consecutive months, and the crop NDVI decreases by more than 15%, a Level 2 warning is triggered, indicating that the fertility of the arable land and crop growth may be seriously affected. Warning signal levels can be graded based on the severity of the arable land degradation. For example, a Level 1 warning indicates a high risk, requiring immediate action; a Level 2 warning indicates a moderate risk, requiring enhanced monitoring and the development of a remediation plan. Warning signal generation relies not only on real-time monitoring data but also incorporates historical data and trend analysis to ensure the accuracy and timeliness of warnings.

[0123] Preferably, the dynamic early warning module can flexibly adjust the early warning threshold according to the characteristics of cultivated land and management needs in different regions. For example, in some ecologically fragile areas, the early warning threshold of heavy metal concentration can be set lower to detect potential pollution risks earlier. For the early warning conditions of soil organic matter content and crop growth indicators, more dynamic factors can be introduced, such as seasonal changes and regional differences. For example, in arid areas, the decline threshold of crop NDVI can be appropriately adjusted to take into account the impact of water stress on crop growth. In addition, the dynamic early warning module can also combine geographic information system (GIS) technology to associate early warning signals with specific geographical locations and generate visual maps to help managers understand the distribution of cultivated land quality and risk areas more intuitively. The system can also set an automatic notification function to send early warning information to relevant managers via SMS, email or mobile application to ensure timely response.

[0124] In some embodiments, the quality traceability blockchain module includes:

[0125] The farmland quality fingerprint unit generates verifiable credentials including soil parameter hash values, remote sensing image feature values, and early warning records;

[0126] Smart contract unit, implements the following farmland quality constraint rules:

[0127] When the quality grade of newly added cultivated land is lower than that of occupied cultivated land, the land change approval process will be frozen.

[0128] When the progress of the soil remediation project does not reach 90% of the planned value, the ecological compensation deposit will be automatically allocated;

[0129] Distributed storage unit, synchronously stores the original data of quality assessment at the provincial farmland protection node.

[0130] It should be noted that the quality traceability blockchain module is a crucial component of this system for achieving verifiable evidence storage of cultivated land monitoring data. This module ensures the authenticity, integrity, and immutability of monitoring data through cultivated land quality fingerprint units, smart contract units, and distributed storage units. The cultivated land quality fingerprint unit generates verifiable credentials including soil parameter hash values, remote sensing image feature values, and early warning records, providing a unique digital identifier for cultivated land quality. The smart contract unit automatically executes relevant operations based on preset cultivated land quality constraint rules, such as freezing the land change approval process or automatically allocating ecological compensation deposits. The distributed storage unit synchronously stores raw quality assessment data at provincial cultivated land protection nodes to ensure data transparency and traceability. Through the application of blockchain technology, the system can provide reliable technical support for the long-term supervision of cultivated land quality and enhance public trust in cultivated land quality information.

[0131] Specifically, the arable land quality fingerprint unit compresses continuous monitoring data into a feature matrix and combines it with the vegetation anomaly index from drone multispectral data to generate a unique fingerprint. The feature matrix efficiently compresses monitoring data, preserving key information and facilitating subsequent comparison and verification. The vegetation anomaly index is calculated by calculating the difference between the current Normalized Difference Vegetation Index (NDVI) and the historical baseline NDVI, reflecting changes in crop growth status. A hash algorithm (such as SHA256) is used to encrypt the feature matrix, vegetation anomaly index, and blockchain timestamp to generate an unalterable unique fingerprint. The smart contract unit automatically executes operations according to pre-set rules. For example, if the quality grade of newly added arable land is lower than that of occupied land, the land change approval process is automatically frozen to ensure that arable land quality is not degraded. If the progress of a soil remediation project falls below 90% of the planned value, an ecological compensation deposit is automatically allocated to ensure the smooth progress of remediation work. The distributed storage unit synchronously stores raw data at provincial nodes, ensuring extensive data backup and traceability.

[0132] Preferably, the cultivated land quality fingerprint unit can further optimize the data compression and feature extraction process. For example, a more advanced dimensionality reduction algorithm (such as principal component analysis PCA) is introduced to pre-process the monitoring data to improve the quality and representativeness of the feature matrix. In the calculation of the vegetation anomaly index, multiple vegetation indices (such as EVI, SAVI, etc.) can be combined to more comprehensively reflect the crop growth status. For the smart contract unit, a more complex rule engine can be introduced to support dynamic adjustment of constraint rules to adapt to the cultivated land management policies in different regions. For example, the judgment criteria for soil quality grades can be adjusted according to the soil types and climatic conditions in different regions. The distributed storage unit can adopt the alliance chain architecture of the blockchain, allowing cultivated land protection departments in different regions to participate in data storage and verification as nodes, thereby improving the system's decentralization and data security.

[0133] In some embodiments, the method for generating a farmland quality fingerprint unit includes:

[0134] The farmland monitoring data of the in-situ sensor array unit for 30 consecutive days is compressed into a feature matrix M sensor , where M sensor It is an m×n dimensional matrix, where m represents the number of monitoring days and n represents the number of sensor types;

[0135] Calculate the vegetation anomaly index from UAV multispectral data:

[0136]

[0137] Among them, NDVI current is the current normalized vegetation index, NDVI baseline is the normalized difference vegetation index of the historical base period;

[0138] Generate a unique fingerprint:

[0139] F print =SHA256(M sensor ||V anomaly ||t block )

[0140] Among them, || represents the data splicing operation, t block It is the blockchain timestamp and SHA256 is the secure hash algorithm.

[0141] It should be noted that the arable land quality fingerprint unit is a key component of the quality traceability blockchain module. Its function is to generate verifiable credentials including soil parameter hash values, remote sensing image feature values, and early warning records, providing a unique digital identifier for arable land quality. This process compresses the arable land data continuously monitored by the in-situ sensor array unit into a feature matrix and combines it with the vegetation anomaly index of the drone's multispectral data to generate a unique fingerprint identifier. The feature matrix is an efficient compression of the monitoring data, which can retain key information and facilitate subsequent comparison and verification. The vegetation anomaly index is calculated by calculating the difference between the current normalized difference vegetation index (NDVI) and the historical baseline NDVI, and can reflect changes in crop growth status. This data is encrypted using the SHA256 hash algorithm to generate a unique, tamper-proof fingerprint identifier, ensuring the authenticity and integrity of the arable land quality data.

[0142] Specifically, the implementation of the farmland quality fingerprint unit involves multiple steps. First, 30 consecutive days of farmland monitoring data from the in-situ sensor array unit is compressed into a feature matrix, an m×n dimensional matrix, where m represents the number of monitoring days and n represents the number of sensor types. This process can be implemented using data dimensionality reduction algorithms, such as principal component analysis (PCA), to reduce data redundancy and extract key features. The vegetation anomaly index is calculated as follows: Vegetation anomaly index = (current NDVI - historical baseline NDVI) / historical baseline NDVI × 100%, where the current NDVI is calculated using drone multispectral data, and the historical baseline NDVI is a normalized vegetation index based on past healthy crop growth conditions. The SHA256 hash algorithm is used to encrypt the feature matrix, vegetation anomaly index, and blockchain timestamp to generate a unique fingerprint identifier. This identifier not only includes key information on farmland quality but also ensures the chronological order and immutability of the data through the blockchain timestamp.

[0143] Preferably, the cultivated land quality fingerprint unit can further optimize the data compression and feature extraction process. For example, a more advanced dimensionality reduction algorithm (such as t-SNE or autoencoder) is introduced to preprocess the monitoring data to improve the quality and representativeness of the feature matrix. In the calculation of the vegetation anomaly index, multiple vegetation indices (such as EVI, SAVI, etc.) can be combined to more comprehensively reflect the crop growth status. In addition, a dynamic time window can be introduced to adjust the time span of the monitoring data according to the crop growth cycle and seasonal changes. For example, the monitoring cycle can be shortened to 15 days during the peak crop growth period to reflect the changes in cultivated land quality more promptly. For hash algorithms, in addition to SHA256, other encryption algorithms (such as SHA3 or BLAKE2) can also be considered to improve security. At the same time, to ensure the traceability of the data, the operation log of each fingerprint identification generated can be recorded in the blockchain, including information such as operation time, operator and data source.

[0144] In some embodiments, the repair decision module includes:

[0145] The salt-alkali treatment submodule generates parameters for the underground pipe salt drainage project based on the predicted value of the farmland quality degradation risk:

[0146]

[0147] Where, L is the distance between concealed pipes (meters), K is the soil permeability coefficient (m / d), t is the design drainage cycle (days), H0 is the initial groundwater level (meters), H t is the target groundwater level (m), ΔS is the salt concentration gradient correction factor;

[0148] The heavy metal remediation submodule calculates the ratio between the amount of biochar applied and the chelating agent spraying concentration based on the predicted value of the farmland quality degradation risk.

[0149] It should be noted that the remediation decision-making module is a key component of this system, used to generate technical parameters for soil improvement projects and ecological compensation plans based on the predicted risk of cultivated land quality degradation. This module includes a salinization and alkali treatment submodule and a heavy metal remediation submodule, each of which provides targeted remediation strategies for different types of cultivated land quality issues. The salinization and alkali treatment submodule provides technical support for saline-alkali land remediation by calculating the parameters of the concealed pipe salt drainage project; the heavy metal remediation submodule provides a scientific basis for the remediation of heavy metal-contaminated soil by calculating the ratio between the amount of biochar applied and the chelating agent spray concentration. The design of these two submodules is based on the specific characteristics of soil quality degradation and can provide precise remediation plans for the restoration of cultivated land quality.

[0150] Specifically, the calculation formula for the parameters of the hidden pipe salt drainage engineering of the salt-alkali treatment submodule is:

[0151] L=4πKt / ln(H O / H t )×1 / (1+0.1ΔC)

[0152] Where, L is the distance between concealed pipes (in meters), K is the soil permeability coefficient (in m / d), t is the design drainage period (in days), H O is the initial groundwater level (in meters), H t is the target groundwater level (in meters), ΔC is the salt concentration gradient correction factor. The soil permeability coefficient K can be determined by soil texture analysis and field tests. The designed drainage period t is set according to the local precipitation pattern and irrigation demand. The initial groundwater level H O and target groundwater level H tThe amount of biochar applied in the heavy metal remediation submodule is calculated based on groundwater level monitoring data and remediation objectives. The equilibrium adsorption amount is calculated using the Freundlich adsorption isotherm, and the minimum application rate is calculated based on the target remediation concentration. These parameters need to be adjusted based on the specific soil type, contamination level, and remediation objectives.

[0153] Preferably, the parameters of the concealed pipe salt drainage project of the salinization and alkali treatment submodule can be optimized according to the soil texture and climatic conditions of different regions. For example, in areas with poor soil permeability, the concealed pipe spacing L can be appropriately reduced to improve drainage efficiency; in arid areas, the designed drainage period t can be appropriately extended to reduce water waste. For the heavy metal remediation submodule, the amount of biochar applied can be adjusted according to the organic matter content of the soil and the degree of heavy metal pollution. For example, in soils with low organic matter content, the amount of biochar applied can be increased to improve the soil adsorption capacity. In addition, the ratio of the chelating agent spray concentration can be optimized according to the type of heavy metal and the pollution concentration. For example, for soils with severe cadmium pollution, the chelating agent concentration can be appropriately increased to improve the remediation efficiency. At the same time, the remediation decision module can also combine geographic information system (GIS) technology to generate a visual layout map of the remediation project, providing more intuitive guidance for actual operations.

[0154] In some embodiments, the biochar application amount calculation includes:

[0155] Establishment of heavy metal adsorption kinetic model:

[0156]

[0157] Among them, Q e is the equilibrium adsorption capacity (mg / g), C e is the equilibrium concentration of heavy metals (mg / L), k f is the Freundlich adsorption coefficient, n is the nonlinear exponent;

[0158] According to the target repair concentration C target Reverse the minimum application rate:

[0159]

[0160] Among them, V soil is the volume of treated soil (m 3 ), ρ is soil bulk density (g / cm 3 ), C0 is the initial pollution concentration (mg / kg).

[0161] It should be noted that this section involves the calculation method for the amount of biochar applied in the heavy metal remediation submodule, which aims to provide a scientific basis for the remediation of heavy metal pollution in soil. Biochar is a carbon-rich porous material with good heavy metal adsorption properties, which can effectively reduce the bioavailability of heavy metals in the soil. This calculation method is based on the heavy metal adsorption kinetic model. It uses the Freundlich adsorption isotherm to describe the adsorption behavior of heavy metals on the biochar surface, and reversely infers the minimum application amount of biochar based on the target remediation concentration. This method can provide quantitative technical support for soil remediation projects and ensure the scientific nature and effectiveness of the remediation process.

[0162] Specifically, the calculation process of biochar application amount is as follows: First, a heavy metal adsorption kinetic model is established, where the Freundlich adsorption isotherm is: where q e represents the equilibrium adsorption capacity (in mg / g), C e Indicates the equilibrium concentration of heavy metals (in mg / L), K n and n are the Freundlich adsorption coefficient and nonlinear index, respectively. These two parameters can be obtained through experimental determination. Usually, adsorption isotherm experiments need to be conducted under laboratory conditions to determine the adsorption characteristics of biochar for specific heavy metals. According to the target remediation concentration C tmt , the minimum application amount of biochar can be inferred, and the calculation formula is:

[0163] q=(V×ρ×(C O -C tmt )) / K n

[0164] Where V represents the volume of treated soil (unit: m 3 ), ρ represents the soil bulk density (unit: g / cm 3 ), C O is the initial pollution concentration (in mg / kg). The settings of these parameters need to be adjusted according to the specific soil type, heavy metal type and pollution degree to ensure that the remediation effect achieves the expected goal.

[0165] Preferably, in order to improve the accuracy and applicability of the calculation of biochar application amount, the following optimization schemes can be considered: First, for different types of biochar (such as rice husk charcoal, charcoal, etc.), there may be differences in their adsorption properties, so when calculating the application amount, it is necessary to adjust it according to the specific source and characteristics of the biochar. For example, the adsorption capacity of rice husk charcoal may be relatively low, and the application amount needs to be appropriately increased. Secondly, a multi-component adsorption model can be introduced to consider the synergistic adsorption effect of multiple heavy metals in the soil. For example, when cadmium and lead are present in the soil at the same time, their adsorption behavior may be different from that of a single heavy metal, and the adsorption coefficient and nonlinear index need to be adjusted based on experimental data. In addition, soil microbial remediation technology can be combined to improve the removal efficiency of heavy metals by adding microbial agents to work synergistically with biochar, thereby reducing the application amount of biochar.

[0166] The aforementioned embodiments of the present invention have the following beneficial effects: The present invention integrates multiple sensors and remote sensing technologies through a multi-source perception module, enabling real-time collection of multi-dimensional data on cultivated land quality, covering key parameters such as soil pH, organic matter content, heavy metal concentrations, and crop growth indicators, thereby comprehensively and dynamically reflecting changes in cultivated land quality. The heterogeneous transmission module encrypts and encapsulates the collected data and standardizes its format, ensuring data security and consistency during transmission and providing a reliable foundation for subsequent processing. The spatiotemporal database module stores processed data and its spatiotemporal correlation information, providing powerful support for spatiotemporal analysis of cultivated land quality. The quality degradation prediction module, combining spatiotemporal correlation analysis with a multi-scale convolutional gated network, accurately predicts cultivated land quality degradation risks, provides early warning of future trends, and provides a basis for scientific decision-making. The dynamic early warning module generates multi-level early warning signals based on the predicted values, which can promptly trigger responses from the restoration decision module, enabling precise management of cultivated land quality and ecological compensation. Based on the degradation risk prediction values, the restoration decision module outputs soil improvement engineering technical parameters and ecological compensation plans, providing scientific guidance for precise restoration of cultivated land quality and enhancing the scientific nature and effectiveness of cultivated land protection.

[0167] Furthermore, the quality traceability blockchain module ensures the authenticity, integrity, and immutability of cultivated land quality data by generating farmland quality fingerprints and smart contract units, thereby enhancing public trust in cultivated land quality information. The cultivated land quality fingerprint unit combines continuous monitoring data and the vegetation anomaly index to generate a unique fingerprint identifier, providing accurate digital evidence of cultivated land quality. The smart contract unit enforces cultivated land quality constraints, automatically freezing the land change approval process or allocating ecological compensation deposits to ensure the effective implementation of cultivated land quality protection measures. Distributed storage units synchronously store raw quality assessment data at provincial cultivated land protection nodes, further enhancing data security and transparency, and providing strong support for the long-term supervision and sustainable use of cultivated land quality.

[0168] Furthermore, the storage medium of the embodiment of the present application stores program instructions that can implement all the above methods, wherein the program instructions can be stored in the above storage medium in the form of a software product, including a number of instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or a terminal device such as a computer, a server, a mobile phone, or a tablet.

[0169] The above descriptions are merely some preferred embodiments of the present invention and an illustration of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present invention is not limited to the technical solutions formed by a specific combination of the above-mentioned technical features, but should also encompass other technical solutions formed by any combination of the above-mentioned technical features or their equivalents without departing from the above-mentioned inventive concept. For example, a technical solution formed by mutually replacing the above-mentioned features with (but not limited to) technical features having similar functions disclosed in the embodiments of the present invention.

Claims

1. A farmland quality monitoring system based on farmland protection, characterized in that: include: A multi-source sensing module is used to collect farmland monitoring data, including soil pH, organic matter content, heavy metal concentration, crop canopy temperature, and vegetation index data; A heterogeneous transmission module, used for encrypting, encapsulating and formatting the farmland monitoring data output by the multi-source perception module; A spatiotemporal database module, configured to store the farmland monitoring data processed by the heterogeneous transmission module and its spatiotemporal correlation information, wherein the spatiotemporal correlation information includes Beidou coordinates of monitoring points, data acquisition timestamps, and multi-source data mapping relationships; a quality degradation prediction module, configured to fuse the processed farmland monitoring data and its spatiotemporal correlation information to generate multi-dimensional farmland monitoring data, and to generate a farmland quality degradation risk prediction value based on the multi-dimensional farmland monitoring data; Dynamic early warning module, used to generate multi-level early warning signals based on the predicted value of cultivated land quality degradation risk and preset cultivated land quality thresholds; Restoration decision-making module, used to output soil improvement engineering technical parameters and ecological compensation plans based on the predicted value of cultivated land quality degradation risk; The quality traceability blockchain module is used to achieve verifiable storage of cultivated land monitoring data.

2. The system according to claim 1, wherein: The multi-source perception module includes: In-situ sensor array units, deployed at farmland monitoring points, include: Ring electrode conductivity sensor, used to measure soil salt concentration; Laser-induced breakdown spectroscopy sensor for detecting the content of arsenic, cadmium, lead, mercury, and chromium in soil; Near-infrared spectral reflectance probe, used to continuously monitor the dynamic changes of soil organic matter; UAV remote sensing unit equipped with a hyperspectral imager to obtain crop stress characteristic data; Satellite analysis unit, used to receive and process NDVI index, surface temperature and soil moisture inversion data.

3. The system according to claim 2, characterized in that The heterogeneous transmission module performs: The cultivated land monitoring data output by the in-situ sensor array unit is subjected to wavelet threshold denoising processing, and the calculation formula is: in, is the denoised signal, ψ j,k is the Daubechies wavelet basis function, j is the number of decomposition levels, k is the translation coefficient, T is the adaptive threshold, σ is the noise standard deviation, N is the signal length, and I(·) is the indicator function; The cultivated land monitoring data output by the UAV remote sensing unit is geo-referenced to generate an encrypted data packet with Beidou coordinates.

4. The system according to claim 1, wherein: The quality degradation prediction module includes: The spatiotemporal correlation analysis submodule is used to calculate the coupling relationship between soil parameters and crop growth indicators: Among them, R sc is the soil-crop correlation, S i is the i-th soil parameter, is the mean value of soil parameters, C i is the corresponding crop growth index, is the mean value of crop indicators; The degradation rate prediction submodule uses a multi-scale convolutional gated network to process the multi-dimensional cultivated land monitoring data and output the cultivated land quality change rate in the next 6 months.

5. The system according to claim 4, characterized in that The calculation steps of the multi-scale convolutional gated network include: (a) Input the soil conductivity time series data in the spatiotemporal database module into the temporal convolution layer to extract the interannual variation feature vector H temp ; (b) The UAV hyperspectral data in the spatiotemporal database module is input into the spatial convolution layer to extract the field-level spatial heterogeneity feature vector H spat ; (c) Generate degradation risk index through gated fusion unit: G t =σ(W g ·[H temp ,H spat ]) Among them, G t is the time gating weight vector, W g is the trainable parameter matrix, and σ is the sigmoid function.

6. The system according to claim 1, wherein: The dynamic warning module performs the following steps: When the heavy metal concentration in the predicted value of farmland quality degradation risk exceeds a preset screening value, a first-level warning signal is generated; When the predicted value of soil organic matter content decreases by more than 5% for three consecutive months and the crop NDVI decreases by more than 15%, a second-level warning signal is generated; The early warning signal triggers the cross-region compensation calculation of the repair decision module.

7. The system according to claim 1, wherein: The quality traceability blockchain module includes: The farmland quality fingerprint unit is used to generate verifiable credentials including soil parameter hash values, remote sensing image feature values, and early warning records; Smart contract unit, used to implement the following farmland quality constraint rules: When the quality grade of newly added arable land is lower than that of occupied arable land, the land change approval process will be frozen; When the progress of the soil remediation project does not reach 90% of the planned value, the ecological compensation deposit will be automatically allocated; Distributed storage unit, synchronously stores the original data of quality assessment at the provincial farmland protection node.

8. The system according to claim 7, characterized in that The method for generating a farmland quality fingerprint unit comprises: The farmland monitoring data of the in-situ sensor array unit for 30 consecutive days is compressed into a feature matrix M sensor , where M sensor It is an m×n dimensional matrix, where m represents the number of monitoring days and n represents the number of sensor types; Calculate the vegetation anomaly index from UAV multispectral data: Among them, NDVI current is the current normalized vegetation index, NDVI baseline is the normalized vegetation index of the historical base period, V anomaly is the vegetation anomaly index; Generate a unique fingerprint: F print =SHA256(M sensor ||V anomaly ||t block ) Among them, || represents the data splicing operation, t block It is the blockchain timestamp and SHA256 is the secure hash algorithm.

9. The system according to claim 1, wherein: The repair decision module includes: The salt-alkali treatment submodule is used to generate parameters for the underground pipe salt drainage project based on the predicted value of the farmland quality degradation risk: Where L is the distance between concealed pipes, K is the soil permeability coefficient, t is the designed drainage period, H0 is the initial groundwater level, H t is the target groundwater level, ΔS is the salt concentration gradient correction factor; The heavy metal remediation submodule is used to calculate the ratio between the amount of biochar applied and the chelating agent spraying concentration based on the predicted value of the farmland quality degradation risk.

10. The system according to claim 9, characterized in that The biochar application amount calculation includes: Establishment of heavy metal adsorption kinetic model: Among them, Q e is the equilibrium adsorption capacity, C e is the equilibrium concentration of heavy metals, k f is the Freundlich adsorption coefficient, n is the nonlinear exponent; According to the target repair concentration C target Reverse the minimum application rate: Among them, V soil is the volume of treated soil, ρ is the soil bulk density, and C0 is the initial pollution concentration.

Images