Irrigation prescription map data fusion method and system based on multi-source data

Abstract

The present application relates to the technical field of irrigation prescription map data fusion, and particularly relates to an irrigation prescription map data fusion method and system based on multi-source data. The method comprises the following steps: obtaining multi-source data of a target area, preprocessing the multi-source data, and generating a standard grid data set; performing feature extraction and normalization processing on the standard grid data set to generate a plurality of sets of normalized feature data; performing weighted fusion calculation on the plurality of sets of normalized feature data using a preset weight to generate comprehensive characteristic value data; mapping the comprehensive characteristic value data to irrigation decision levels and performing spatial visualization expression to generate an irrigation prescription map; the present application provides precise irrigation operation through irrigation prescription map data fusion processing, realizes complete conversion of irrigation decision from data calculation to spatial expression, and makes irrigation decision have a clear spatial position and level distinction.

Description

Technical Field

[0001] This invention relates to the field of irrigation prescription map data fusion technology, and in particular to an irrigation prescription map data fusion method and system based on multi-source data. Background Technology

[0002] Precision irrigation is a core technology for water conservation and efficiency improvement in modern agriculture. Irrigation prescription maps, as the core basis for precision irrigation, directly determine the scientific validity and rationality of irrigation operations. Currently, irrigation prescription map generation often relies on a single data source, resulting in data bias and an inability to comprehensively reflect the combined water requirements of crops and their growing environment. Some methods using multi-source data suffer from spatial misalignment and inconsistent dimensions due to a lack of standardized spatial unification and format processing for data from different sources, affecting the accuracy of the fusion results. Furthermore, existing multi-source data fusion methods often lack clear feature extraction and normalization processes, making it difficult to achieve effective compatibility of different types of feature parameters. Ultimately, this leads to irrigation prescription maps with low spatial resolution and insufficient decision-making accuracy, failing to meet the actual needs of large-scale precision irrigation in farmland. Summary of the Invention

[0003] Therefore, it is necessary to provide a method and system for fusion of irrigation prescription map data based on multi-source data to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, an irrigation prescription map data fusion method based on multi-source data includes the following steps:

[0005] Step S1: Obtain multi-source data for the target area, preprocess the multi-source data, and generate a standard grid dataset;

[0006] Step S2: Perform feature extraction and normalization on the standard grid dataset to generate multiple sets of normalized feature data;

[0007] Step S3: Using preset weights, perform weighted fusion calculations on multiple sets of normalized feature data to generate comprehensive feature value data;

[0008] Step S4: Map the comprehensive feature value data to irrigation decision levels and express them spatially to generate an irrigation prescription map.

[0009] The present invention also provides an irrigation prescription map data fusion system based on multi-source data, used to execute the irrigation prescription map data fusion method based on multi-source data as described above. The irrigation prescription map data fusion system based on multi-source data includes:

[0010] The preprocessing module is used to acquire multi-source data of the target area, preprocess the multi-source data, and generate a standard grid dataset.

[0011] The normalization module is used to extract and normalize features from a standard grid dataset, generating multiple sets of normalized feature data.

[0012] The weighted fusion calculation module is used to perform weighted fusion calculation on multiple sets of normalized feature data using preset weights to generate comprehensive feature value data.

[0013] The visualization module is used to map comprehensive feature value data to irrigation decision levels and perform spatial visualization to generate irrigation prescription maps.

[0014] The beneficial effects of this invention are as follows: By acquiring and preprocessing multi-source data, remote sensing images, soil data, and meteorological data are unified to the same geographic coordinate system and grid resolution, eliminating spatial misalignment and format differences between data from different sources, and ensuring spatial consistency and integrity of subsequent data processing. Feature extraction and normalization are performed on the standard grid dataset, mapping feature parameters of different dimensions and numerical ranges to the [0,1] interval, reducing the interference of feature numerical differences on the fusion results, and improving the comparability and compatibility of multi-source features. Preset weights are used to perform grid-by-grid weighted fusion of multiple sets of normalized feature data, which can stably and quantitatively integrate crop growth and soil moisture information, enabling the comprehensive feature values ​​to objectively reflect the spatial distribution of crop water deficit status within the region. By mapping the comprehensive feature values ​​to irrigation decision levels and achieving spatial visualization, standardized irrigation prescription maps are directly generated, realizing the complete transformation of irrigation decisions from data calculation to spatial expression, giving irrigation decisions clear spatial location and level distinctions, providing an intuitive and directly executable basis for precision irrigation operations, and improving the refinement and standardization of irrigation management. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the steps of a method for fusing irrigation prescription map data based on multi-source data.

[0016] Figure 2 This is a schematic diagram of the process of an irrigation prescription map data fusion system based on multi-source data;

[0017] Figure 3 A diagram showing irrigation prescriptions for farmland;

[0018] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0019] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0020] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0021] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0022] To achieve the above objectives, please refer to Figures 1 to 3 A data fusion method for irrigation prescription maps based on multi-source data includes the following steps:

[0023] Step S1: Obtain multi-source data for the target area, preprocess the multi-source data, and generate a standard grid dataset;

[0024] Step S2: Perform feature extraction and normalization on the standard grid dataset to generate multiple sets of normalized feature data;

[0025] Step S3: Using preset weights, perform weighted fusion calculations on multiple sets of normalized feature data to generate comprehensive feature value data;

[0026] Step S4: Map the comprehensive feature value data to irrigation decision levels and express them spatially to generate an irrigation prescription map.

[0027] In this embodiment of the invention, a target farmland area is used as the processing object. Temporal multispectral remote sensing images of the area are acquired via a satellite platform and identified as the original remote sensing images. Soil moisture and soil temperature data are collected through an IoT sensor network deployed within the target area and identified as the original soil data. Temperature, precipitation, and evaporation data of the target area are collected through ground meteorological stations and identified as the original meteorological data. The original remote sensing images, original soil data, and original meteorological data are integrated into multi-source data for the target area. Using the geographic coordinate system and spatial resolution of the original remote sensing images as a reference, geometric registration is performed on the original soil data and original meteorological data to eliminate spatial offsets between the multi-source data. Then, resampling is performed on the registered data to ensure consistent grid cell size and grid row and column counts across the multi-source data. The data storage format, data encoding method, and spatial indexing rules are unified for the data that has undergone geometric registration and resampling. Outliers, missing values, and redundant information are removed from the data, generating a standard grid dataset covering the entire target area with unified spatial reference, grid specifications, and format. This standard grid dataset provides the basic data carrier for subsequent feature extraction operations.

[0028] Read the standard grid dataset generated in step S1. Based on the remote sensing image data in the standard grid dataset, calculate the normalized vegetation index (NVI) using band operations. The NVI is then used as a crop growth characteristic parameter. Based on the soil data in the standard grid dataset, extract the soil moisture content value and use it as a soil moisture characteristic parameter. This completes the extraction of characteristic parameters related to crop moisture status. Traverse all grid cells in the entire target area and calculate the maximum and minimum values ​​of the crop growth characteristic parameter and soil moisture characteristic parameter within all grid cells. Using extreme value normalization, map the original values ​​of the crop growth characteristic parameter and soil moisture characteristic parameter to the [0,1] value range. Specifically, subtract the minimum value of the corresponding characteristic parameter from the original value and then divide by the difference between the maximum and minimum values ​​of the corresponding characteristic parameter to complete the numerical normalization operation. This generates normalized crop growth characteristic data and normalized soil moisture characteristic data. The two sets of normalized characteristic data maintain the same spatial grid structure as the standard grid dataset, providing input data with a unified dimension for subsequent weighted fusion operations.

[0029] Based on the contribution of crop growth characteristics and soil moisture characteristics to irrigation decisions, a first fusion weight is assigned to the normalized crop growth characteristic data, and a second fusion weight is assigned to the normalized soil moisture characteristic data. Both weight values ​​are within the range [0,1], and the sum of the two weight values ​​is 1. Using a single grid cell of the standard grid dataset as the basic computational unit, a weighted summation operation is performed cell by cell. The value of the normalized crop growth characteristic data in that grid cell is multiplied by the first fusion weight, and the value of the normalized soil moisture characteristic data in that grid cell is multiplied by the second fusion weight. The two product results are added together to obtain the comprehensive characteristic value corresponding to that grid cell. Following the above computational logic, all grid cells in the target area are processed. The comprehensive characteristic values ​​of all grid cells are combined to form comprehensive characteristic value data covering the entire target area. The comprehensive characteristic value data maintains a spatial grid distribution consistent with the standard grid dataset, providing a basis for subsequent irrigation decision level mapping.

[0030] A one-to-one correspondence is established between comprehensive feature values ​​and irrigation decision levels. Continuous intervals of comprehensive feature values ​​are divided, with each interval corresponding to a unique irrigation level. The value of each grid cell in the comprehensive feature value data is substituted into the correspondence to match the irrigation decision level to which that cell belongs, completing the conversion from comprehensive feature value data to irrigation decision levels and forming irrigation level raster data. Map adjustments are performed on the irrigation level raster data to standardize the color coding, raster boundary style, and spatial labeling rules for irrigation levels, while preserving the geographic coordinates and spatial extent information of the standard grid dataset. The adjusted irrigation level raster data is then spatially visualized to form an irrigation prescription map that includes spatial location information, irrigation level information, and a visual representation. This irrigation prescription map provides direct data support for precision irrigation operations in the target area.

[0031] Please refer to [link / reference needed] for further information. Figure 3 A standard grid dataset generated from a global grid, with different colors marking irrigation levels 1-3, corresponding to the generated irrigation level raster maps.

[0032] Preferably, step S1 includes the following steps:

[0033] Step S11: Acquire the original remote sensing images, original soil data and original meteorological data of the target area as multi-source data of the target area;

[0034] Step S12: Convert the multi-source data of the target area to the same geographic coordinate system and grid resolution to generate a standard grid dataset.

[0035] In this embodiment of the invention, a satellite-borne multispectral imaging device is used to acquire on-orbit imaging data of the target farmland area, obtaining temporal multispectral image data including blue, green, red, and near-infrared bands. This image data carries surface reflectance values ​​and imaging time information, and is identified as the original remote sensing image of the target area. An IoT soil sensor network deployed within the target area collects real-time data on volumetric water content and surface temperature within the soil profile. The sensors collect data at fixed time intervals and upload it to a data receiving terminal, identifying this set of soil water content and temperature data as the original soil data. Ground-based meteorological observation stations within the target area collect near-surface temperature, precipitation, and water evaporation values. The meteorological observation equipment continuously collects data at a fixed sampling frequency, forming a continuous observation sequence, and this set of temperature, precipitation, and evaporation data is identified as the original meteorological data. The original remote sensing image, original soil data, and original meteorological data are aggregated and collected according to spatial location identifiers to form multi-source data for the target area. This multi-source data provides the original input data for subsequent spatial unified conversion processing.

[0036] Using the original remote sensing image obtained in step S11 as the spatial reference, the geographic coordinate system parameters, projection parameters, and grid resolution parameters inherent in the original remote sensing image are extracted, and this set of parameters is determined as the reference parameters for unified transformation of multi-source data. Geometric registration processing is performed on the original soil data and the original meteorological data. Based on the spatial coordinate information of the original remote sensing image, the spatial coordinates of the original soil data and the original meteorological data are corrected point by point to eliminate the spatial offset between the original soil data, the original meteorological data, and the original remote sensing image. After completing the geometric registration, resampling processing is performed on the original soil data and the original meteorological data. A bilinear interpolation algorithm is used to adjust the grid size of the original soil data and the original meteorological data to a value that is completely consistent with the grid resolution of the original remote sensing image, so that the multi-source data are consistent in the number of rows and columns of spatial grids, the size of grid cells, and the spatial coverage. The multi-source data that has undergone geometric registration and resampling is subjected to format standardization processing to unify the data storage encoding method, data bit depth and raster organization form, and to remove abnormal sampling points and blank missing values ​​in the data sequence. Finally, a standard grid dataset with unified spatial reference, unified grid size and unified format is generated. This standard grid dataset is directly used for subsequent feature extraction processing.

[0037] Preferably, step S11 includes the following steps:

[0038] Step S111: Acquire temporal multispectral remote sensing images of the target area via a satellite platform as the raw remote sensing images;

[0039] Step S112: Read soil moisture and temperature data from the IoT sensor network deployed in the target area, and obtain air temperature, precipitation and evaporation data, which are used as raw soil data and raw meteorological data, respectively.

[0040] Step S113: Integrate and process the original remote sensing images, original soil data, and original meteorological data to form multi-source data for the target area.

[0041] In this embodiment of the invention, the satellite platform performs on-orbit imaging of the target area along a predetermined orbit. Multi-band surface reflectance information of the target area is acquired through the onboard multispectral imaging payload. The acquisition bands cover the blue, green, red, and near-infrared bands. Multiple image data are continuously acquired at fixed time intervals to form temporal multispectral remote sensing image data. This image data carries corresponding geographic coordinate information, imaging time information, and surface reflectance values. The data carrier is in raster form, with each raster cell storing the surface reflectance value of the corresponding band. The same raster cell corresponds to different reflectance values ​​at different acquisition times. This temporal multispectral remote sensing image is directly used as the original remote sensing image of the target area, providing raw remote sensing input data for subsequent data integration and processing.

[0042] An IoT sensor network is deployed within the target area. Soil moisture and temperature sensors are buried at fixed depths within the soil profile of the target area. Soil volumetric water content and soil temperature values ​​are collected at a fixed sampling frequency. The collected signals are uploaded to a data aggregation node via wired transmission, forming a soil moisture and temperature data sequence. This data sequence carries the coordinates of the corresponding sampling point and the sampling time, serving as the raw soil data. Meteorological observation equipment is deployed in and around the target area. Air temperature sensors collect near-surface atmospheric temperature values, rainfall sensors collect accumulated precipitation values ​​over a period of time, and evaporation sensors collect water surface evaporation values. These values ​​are arranged at fixed time intervals to form a continuous observation sequence. This observation sequence carries the corresponding observation location and observation time, serving as the raw meteorological data. Both the raw soil and raw meteorological data are stored in numerical sequence form, providing ground-based observational raw input data for subsequent data integration and processing.

[0043] The original remote sensing imagery acquired in step S111, the original soil data acquired in step S112, and the original meteorological data are read. Spatial matching processing is performed on the three types of data based on spatial location identifiers. The raster coordinates of the original remote sensing imagery, the sampling point coordinates of the original soil data, and the observation point coordinates of the original meteorological data are uniformly associated. The three types of data are truncated and aggregated according to the spatial boundary range of the target area. Invalid data entries that exceed the spatial range of the target area are removed, and the original remote sensing imagery data, original soil data entries, and original meteorological data entries that completely correspond to the target area are retained. The three types of data after spatial matching and range truncation are centrally stored to form a unified data set. This data set is the multi-source data of the target area. This multi-source data provides complete input data for the subsequent unified conversion to the same geographic coordinate system and grid resolution.

[0044] Preferably, step S12 includes the following steps:

[0045] Step S121: Using the original remote sensing image as a spatial reference, perform geometric registration and resampling operations on the original soil data and original meteorological data;

[0046] Step S122: Standardize the format of the spatially registered data to generate a standard grid dataset.

[0047] In this embodiment of the invention, the original remote sensing image is extracted from the multi-source data obtained in step S113. The projection coordinate system parameters, ellipsoid parameters, central meridian parameters, and pixel resolution parameters of the original remote sensing image are read and determined as unified spatial reference parameters. The coordinates of the sampling points corresponding to the original soil data and the coordinates of the observation points corresponding to the original meteorological data are substituted into the projection transformation formula of the original remote sensing image to perform the conversion operation between geodetic coordinates and plane coordinates. Based on the converted plane coordinate values, the original soil data and the original meteorological data are located in the same spatial plane of the original remote sensing image, completing the geometric registration operation of the original soil data and the original meteorological data, and eliminating the spatial offset between the original soil data, the original meteorological data and the original remote sensing image. Based on the pixel resolution of the original remote sensing image, a grid boundary of fixed size is set. The bilinear interpolation algorithm is used to perform grid-by-grid interpolation on the geometrically registered original soil data and original meteorological data. The original soil data and original meteorological data in discrete point form are converted into raster data in the same row and column form as the original remote sensing image. The resampling operation of the original soil data and original meteorological data is completed, so that the original remote sensing image, original soil data and original meteorological data are completely consistent in terms of spatial coverage, number of grid rows and columns and grid cell size.

[0048] After geometric registration and resampling in step S121, the original remote sensing images, original soil data raster, and original meteorological data raster are standardized in format. The storage encoding method for all three types of data is unified to integer encoding, the data bit depth is unified to 16 bits, the raster data organization is unified to row-major storage mode, and missing value identifiers are unified to fixed values. All grid cells are traversed to identify and remove outliers exceeding the physical value range. Grid cells with missing values ​​are filled using neighborhood mean imputation to ensure that all grid cells contain valid values. The original remote sensing image raster data, original soil data raster data, and original meteorological data raster data, after format standardization, outlier removal, and missing value imputation, are combined and encapsulated. Each type of data maintains a one-to-one correspondence within the same spatial grid, ultimately generating a standard grid dataset that covers the entire target area, has a consistent spatial reference, consistent grid specifications, consistent data format, and valid and complete values. This standard grid dataset is directly used as input data for subsequent feature extraction and normalization processing.

[0049] Preferably, step S2 includes the following steps:

[0050] Step S21: Extract the feature parameters related to crop water status from the standard grid dataset;

[0051] Step S22: Perform numerical normalization on each set of feature parameters to generate multiple sets of normalized feature data.

[0052] In this embodiment of the invention, the standard grid dataset generated in step S122 is read, and the red band reflectance and near-infrared band reflectance values ​​corresponding to the original remote sensing images in the standard grid dataset are extracted. A ratio calculation relationship is constructed based on the difference and summation results of the near-infrared band reflectance and red band reflectance values. The ratio calculation result is determined as the normalized vegetation index (NVI) value, which directly reflects the crop surface cover status and growth level, and serves as a crop growth characteristic parameter. Soil data raster cells after geometric registration and resampling are extracted from the standard grid dataset, and the soil volumetric water content value within each grid cell is read. This value directly reflects the soil's internal water storage state and serves as a soil moisture characteristic parameter. Both the crop growth characteristic parameter and the soil moisture characteristic parameter maintain the same spatial grid structure, number of grid rows and columns, and spatial coverage as the standard grid dataset. These two types of characteristic parameters together constitute a set of characteristic parameters directly related to the crop water status, providing a fixed input object for subsequent numerical normalization processing.

[0053] The algorithm iterates through all grid cells in the standard grid dataset, sequentially checking the corresponding values ​​of crop growth characteristic parameters within each grid cell. The maximum and minimum values ​​of these parameters are determined through global statistical operations. Similarly, the algorithm iterates through the corresponding values ​​of soil moisture characteristic parameters within each grid cell, again using global statistical operations to determine their maximum and minimum values. The original values ​​of the crop growth characteristic parameters are then subtracted from their corresponding minimum values, and divided by the difference between the maximum and minimum values. This linear mapping operation maps all crop growth characteristic parameters to the [0,1] interval, resulting in normalized crop growth characteristic data. The same linear mapping operation is then performed on the soil moisture characteristic parameters, subtracting their corresponding minimum values ​​and dividing by the difference between their maximum and minimum values. This also maps all soil moisture characteristic parameters to the [0,1] interval, resulting in normalized soil moisture characteristic data. Both sets of normalized feature data maintain the same spatial grid structure and grid position correspondence as the standard grid dataset, ultimately generating multiple sets of normalized feature data with unified dimensions, unified numerical ranges, and unified spatial structure. This set of normalized feature data is directly used as the input data for subsequent weighted fusion calculations.

[0054] Preferably, step S21 includes the following steps:

[0055] Step S211: Calculate the normalized vegetation index as a crop growth characteristic based on remote sensing images in the standard grid dataset;

[0056] Step S212: Calculate soil moisture content as a soil moisture feature based on soil data in the standard grid dataset.

[0057] In this embodiment of the invention, the standard grid dataset generated in step S122 is read, and the temporal multispectral remote sensing image raster data after geometric registration, resampling, and format standardization is extracted from the standard grid dataset. The near-infrared band surface reflectance value and the red band surface reflectance value corresponding to each grid cell in the remote sensing image raster data are read. Taking a single grid cell as the calculation unit, the normalized vegetation index is calculated by using the ratio of band difference to band sum. The specific calculation logic is as follows: the near-infrared band reflectance value is subtracted from the red band reflectance value to obtain the difference result; the near-infrared band reflectance value and the red band reflectance value are added to obtain the sum result; and the difference result and the sum result are divided to obtain the normalized vegetation index value corresponding to the grid cell. According to the above calculation logic, all grid cells within the target area are traversed to complete the unit-by-unit calculation of the normalized vegetation index across the entire area. The normalized vegetation index values ​​corresponding to all grid cells are combined to form raster feature data with the same spatial structure as the standard grid dataset. This normalized vegetation index raster data is directly used as crop growth features, providing remote sensing feature input data for subsequent feature parameter normalization processing.

[0058] Read the standard grid dataset generated in step S122, and extract the soil raster data from the standard grid dataset after geometric registration, resampling, and format standardization. The soil raster data uses the spatial grid of the standard grid dataset as a carrier, and stores the original soil observation values ​​at the corresponding location in each grid cell. The original observation values ​​are converted using the soil volumetric water content conversion formula. During the conversion process, a fixed physical calibration relationship is used to convert the original electrical signal values ​​output by the sensor into soil volumetric water content values. All grid cells in the target area are traversed to complete the unit-by-unit conversion of soil volumetric water content in the entire area. The soil volumetric water content values ​​corresponding to all grid cells are combined to form raster feature data with the same spatial structure as the standard grid dataset. This soil volumetric water content raster data is directly used as soil moisture feature, providing soil-type feature input data for subsequent feature parameter normalization processing.

[0059] Preferably, step S22 includes the following steps:

[0060] Step S221: For each extracted set of feature parameters, calculate its maximum and minimum values ​​within the entire target region;

[0061] Step S222: Based on the maximum and minimum values ​​of each set of feature parameters, map the feature values ​​to the [0,1] interval to generate multiple sets of normalized feature data.

[0062] In this embodiment of the invention, two sets of feature parameters are read: the Normalized Difference Vegetation Index (NDI) obtained in step S211 and the Soil Moisture Content obtained in step S212. Using the target area defined by the standard grid dataset as the statistical boundary, all grid cells within the target area are traversed row by row and column by column. The corresponding value of the NDI within each grid cell is extracted. Through a global point-by-point comparison operation, the maximum and minimum values ​​encountered during the traversal are retained. The maximum value is determined as the global maximum value of the NDI within the target area, and the minimum value is determined as the global minimum value of the NDI within the target area. Following the same traversal path and comparison logic, all grid cells within the target area are traversed row by row and column by column. The corresponding value of the Soil Moisture Content within each grid cell is extracted. Through a global point-by-point comparison operation, the maximum and minimum values ​​encountered during the traversal are retained. The maximum value is determined as the global maximum value of the Soil Moisture Content within the target area, and the minimum value is determined as the global minimum value of the Soil Moisture Content within the target area. The global maximum and global minimum values ​​of the two sets of feature parameters are fixed values, providing fixed extreme value constraints for subsequent interval mapping operations.

[0063] Based on the global maximum and minimum values ​​of the Normalized Difference Vegetation Index (NDVI) and the global maximum and minimum values ​​of Soil Moisture Content determined in step S221, an extreme value linear mapping algorithm is used to perform interval transformation operations on the feature parameters of each grid cell within the target area. For the NDVI, the global minimum value is subtracted from the original feature value in a single grid cell, and the result is divided by the difference between the global maximum and the global minimum value. The result obtained after the operation is limited to the interval [0,1]. The above operation is performed on grid cells one by one to form NDVI normalized feature data with a spatial structure consistent with the standard grid dataset. For Soil Moisture Content, the same operation logic is used: the global minimum value is subtracted from the original feature value in a single grid cell, and the result is divided by the difference between the global maximum and the global minimum value to obtain a mapping result within the interval [0,1]. The above operation is performed on grid cells one by one to form Soil Moisture Normalized Feature Data with a spatial structure consistent with the standard grid dataset. The two sets of mapped raster data maintain uniformity in spatial location, grid size, and data format, together forming multiple sets of normalized feature data. These multiple sets of normalized feature data are directly used as input data for subsequent weighted fusion calculations.

[0064] Preferably, step S3 includes the following steps:

[0065] Step S31: Assign corresponding fusion weights to each set of features in the multiple sets of normalized feature data;

[0066] Step S32: Based on the fusion weights, perform weighted summation on a grid-by-grid basis on multiple sets of normalized feature data to generate comprehensive feature value data.

[0067] In this embodiment of the invention, multiple sets of normalized feature data generated in step S222 are read. These sets of normalized feature data include normalized vegetation index (NVI) normalized feature data and soil moisture content normalized feature data. Both sets of normalized feature data use values ​​in the range [0,1] and maintain the same spatial grid structure. The first fusion weight corresponding to the NNV normalized feature data is set to a fixed value of 0.4, and the second fusion weight corresponding to the soil moisture content normalized feature data is set to a fixed value of 0.6. Both the first and second fusion weights are within the range [0,1], and the sum of the first and second fusion weights equals 1. A unique binding relationship is established between the two sets of fixed weights and their corresponding normalized feature data, ensuring that the NNV normalized feature data uses only the first fusion weight in subsequent calculations, and the soil moisture content normalized feature data uses only the second fusion weight in subsequent calculations. The two sets of fusion weights are stored as fixed parameters, providing constant weight constraints for subsequent weighted summation calculations per grid cell.

[0068] Based on the first and second fusion weights determined in step S31, and using a single grid cell within the target area as the basic operational unit, the normalized feature data within each grid cell is weighted and summed sequentially according to the row and column traversal order of the standard grid dataset. For a single grid cell, the mapping value of the normalized vegetation index normalized feature data within that cell is extracted, and this value is multiplied by the first fusion weight to obtain the first weighted result; the mapping value of the soil moisture content normalized feature data within that cell is extracted, and this value is multiplied by the second fusion weight to obtain the second weighted result; the first weighted result and the second weighted result are added to obtain the comprehensive feature value corresponding to that grid cell. Following the above operational process, all grid cells within the target area are traversed, and the comprehensive feature values ​​of all grid cells are calculated unit by unit. The comprehensive feature values ​​corresponding to all grid cells are combined according to the spatial arrangement of the standard grid dataset to form comprehensive feature value data covering the entire target area and spatially corresponding one-to-one with the standard grid dataset. This comprehensive feature value data is directly used as input data for subsequent irrigation decision level mapping and spatial visualization.

[0069] Preferably, step S4 includes the following steps:

[0070] Step S41: Establish the correspondence between comprehensive feature value data and different irrigation volume levels;

[0071] Step S42: Based on the correspondence, convert the comprehensive feature value data into an irrigation level raster map and adjust the map to generate an irrigation prescription map.

[0072] In this embodiment of the invention, the comprehensive feature value data generated in step S32 is read. This comprehensive feature value data is distributed in the interval [0,1] and has a spatial grid structure consistent with the standard grid dataset. The interval [0,1] is divided into four continuous and non-overlapping numerical intervals at fixed intervals: the first interval is [0,0.25], the second interval is (0.25,0.50], the third interval is (0.50,0.75], and the fourth interval is (0.75,1.00). The first numerical interval corresponds to the first irrigation level, which is the minimum irrigation water volume per unit area; the second numerical interval corresponds to the second minimum irrigation level, which is the second minimum irrigation water volume per unit area; the third numerical interval corresponds to the third irrigation level, which is the second maximum irrigation water volume per unit area; and the fourth numerical interval corresponds to the fourth irrigation level, which is the maximum irrigation water volume per unit area. These four numerical intervals and four irrigation levels form a one-to-one fixed mapping relationship. This mapping relationship is stored as a combination of numerical intervals and level identifiers, providing a unique basis for subsequent conversion of comprehensive feature value data into irrigation levels.

[0073] Based on the correspondence between the comprehensive feature values ​​and irrigation levels established in step S41, the corresponding values ​​in the comprehensive feature value data are extracted from each grid cell according to the row and column traversal order of the standard grid dataset. These values ​​are then substituted into the corresponding value range for determination, and the unique irrigation level assigned to each grid cell is obtained. This irrigation level is recorded in the corresponding grid cell using a fixed code. After traversal, the irrigation level codes of all grid cells are combined to form irrigation level raster data. This irrigation level raster data retains the geographic coordinate system, grid resolution, and spatial coverage of the standard grid dataset. Map adjustment operations are performed on the irrigation level raster data to unify the coding identifiers, raster filling rules, and boundary representations of different irrigation levels. The geographic coordinate information and irrigation level coding information of each grid cell are retained, and redundant labels and invalid codes in the raster data are removed. The adjusted irrigation level raster data is then spatially visualized and encapsulated to form an irrigation prescription map containing spatial location information, irrigation volume information, grid structure information, and visualization representation information. This irrigation prescription map provides a direct data carrier for irrigation operations in the target area.

[0074] The present invention also provides an irrigation prescription map data fusion system based on multi-source data, used to execute the irrigation prescription map data fusion method based on multi-source data as described above. The irrigation prescription map data fusion system based on multi-source data includes:

[0075] Preprocessing module 101 is used to acquire multi-source data of the target area, preprocess the multi-source data, and generate a standard grid dataset;

[0076] The normalization processing module 102 is used to perform feature extraction and normalization processing on the standard grid dataset to generate multiple sets of normalized feature data;

[0077] The weighted fusion calculation module 103 is used to perform weighted fusion calculation on multiple sets of normalized feature data using preset weights to generate comprehensive feature value data.

[0078] The visualization module 104 is used to map the comprehensive feature value data to irrigation decision levels and perform spatial visualization to generate irrigation prescription maps.

[0079] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for fusing irrigation prescription map data based on multi-source data, characterized in that, Includes the following steps: Step S1: Obtain multi-source data for the target area, preprocess the multi-source data, and generate a standard grid dataset; Step S2: Perform feature extraction and normalization on the standard grid dataset to generate multiple sets of normalized feature data; Step S3: Using preset weights, perform weighted fusion calculations on multiple sets of normalized feature data to generate comprehensive feature value data; Step S4: Map the comprehensive feature value data to irrigation decision levels and express them spatially to generate an irrigation prescription map.

2. The irrigation prescription map data fusion method based on multi-source data according to claim 1, characterized in that, Step S1 includes the following steps: Step S11: Acquire the original remote sensing images, original soil data and original meteorological data of the target area as multi-source data of the target area; Step S12: Convert the multi-source data of the target area to the same geographic coordinate system and grid resolution to generate a standard grid dataset.

3. The irrigation prescription map data fusion method based on multi-source data according to claim 2, characterized in that, Step S11 includes the following steps: Step S111: Acquire temporal multispectral remote sensing images of the target area via a satellite platform as the raw remote sensing images; Step S112: Read soil moisture and temperature data from the IoT sensor network deployed in the target area, and obtain air temperature, precipitation and evaporation data, which are used as raw soil data and raw meteorological data, respectively. Step S113: Integrate and process the original remote sensing images, original soil data, and original meteorological data to form multi-source data for the target area.

4. The irrigation prescription map data fusion method based on multi-source data according to claim 2, characterized in that, Step S12 includes the following steps: Step S121: Using the original remote sensing image as a spatial reference, perform geometric registration and resampling operations on the original soil data and original meteorological data; Step S122: Standardize the format of the spatially registered data to generate a standard grid dataset.

5. The irrigation prescription map data fusion method based on multi-source data according to claim 1, characterized in that, Step S2 includes the following steps: Step S21: Extract the feature parameters related to crop water status from the standard grid dataset; Step S22: Perform numerical normalization on each set of feature parameters to generate multiple sets of normalized feature data.

6. The irrigation prescription map data fusion method based on multi-source data according to claim 5, characterized in that, Step S21 includes the following steps: Step S211: Calculate the normalized vegetation index as a crop growth characteristic based on remote sensing images in the standard grid dataset; Step S212: Calculate soil moisture content as a soil moisture feature based on soil data in the standard grid dataset.

7. The irrigation prescription map data fusion method based on multi-source data according to claim 5, characterized in that, Step S22 includes the following steps: Step S221: For each extracted set of feature parameters, calculate its maximum and minimum values ​​within the entire target region; Step S222: Based on the maximum and minimum values ​​of each set of feature parameters, map the feature values ​​to the [0,1] interval to generate multiple sets of normalized feature data.

8. The irrigation prescription map data fusion method based on multi-source data according to claim 1, characterized in that, Step S3 includes the following steps: Step S31: Assign corresponding fusion weights to each set of features in the multiple sets of normalized feature data; Step S32: Based on the fusion weights, perform weighted summation on a grid-by-grid basis on multiple sets of normalized feature data to generate comprehensive feature value data.

9. The irrigation prescription map data fusion method based on multi-source data according to claim 1, characterized in that, Step S4 includes the following steps: Step S41: Establish the correspondence between comprehensive feature value data and different irrigation volume levels; Step S42: Based on the correspondence, convert the comprehensive feature value data into an irrigation level raster map and adjust the map to generate an irrigation prescription map.

10. A data fusion system for irrigation prescription maps based on multi-source data, characterized in that, For executing the irrigation prescription map data fusion method based on multi-source data as described in claim 1, the irrigation prescription map data fusion system based on multi-source data includes: The preprocessing module is used to acquire multi-source data of the target area, preprocess the multi-source data, and generate a standard grid dataset. The normalization module is used to extract and normalize features from a standard grid dataset, generating multiple sets of normalized feature data. The weighted fusion calculation module is used to perform weighted fusion calculation on multiple sets of normalized feature data using preset weights to generate comprehensive feature value data. The visualization module is used to map comprehensive feature value data to irrigation decision levels and perform spatial visualization to generate irrigation prescription maps.

Images