A multi-mode cooperative driving adaptive remote sensing drought monitoring method and system

Abstract

The application provides a kind of multi-mode cooperative driving's adaptive remote sensing drought monitoring method and system, according to the important characteristic index of drought, set up a variety of drought mode, select a space division standard for each mode, combine standard cross division, determine several monitoring units;Select representative drought monitoring index as candidate factor for each mode;Calculate the correlation of various factors in monitoring unit, select the top factor for each mode;According to the drought grade corresponding table of factor, produce single mode drought grade sequence for each mode;Production measured drought grade sequence, construct the adaptive evaluation model of each mode drought grade sequence and measured drought grade sequence, realize the adaptive drought grade evaluation of monitoring unit under multi-mode cooperative driving.The application considers geographical features to construct drought monitoring unit, and uses factor similarity index to realize drought factor screening, breaks through the limitation of single mode drought monitoring, and solves the problem of integrated monitoring of multi-scale and multi-type drought.

Description

Technical Field

[0001] This invention belongs to the field of drought monitoring technology. Specifically, it proposes a multi-mode collaborative driven adaptive remote sensing drought monitoring method and system. Background Technology

[0002] Drought is one of the major natural disasters worldwide and the most significant natural disaster leading to reduced food production.

[0003] Drought, with its wide reach, long duration, and numerous influencing factors, has always been a global challenge for monitoring and early warning. The difficulty in drought monitoring lies in the randomness of drought as a natural phenomenon and the complexity of it as a natural-social symbiotic event. Long-term, comprehensive, and multi-method-integrated monitoring and analysis are necessary to understand its occurrence and development patterns, enabling early warning, and the formulation of countermeasures to prevent severe drought disasters and reduce losses. Currently, drought monitoring mainly relies on ground stations or remote sensing, but single-mode monitoring cannot simultaneously guarantee accuracy and efficiency. The operational application of remote sensing drought monitoring currently faces several problems. On the one hand, while traditional methods are comprehensive and remote sensing drought monitoring methods are numerous, each type of remote sensing drought index has its applicable conditions, failing to meet the requirements of long-term, large-scale drought monitoring. On the other hand, current methods primarily characterize and infer drought conditions and their development by acquiring soil moisture, precipitation, or runoff data. However, the mechanisms of drought formation are extremely complex, and these methods cannot comprehensively, accurately, and timely reflect the drought situation, falling short of the needs of my country's water conservancy industry in combating complex drought disasters.

[0004] This patent focuses on tackling key technologies for adaptive remote sensing drought monitoring driven by multi-mode synergy. Important drought characteristics, such as soil moisture, water content changes, and surface temperature, are defined as drought monitoring modes. Depending on different spatiotemporal environments and data availability, one or more modes are selected for synergy to explore drought remote sensing inversion based on multi-mode synergy. By coordinating multiple modes, including water body characteristic change monitoring, soil moisture content inversion, and surface temperature inversion, drought monitoring and assessment can be achieved. This not only enriches and innovates drought monitoring methods and overcomes the limitations of single-mode monitoring, but also improves the reliability, timeliness, and forward-looking nature of monitoring, providing an important theoretical and methodological foundation for reliable monitoring of long-term, large-scale droughts. It has significant research and application value in drought relief and disaster reduction, and scientific management and allocation of water resources.

[0005] By integrating multiple drought models such as soil moisture, temperature, and water body characteristics, the optimal single-model drought monitoring factors for each monitoring unit are selected. The multi-model collaborative driving mechanism enables adaptive remote sensing drought level assessment of the monitoring unit and monitors the entire process of drought occurrence and development. This is a novel multi-model collaborative driving monitoring method. Summary of the Invention

[0006] This invention combines the spatiotemporal adaptability of different remote sensing drought indices under multiple modes to select the optimal drought factor, and proposes an adaptive drought level assessment method. It realizes adaptive assessment of remote sensing drought driven by multi-mode collaboration, and solves the problem that a single drought index is difficult to reflect the multi-scale and multi-type drought situation. It provides a complete evaluation system and reliable technical solution for scientific and systematic drought monitoring.

[0007] To address the aforementioned technical problems, the present invention provides a multi-mode collaboratively driven adaptive remote sensing drought monitoring method, comprising the following steps:

[0008] Step 1: Set up multiple drought conditions based on key drought characteristics indicators, select a spatial division standard for each model, and determine several monitoring units by combining the cross-division of the standards.

[0009] Step 2: Select representative drought monitoring indices as candidate factors for each model;

[0010] Step 3: Calculate the correlation of various factors within the monitoring unit, and select the top-ranked factors for each pattern based on the factor similarity index.

[0011] Step 4: Based on the drought level correspondence table of factors, generate the drought level sequence for each production model.

[0012] Step 5: Generate the measured drought level sequence, construct an adaptive evaluation model of the drought level sequence of each model and the measured drought level sequence, and realize the adaptive drought level evaluation of the monitoring unit under the collaborative drive of multiple models.

[0013] Furthermore, the system sets up multiple drought conditions, including three modes: soil moisture, surface temperature, and water body characteristics.

[0014] Furthermore, step 1 is implemented by including the following sub-steps:

[0015] Step 1.1: Select any spatial partitioning standard to perform primary partitioning of the space;

[0016] Step 1.2: Select any of the remaining drought monitoring models, and further divide the space based on the primary zoning and the zoning criteria of the model to obtain the secondary zoning;

[0017] Step 1.3: Traverse other patterns and repeat Step 1.2 until all spatial division criteria of all patterns have been traversed, and the final spatial division unit is obtained, which serves as the monitoring unit for adaptive drought monitoring.

[0018] Furthermore, representative drought monitoring indices include, but are not limited to, meteorological drought index, hydrological drought index, and agricultural drought index; at least one drought index is selected as a candidate factor for each type of cold weather model.

[0019] Furthermore, step 3 is implemented by including the following sub-steps:

[0020] Step 3.1: Calculate the Pearson correlation coefficient for each factor.

[0021]

[0022] In the formula, cov is the covariance, σ is the standard deviation, and D and D' are the drought index sequences corresponding to the two candidate factors, respectively. i and D′ i This represents the i-th time-series value of the drought index, where n is the number of time-series values. and σ represents the average value of the drought index series. D Let σ be the standard deviation of the drought index series D. D′ The standard deviation of the drought index series D';

[0023] Step 3.2: Calculate the maximum mutual information coefficient (MIC) for each factor;

[0024] Step 3.3: Construct a factor similarity index S by calculating the mean of the two, S = (R + MIC) / 2. Based on the ascending order of the factor similarity index, select the top-ranked drought index for each pattern.

[0025] Furthermore, step 5 is implemented by including the following sub-steps:

[0026] Step 5.1: Collect measured or actual disaster data and calculate the actual drought index;

[0027] Step 5.2: Generate a sequence of measured drought levels based on the drought level comparison table;

[0028] Step 5.3: Establish an adaptive evaluation model for the measured drought level sequence of the monitoring unit and the drought level sequence of the single model;

[0029] Step 5.4: Based on the adaptive assessment method, produce the final drought level product of the assessment unit to realize adaptive drought monitoring driven by multi-mode collaboration.

[0030] On the other hand, the present invention also provides a multi-mode collaborative driven adaptive remote sensing drought monitoring system for implementing the multi-mode collaborative driven adaptive remote sensing drought monitoring method described above.

[0031] Moreover, it includes the following modules,

[0032] The first module is used to set up multiple drought conditions based on important drought characteristic indicators, select a spatial division standard for each model, and determine several monitoring units by combining the cross division of the standards.

[0033] The second module is used to select representative drought monitoring indices as candidate factors for each model;

[0034] The third module is used to calculate the correlation of various factors within the monitoring unit and to select the top-ranked factors for each pattern based on the factor similarity index.

[0035] The fourth module is used to generate single-mode drought level sequences for each model based on the drought level correspondence table of factors;

[0036] The fifth module is used to generate measured drought level sequences, construct an adaptive evaluation model of drought level sequences of each model and measured drought level sequences, and realize adaptive drought level evaluation of monitoring units under multi-mode collaborative driving.

[0037] Alternatively, it may include a processor and a memory, with the memory used to store program instructions and the processor used to call the stored instructions in the memory to execute a multi-mode collaboratively driven adaptive remote sensing drought monitoring method as described above.

[0038] Alternatively, it may include a readable storage medium storing a computer program that, when executed, implements a multi-mode collaboratively driven adaptive remote sensing drought monitoring method as described above.

[0039] Compared with the prior art, the present invention has the following features and beneficial effects:

[0040] 1. To address the uncertainty in reflecting objective drought conditions due to the spatiotemporal heterogeneity of drought factors, a drought monitoring unit is constructed by spatially partitioning the underlying surface based on different geographical environments. The factor similarity index is used to achieve rapid and accurate extraction and screening of drought factor characteristics, which reduces the impact of spatiotemporal differences in large-scale long-term drought monitoring and improves the spatiotemporal adaptability and sensitivity of drought characterization.

[0041] 2. To overcome the challenge of adaptive monitoring of drought conditions at multiple scales and of multiple types, a multi-mode collaborative monitoring mechanism was proposed, focusing on soil moisture, water body characteristics and surface temperature. Based on underlying surface conditions such as land cover type and the long-term drought evolution pattern in the region, adaptive remote sensing monitoring of drought conditions in different regions was achieved, breaking through the limitations of single-mode drought monitoring.

[0042] 3. Combining measured data and integrating drought level assessment information from multiple models, a remote sensing drought level classification method based on BP neural network was proposed, and a comprehensive level assessment standard was established for adaptive drought monitoring. Attached Figure Description

[0043] Figure 1 This is a flowchart of the adaptive drought monitoring assessment method proposed in this embodiment of the invention;

[0044] Figure 2 This is an example diagram of a drought level sequence according to an embodiment of the present invention;

[0045] Figure 3 This is a schematic diagram of the adaptive evaluation of the BP neural network in an embodiment of the present invention. Detailed Implementation

[0046] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0047] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0048] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but this is not intended to limit the present invention.

[0049] This invention constructs drought monitoring units by considering geographical features, uses factor similarity indices to screen drought factors, and builds a multi-mode collaborative adaptive drought assessment model. This overcomes the limitations of single-mode drought monitoring, solves the problem of integrated monitoring of multi-scale and multi-type droughts, and improves the model's adaptability in complex drought environments and the foresight, reliability, and timeliness of operational drought monitoring. Its main steps include: dividing the space into monitoring units according to each model; selecting representative candidate factors in each monitoring unit; selecting top-ranked factors for each single-mode factor based on factor similarity indices to generate single-mode drought level sequences; and constructing a classification decision model based on measured drought level sequences to fuse the single-mode drought level sequences, thereby achieving adaptive drought monitoring of the monitoring units.

[0050] like Figure 1 As shown in the figure, this invention discloses a multi-mode collaborative adaptive remote sensing drought monitoring method, which includes the following steps:

[0051] Step 1: Select a spatial division standard for each pattern, and combine multiple standards to cross-divide and determine several evaluation units.

[0052] The term "model" refers to a general term commonly used as an index of drought conditions, including but not limited to soil moisture, surface temperature, and precipitation.

[0053] Preferably, step 1 specifically includes:

[0054] Step 1.1: Select any spatial partitioning standard to divide the space into primary partitions.

[0055] Step 1.2: Select any of the remaining drought monitoring modes, and further divide the space based on the primary zoning and the zoning criteria of the mode to obtain the secondary zoning.

[0056] Step 1.3: Traverse other modes and repeat step 1.2 until all spatial division criteria of all modes have been considered, and the final spatial division unit is obtained, which serves as the monitoring unit for adaptive drought monitoring.

[0057] In this embodiment, three models were selected: soil moisture, surface temperature, and water body characteristics. The spatial classification criteria for soil moisture were vegetation cover zoning, for surface temperature were climate zoning, and for water body characteristics were watershed zoning.

[0058] Step 1.1: Select a soil moisture pattern and determine the surface vegetation cover zoning standard. Identify the surface vegetation cover type as the primary zoning, distinguishing between cultivated land, forest land, water area, urban and rural areas, wetland, and ground cover.

[0059] Step 1.2 involves traversing surface temperature models to determine the secondary zoning standard as climate zoning. Based on the primary zoning, climate types are identified as secondary zoning, distinguishing between temperate continental climate zones, temperate monsoon climate zones, subtropical monsoon climate zones, tropical monsoon climate zones, and plateau / alpine climate zones.

[0060] Step 1.3: Finally, traverse the water body characteristic patterns to determine the watershed zoning standards as the third-level zoning, that is, add watershed conditions to the second-level zoning to form the third-level zoning, including 10 major watersheds: Yangtze River Basin, Yellow River Basin, Pearl River Basin, Hai River Basin, Huai River Basin, Songhua River Basin, Liao River Basin, Southwest Basin, Northwest Basin, and Southeast Basin.

[0061] Step 2: Select a wide range of representative drought monitoring indices as candidate factors for each model.

[0062] Preferably, a wide range of drought indices include, but are not limited to, meteorological drought indices, hydrological drought indices, and agricultural drought indices. For each selected drought model, at least one drought index is selected as a candidate factor for each model.

[0063] The examples selected various representative drought monitoring factors as candidates for three single models: soil moisture model, surface temperature model, and water body characteristic model.

[0064] Step 2.1: For the soil moisture model, the example selected three drought monitoring factors: Soil Moisture Deficit Index (SWDI), Relative Soil Moisture (RSM), and Soil Moisture Index (SMA). Table 1 lists the calculation formulas for the three drought monitoring factors.

[0065] Table 1 Candidate soil moisture factors

[0066]

[0067]

[0068]

[0069] Step 2.2, for the surface temperature model, the example selected three drought monitoring factors: the Shortwave Infrared Moisture Stress Index (MSIWSI), the Temperature State Index (TCI), and the Temperature Vegetation Drought Index (TVDI). Table 1 lists the calculation formulas for the three drought monitoring factors.

[0070] Step 2.3, for the water body characteristic model, the example selected two drought monitoring factors: the Standardized Reservoir Supply Index (SRSI) and the Standardized Water Level Index (SWI). Table 1 lists their calculation formulas.

[0071] Step 3: Optimization and screening of drought factors. For each model, calculate the correlation index between two factors, and select the top-ranked factors for each model based on the comprehensive evaluation index. If there are only two factors, then select one of them.

[0072] Preferably, step 3 specifically includes:

[0073] Step 3.1, calculate the Pearson correlation coefficient of the index.

[0074]

[0075] In the formula, cov() represents the covariance, and σ represents the standard deviation. D and D' are the drought index sequences corresponding to the two candidate factors. i and D′ i This represents the i-th time series value of the drought index, where n is the number of time series values. and σ represents the average value of the drought index series. D Let σ be the standard deviation of the drought index series D. D′ denoted as the standard deviation of the drought index sequence D'.

[0076] Step 3.2, calculate the maximum mutual information coefficient (MIC) of the exponent.

[0077]

[0078] In the formula, p(D,D') is the joint probability density function of random exponential variables D and D', and p(D) and p(D') are the marginal probability density functions of variables D and D', respectively. The higher the mutual information between two random variables, the greater their correlation; conversely, the lower the mutual information, the smaller the correlation between the variables. In the example, the random exponential variables D and D' are also the drought index sequences corresponding to the two candidate factors.

[0079] Step 3.3: Construct a factor similarity index S by calculating the mean of the two. Then, sort the factor similarity indices in ascending order and select the top-ranked drought index for each pattern.

[0080] S=(R+MIC) / 2#(3)

[0081] The specific implementation in the example is as follows:

[0082] Step 3.1: For the soil moisture model, calculate the R-values ​​(Equation 1) and MIC values ​​(Equation 2) between SWDI and RSM, SWDI and SMA, and RSM and SMA. Further calculate the factor similarity index for the average indicators and rank them. After this screening step, SWDI is retained as input data for multi-mode collaborative driving.

[0083] Step 3.2: For the temperature model, calculate the R-values ​​(Equation 1) and MIC values ​​(Equation 2) between MSIWSI and TCI, MSIWSI and TVDI, and TCI and WSDI. Further calculate the comprehensive evaluation index by averaging the indicators and rank them. After this screening step, A is retained as input data for multi-mode collaborative driving.

[0084] Step 3.3: For the soil moisture model, since two indicators A have been selected, no further filtering is needed. Either one can be selected and kept directly as input data for multi-mode collaborative driving. In this case, SRSI is selected.

[0085] Step 4: Reclassify based on the empirical drought level assessment thresholds of each factor to generate a single-mode drought level sequence.

[0086] In this embodiment, the drought severity level sequence is a raster product stored in GeoTIFF format. Each pixel unit of this raster product stores drought severity level information extracted according to drought severity level classification rules (including five levels: no drought, mild drought, moderate drought, severe drought, and extreme drought), thereby assessing the severity of drought in different regions. In this embodiment, the single-mode drought severity level sequence generated by each drought index is provided on a monthly basis. By using hierarchical color coding on the drought severity level product, a thematic map of the drought severity level sequence can be created (Figure 2).

[0087] Preferably, step 4 specifically includes:

[0088] Step 4.1: Reclassify according to the drought level correspondence table of Soil Moisture Deficit Index (SWDI) (Table 2) to generate a single-mode drought level sequence of soil moisture.

[0089] Table 2. Soil Moisture Deficit Index (SWDI) Drought Level Classification

[0090] Drought level No drought mild drought drought Severe drought severe drought SWDI 0≤SWDI -2≤SWDI≤0 -5≤SWDI≤-2 -10≤SWDI≤-5 SWDI≤-10

[0091] Step 4.2: Reclassify according to the drought level correspondence table of Temperature State Index (TCI) (Table 3) to generate a single-mode drought level sequence of production surface temperature.

[0092] Table 3. Correspondence between Temperature State Index (TCI) Drought Levels

[0093]

[0094]

[0095] Step 4.3: Reclassify according to the drought level correspondence table of the Standardized Reservoir Water Supply Index (SRSI) (Table 4) to generate a single-mode drought level sequence of production water body characteristics.

[0096] Table 4. Standardized Reservoir Water Supply Index (SRSI) and Drought Level Correspondence Table

[0097]

[0098] Step 5: Calculate the measured drought level sequence, construct a classification model of drought level sequence of each mode and measured drought level sequence, and realize the adaptive evaluation of multiple modes of monitoring units.

[0099] Preferably, step 5 specifically includes:

[0100] Step 5.1: Collect measured or actual disaster data and calculate the actual drought index.

[0101] Step 5.2: Reclassify the actual drought indicators according to the drought level comparison table (Table 5) to generate a measured drought level sequence.

[0102] Table 5. Correspondence between Comprehensive Yield Reduction Factor and Drought Level (C)

[0103] Drought level No drought mild drought drought Severe drought severe drought C(％) 8≤C 8<C≤16 16<C≤24 24<C≤32 C>32

[0104] Step 5.3: Establish an adaptive evaluation model for the measured drought level sequence of the monitoring unit and the drought level sequence of the single model.

[0105] Step 5.4: Based on the adaptive assessment model, produce the final drought level product of the assessment unit to realize adaptive drought monitoring driven by multi-mode collaboration.

[0106] In this embodiment, the specific implementation is as follows:

[0107] Step 5.1: In the example, the comprehensive reduction rate of production is used as an indicator of the actual drought situation.

[0108] The measured drought severity assessment method uses the comprehensive yield reduction percentage as an indicator. The comprehensive yield reduction percentage is the sum of the products of the proportion of area affected by different degrees of disaster and the average yield reduction percentage over the corresponding area. A preferred implementation example (Zhan Xinye, 2018) uses the following calculation method:

[0109] C=D3×0.9+(D2-D3)×0.55+(D1-D2)×0.2#(4)

[0110] In the formula, C represents the comprehensive yield reduction percentage, and D1, D2, and D3 represent the proportions of the affected area (yield reduction of 10%), the severely affected area (yield reduction of 30%), and the area of ​​total crop failure (yield reduction of 80%) to the total sown area, respectively.

[0111] Step 5.2: According to the measured drought level classification table (Table 5), and in conjunction with Step 5.1, produce the measured drought level sequence.

[0112] Step 5.3: Example: A backpropagation (BP) neural network is selected to construct an adaptive assessment model. Using several parameters from multiple modes as input and measured parameters as output, a BP neural network model is trained to achieve an adaptive assessment of drought conditions. The trained BP neural network model is the adaptive assessment model.

[0113] according to Figure 3 The backpropagation (BP) neural network was configured with a three-layer structure, including an input layer, a hidden layer, and an output layer, transmitting information layer by layer. The input layer had three nodes, representing soil moisture level sequences, surface temperature level sequences, and water body characteristic level sequences from the training sample set. The hidden layer underwent multiple trials to determine the appropriate number of nodes. The output layer had one node, representing the measured drought level sequence determined by the comprehensive yield reduction factor.

[0114] In the input and output layers, the drought severity sequence data comprises five levels, each with a unique label: 0 (no drought), 1 (mild drought), 2 (moderate drought), 3 (severe drought), and 4 (extreme drought). Three node label values ​​from the input layer are randomly selected as input information for the backpropagation (BP) neural network. This input information is passed from the input layer to the hidden layer, processed and computed by the activation function, and then output through the output layer. During training, the sample information is learned by adjusting the connection weights and threshold parameters of each layer based on the desired output label.

[0115] Step 5.4: Based on the trained adaptive assessment model, produce the final drought level product of the assessment unit to realize adaptive drought monitoring driven by multi-mode collaboration.

[0116] In specific implementation, the method proposed in the technical solution of this invention can be automatically executed by those skilled in the art using computer software technology. System devices for implementing the method, such as computer-readable storage media storing the corresponding computer program of the technical solution of this invention and computer equipment including the computer program running the corresponding computer program, should also be within the protection scope of this invention.

[0117] In some possible embodiments, a multi-mode collaboratively driven adaptive remote sensing drought monitoring system is provided, including the following modules:

[0118] The first module is used to set up multiple drought conditions based on important drought characteristic indicators, select a spatial division standard for each model, and determine several monitoring units by combining the cross division of the standards.

[0119] The second module is used to select representative drought monitoring indices as candidate factors for each model;

[0120] The third module is used to calculate the correlation of various factors within the monitoring unit and to select the top-ranked factors for each pattern based on the factor similarity index.

[0121] The fourth module is used to generate single-mode drought level sequences for each model based on the drought level correspondence table of factors;

[0122] The fifth module is used to generate measured drought level sequences, construct an adaptive evaluation model of drought level sequences of each model and measured drought level sequences, and realize adaptive drought level evaluation of monitoring units under multi-mode collaborative driving.

[0123] In some possible embodiments, a multi-mode collaborative adaptive remote sensing drought monitoring system is provided, including a processor and a memory. The memory is used to store program instructions, and the processor is used to call the stored instructions in the memory to execute a multi-mode collaborative adaptive remote sensing drought monitoring method as described above.

[0124] In some possible embodiments, a multi-mode collaborative adaptive remote sensing drought monitoring system is provided, including a readable storage medium on which a computer program is stored. When the computer program is executed, it implements the multi-mode collaborative adaptive remote sensing drought monitoring method described above.

[0125] The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which this invention pertains may make various modifications or additions to the described specific embodiments or use similar methods to substitute them, without departing from the spirit of the invention or exceeding the scope defined by the appended claims.

Claims

1. A multi-mode collaborative adaptive remote sensing drought monitoring method, characterized in that: Includes the following steps, Step 1: Based on key drought characteristics, set up multiple drought models, select a spatial division standard for each model, and determine several monitoring units by combining the cross-division of the standards; the implementation includes the following sub-steps. Step 1.1: Select any spatial partitioning standard to perform primary partitioning of the space; Step 1.2: Select any of the remaining drought monitoring models, and further divide the space based on the primary zoning and the zoning criteria of the model to obtain the secondary zoning; Step 1.3: Traverse other modes and repeat Step 1.2 until all spatial division criteria of all modes have been traversed, and the final spatial division unit is obtained, which serves as the monitoring unit for adaptive drought monitoring. Step 2: Select representative drought monitoring indices as candidate factors for each model; Step 3: Calculate the correlation of various factors within the monitoring unit, and select the top-ranked factors for each pattern based on the factor similarity index. Step 4: Based on the drought level correspondence table of factors, generate the drought level sequence for each production model. Step 5: Generate the measured drought level sequence, construct an adaptive evaluation model of the drought level sequence of each model and the measured drought level sequence, and realize the adaptive drought level evaluation of the monitoring unit under the collaborative drive of multiple models. The implementation method includes the following sub-steps: Step 5.1: Collect measured or actual disaster data and calculate the actual drought index; Step 5.2: Generate a sequence of measured drought levels based on the drought level comparison table; Step 5.3: Establish an adaptive evaluation model for the measured drought level sequence of the monitoring unit and the drought level sequence of the single model; Step 5.4: Based on the adaptive assessment method, produce the final drought level product of the assessment unit to realize adaptive drought monitoring driven by multi-mode collaboration.

2. The multi-mode collaborative adaptive remote sensing drought monitoring method according to claim 1, characterized in that: The system sets up multiple drought conditions, including three modes: soil moisture, surface temperature, and water body characteristics.

3. The multi-mode collaborative adaptive remote sensing drought monitoring method according to claim 1, characterized in that: Representative drought monitoring indices include meteorological drought index, hydrological drought index and agricultural drought index; at least one drought index is selected as a candidate factor for each type of cold weather model.

4. The multi-mode collaborative adaptive remote sensing drought monitoring method according to claim 1, characterized in that: Step 3 is implemented by including the following sub-steps: Step 3.1: Calculate the Pearson correlation coefficient for each factor. In the formula, cov is the covariance, σ is the standard deviation, and D and D' are the drought index sequences corresponding to the two candidate factors, respectively. and This represents the i-th time-series value of the drought index, where n is the number of time-series values. and This represents the average value of the drought index series. The standard deviation of the drought index series D is given. The standard deviation of the drought index series D'; Step 3.2: Calculate the maximum mutual information coefficient (MIC) for each factor; Step 3.3: Construct the factor similarity index S by calculating the mean of the two. Based on the ascending order of factor similarity indicators, the top-ranked drought index is selected for each pattern.

5. A multi-mode collaborative adaptive remote sensing drought monitoring system, characterized in that: This method is used to implement a multi-mode collaboratively driven adaptive remote sensing drought monitoring method as described in any one of claims 1-4.

6. The multi-mode collaborative driven adaptive remote sensing drought monitoring system according to claim 5, characterized in that: Includes the following modules, The first module is used to set up multiple drought conditions based on important drought characteristic indicators, select a spatial division standard for each model, and determine several monitoring units by combining the cross division of the standards. The second module is used to select representative drought monitoring indices as candidate factors for each model; The third module is used to calculate the correlation of various factors within the monitoring unit and to select the top-ranked factors for each pattern based on the factor similarity index. The fourth module is used to generate single-mode drought level sequences for each model based on the drought level correspondence table of factors; The fifth module is used to generate measured drought level sequences, construct an adaptive evaluation model of drought level sequences of each model and measured drought level sequences, and realize adaptive drought level evaluation of monitoring units under multi-mode collaborative driving.

7. The multi-mode collaborative driven adaptive remote sensing drought monitoring system according to claim 5, characterized in that: It includes a processor and a memory, the memory being used to store program instructions, and the processor being used to call the stored instructions in the memory to execute the multi-mode collaborative driven adaptive remote sensing drought monitoring method as described in any one of claims 1-4.

8. The multi-mode collaborative driven adaptive remote sensing drought monitoring system according to claim 5, characterized in that: It includes a readable storage medium on which a computer program is stored, and when the computer program is executed, it implements a multi-mode collaborative driven adaptive remote sensing drought monitoring method as described in any one of claims 1-4.

Images