Crop disease and pest monitoring method and device fused with multi-source satellite remote sensing data

Abstract

The embodiment of the invention relates to the technical field of remote sensing data processing and crop disease and insect pest monitoring, and provides a crop disease and insect pest monitoring method and device fusing multi-source satellite remote sensing data, and the method comprises the steps: obtaining first satellite image data, second satellite image data, field survey data and satellite meteorological data of a target region; extracting target crop distribution data of the target area according to the first satellite image data and the field survey data; performing field parcel segmentation according to the second satellite image data and the target crop distribution data through an edge detection method to obtain field parcel vector data; inputting the satellite meteorological data into a pest and disease damage meteorological index evaluation model to calculate a ten-day-by-day pest and disease damage index of the target area; and performing mask operation processing on the pest and disease attack index in every ten days based on the land parcel vector data to obtain a target pest and disease attack index of each land parcel in each ten days. Therefore, real-time monitoring, classification and early warning of crop diseases and insect pests are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of remote sensing data processing and crop pest and disease monitoring, and in particular to a method and device for monitoring crop pests and diseases by fusing multi-source satellite remote sensing data. Background Art

[0002] Traditional pest and disease monitoring methods mostly rely on ground surveys, which have problems such as high cost, low efficiency, and limited coverage. Remote sensing technology can achieve large-area, rapid, and accurate pest and disease monitoring through the acquisition and analysis of satellite images. However, a single data source often has problems such as incomplete information and insufficient accuracy. Therefore, classifying pest and diseases by fusing multi-source satellite remote sensing data has become a research hotspot.

[0003] Most current methods focus on feature extraction and classification of a single data source, while ignoring the complementarity and redundancy between multi-source data. Existing feature extraction methods often rely on manually designed feature extractors, which may not be able to comprehensively capture complex features related to pest and diseases. In addition, due to the differences in spectral and texture features between different crops and different types of pest and diseases, more refined feature extraction methods are needed. Most current pest and disease classification models are trained based on specific crops in specific regions and lack generalization ability. When applied to other regions or other crops, the classification performance of the model may decrease significantly. Most existing pest and disease classification methods require a long data processing time and cannot meet the needs of real-time monitoring. In agricultural production, the timely detection and prevention of pest and diseases are crucial. Therefore, it is necessary to improve the speed of data processing and classification to achieve pest and disease monitoring. Summary of the Invention

[0004] The present invention provides a method and device for monitoring crop pests and diseases by fusing multi-source satellite remote sensing data, to solve the defects in the prior art that crop pests and diseases cannot be accurately predicted and real-time monitoring cannot be achieved, and to realize real-time monitoring, classification, and early warning of crop pests and diseases.

[0005] The present invention provides a method for monitoring crop pests and diseases by fusing multi-source satellite remote sensing data, including: Obtaining first satellite image data, second satellite image data, field survey data, and satellite meteorological data of a target area, where the first satellite image data and the second satellite image data are remote sensing data obtained by multi-source satellites with different resolutions, and the resolution of the first satellite image data is less than that of the second satellite image data; Extracting target crop distribution data of the target area according to the first satellite image data and the field survey data; By means of edge detection method, perform field block segmentation based on the second satellite image data and the target crop distribution data to obtain the vector data of the field blocks corresponding to each type of crop in multiple study areas; Input the satellite meteorological data into the pest and disease meteorological index evaluation model, and calculate the decadal pest and disease incidence indexes of the target area through the pest and disease meteorological index evaluation model; Perform a masking operation on the decadal pest and disease incidence indexes based on the vector data of the field blocks to obtain the target pest and disease incidence indexes of each field block for each decade.

[0006] In a possible implementation manner, the method further includes: Input the first satellite image data into the supervised classification model, and output the distribution results of each type of crop in the target area through the supervised classification model; Correct and supplement the distribution results of each type of crop in the target area based on the field investigation data to obtain the target crop distribution data of the target area; Wherein, the supervised classification model is trained through the following steps: Obtain the sample investigation data and sample satellite image data of the target area; Select the target areas of each type of crop on the sample satellite image data as training samples according to the sample investigation data; Extract the target features of each type of crop from the sample satellite image data; Train the supervised classification model based on the target features and the training samples.

[0007] In a possible implementation manner, the method further includes: Perform data fusion on the second satellite image data and the target crop distribution data to obtain fusion data; Based on the data characteristics of the second satellite image data, identify the field block boundary information in the fusion data through the edge detection method; Match the field block boundary information with the distribution data of each type of crop to obtain the field block segmentation results of multiple study areas corresponding to each type of crop; Convert the field block segmentation results into vector data to obtain the vector data of the field blocks corresponding to multiple study areas for each type of crop.

[0008] In a possible implementation manner, the method further includes: Obtain the third satellite image data, the fourth satellite image data, and the fifth satellite image data with a target time resolution of the target area. The third satellite image data and the fourth satellite image data are remote sensing data obtained by multi-source satellites with different resolutions; Extract the first pest and disease meteorological data from the fifth satellite image data, and interpolate the meteorological data into target pest and disease meteorological data with a target resolution, where the target pest and disease meteorological data includes temperature, air pressure, specific humidity, precipitation, and wind speed; Construct a pest and disease meteorological index evaluation model based on the target pest and disease meteorological data, the third satellite image data, and the fourth satellite image data.

[0009] In a possible implementation, the method further includes: Extract the second pest and disease meteorological data from the satellite meteorological data; Calculate the daily contribution value of the second pest and disease meteorological data to the crop infection within a period based on the second pest and disease meteorological data; Calculate the ten-day contribution average based on the daily contribution value; Calculate the ten-day pest and disease incidence index of the target area based on the ten-day contribution average using a comprehensive meteorological data evaluation method.

[0010] In a possible implementation, the method further includes: Set a mask layer based on the plot vector data; Perform a masking operation on the ten-day pest and disease incidence index and the mask layer to obtain the target pest and disease incidence index of each plot for each ten-day period.

[0011] In a possible implementation, the method further includes: Attach the target pest and disease incidence index of each plot for each ten-day period as an attribute to the attribute table of the corresponding plot vector data.

[0012] The present invention also provides a crop pest and disease monitoring device that fuses multi-source satellite remote sensing data, including the following modules: An acquisition module for acquiring the first satellite image data, the second satellite image data, on-site investigation data, and satellite meteorological data of a target area, where the first satellite image data and the second satellite image data are remote sensing data obtained by multi-source satellites with different resolutions, and the resolution of the first satellite image data is less than that of the second satellite image data; An extraction module for extracting the target crop distribution data of the target area according to the first satellite image data and the on-site investigation data; A segmentation module for performing field block segmentation on the second satellite image data and the target crop distribution data by an edge detection method to obtain the plot vector data of multiple research areas corresponding to each crop; A calculation module, configured to input the satellite meteorological data into a pest and disease meteorological index evaluation model, and calculate the decadal pest and disease incidence indexes of the target area through the pest and disease meteorological index evaluation model; A processing module, configured to perform a masking operation on the decadal pest and disease incidence indexes based on the plot vector data to obtain the target pest and disease incidence indexes of each plot for each decade.

[0013] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the method for monitoring crop pests and diseases by fusing multi-source satellite remote sensing data as described in any one of the above is implemented.

[0014] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the method for monitoring crop pests and diseases by fusing multi-source satellite remote sensing data as described in any one of the above is implemented.

[0015] The present invention also provides a computer program product, including a computer program. When the computer program is executed by a processor, the method for monitoring crop pests and diseases by fusing multi-source satellite remote sensing data as described in any one of the above is implemented.

[0016] The method and device for monitoring crop pests and diseases by fusing multi-source satellite remote sensing data provided by the present invention obtain the first satellite image data, the second satellite image data, the field survey data, and the satellite meteorological data of the target area. The first satellite image data and the second satellite image data are remote sensing data obtained by multi-source satellites with different resolutions, and the resolution of the first satellite image data is less than that of the second satellite image data; according to the first satellite image data and the field survey data, the target crop distribution data of the target area is extracted; through an edge detection method, the plot vector data of multiple study areas corresponding to each crop is obtained by performing plot segmentation based on the second satellite image data and the target crop distribution data; the satellite meteorological data is input into a pest and disease meteorological index evaluation model, and the decadal pest and disease incidence indexes of the target area are calculated through the pest and disease meteorological index evaluation model; a masking operation is performed on the decadal pest and disease incidence indexes based on the plot vector data to obtain the target pest and disease incidence indexes of each plot for each decade. Compared with the limitations in the current crop pest and disease classification methods, especially the defects of high cost, low efficiency, poor real-time performance in the traditional ground survey method and incomplete information and insufficient accuracy of a single satellite data source, in this solution, large-scale and continuous crop pest and disease information is obtained through satellite remote sensing technology, reducing the frequency and scope of field surveys, improving the economic benefits of monitoring. At the same time, the automated data processing and classification process also further reduces the dependence on professional personnel and improves the efficiency and reliability of monitoring. Description of the Drawings

[0017] To more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0018] Figure 1 It is one of the schematic flowcharts of the crop pest and disease monitoring method that fuses multi-source satellite remote sensing data provided by the present invention.

[0019] Figure 2 It is the second schematic flowchart of the crop pest and disease monitoring method that fuses multi-source satellite remote sensing data provided by the present invention.

[0020] Figure 3 It is the third schematic flowchart of the crop pest and disease monitoring method that fuses multi-source satellite remote sensing data provided by the present invention.

[0021] Figure 4 It is the schematic structural diagram of the crop pest and disease monitoring device that fuses multi-source satellite remote sensing data provided by the present invention.

[0022] Figure 5 It is the schematic structural diagram of the electronic device provided by the present invention. Detailed implementation manners

[0023] To make the objectives, technical solutions and advantages of the present invention clearer, the following will clearly and completely describe the technical solutions in the present invention with reference to the drawings in the present invention. Obviously, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments in the present invention belong to the scope protected by the present invention.

[0024] To facilitate the understanding of the embodiments of the present invention, the following will further explain with specific embodiments with reference to the drawings. The embodiments do not limit the embodiments of the present invention.

[0025] Figure 1 It is one of the schematic flowcharts of the crop pest and disease monitoring method that fuses multi-source satellite remote sensing data provided by the present invention. As Figure 1 shown, the method includes the following: The core objective of the embodiments of the present invention is to solve the limitations existing in the current crop pest and disease classification methods, especially for the problems of high cost, low efficiency, poor real-time performance in traditional ground survey methods and incomplete information and insufficient accuracy in single satellite data sources. A crop pest and disease monitoring method that fuses multi-source satellite remote sensing data is proposed.

[0026] Specifically, by integrating multi-source remote sensing data from different satellite systems and comprehensively utilizing their respective spectral information, spatial resolution, and temporal resolution, etc., the ability to capture crop pest and disease characteristics is enhanced, thereby improving the accuracy of classification. This method of multi-source data fusion can more comprehensively reflect the actual situation of crop pests and diseases and reduce classification errors caused by insufficient single data sources. Enhance the generalization ability of the classification model and construct a pest and disease classification model with strong generalization ability. By training a dataset containing multi-source features, the model can learn more diverse pest and disease characteristics, thus adapting to the classification requirements of pests and diseases in different regions, different crops, and different growth stages. This improvement in generalization ability makes this method more flexible and practical in actual applications. The embodiments of the present invention can also achieve real-time monitoring and early warning of crop pests and diseases. By integrating multi-source satellite remote sensing data and combining advanced data processing and classification algorithms, rapid identification and classification of crop pests and diseases can be achieved. This ability of real-time monitoring can promptly detect and respond to the outbreak of pests and diseases, provide timely prevention and control guidance for agricultural production, and reduce the impact of pests and diseases on crop yields. Compared with traditional ground survey methods, this method can significantly reduce labor, material, and time costs. By obtaining large-scale and continuous crop pest and disease information through satellite remote sensing technology, the frequency and scope of on-site surveys are reduced, and the economic benefits of monitoring are improved. At the same time, the automated data processing and classification process also further reduces the dependence on professional personnel and improves the efficiency and reliability of monitoring.

[0027] The following is a detailed description: S11. Obtain the first satellite image data, the second satellite image data, on-site survey data, and satellite meteorological data of the target area.

[0028] In the embodiments of the present invention, the first satellite image data and the second satellite image data are remote sensing data obtained by multi-source satellites with different resolutions.

[0029] The remote sensing image data of the target area can be obtained by high-resolution satellites and low-resolution satellites respectively. High-resolution satellite images can clearly show the details of ground objects, such as the planting boundaries of crops, the shapes of fields, etc., providing basic information for the subsequent extraction of crop distribution data. For example, satellite images with a resolution of 0.5 meters or higher can more accurately distinguish the planting areas of different crops. Although low-resolution satellite images are not as detailed as high-resolution images, their coverage is wider and they can provide macroscopic information over a larger area. For example, satellite images with a resolution of about 16 meters can be used to analyze land use types and vegetation coverage within the area to assist in field segmentation.

[0030] Obtain on-site investigation data of the target area through on-site inspections. On-site investigators can record information such as the planted varieties, growth stages, and planting densities of crops. These data can provide an accurate reference basis for the extraction of crop distribution data, making up for the deficiencies of satellite image data in crop identification. For example, during on-site investigations, investigators can use GPS positioning devices to record the coordinates of different crop planting areas for subsequent comparative analysis with satellite image data.

[0031] Obtain satellite meteorological data of the target area. These data usually include meteorological elements such as temperature, precipitation, humidity, and wind speed. These meteorological data have an important impact on the occurrence and spread of pests and diseases and are important input data for the meteorological index evaluation model of pests and diseases. For example, the combined conditions of temperature and humidity may affect the hatching and growth rates of certain pests and diseases.

[0032] S12. According to the first satellite image data and the on-site investigation data, extract the target crop distribution data of the target area.

[0033] Input the first satellite image data into a pre-trained supervised classification model, and output the distribution results of each crop in the target area through the supervised classification model; based on the on-site investigation data, correct and supplement the distribution results of each crop in the target area to obtain the target crop distribution data of the target area.

[0034] S13. Through an edge detection method, perform field block segmentation according to the second satellite image data and the target crop distribution data to obtain the vector data of the field blocks of multiple study areas corresponding to each crop.

[0035] Fuse the second satellite image data and the target crop distribution data to obtain fused data; based on the data characteristics of the second satellite image data, identify the field block boundary information in the fused data through an edge detection method; match the field block boundary information with the distribution data of each crop to obtain the field block segmentation results of multiple study areas corresponding to each crop; convert the field block segmentation results into vector data to obtain the vector data of the field blocks of multiple study areas corresponding to each crop. Edge detection algorithms such as the Canny edge detection algorithm can find the boundary lines between field blocks, which are usually formed by differences in color, texture, etc. between different crop planting areas. Multiple study areas correspond to different crops, and corresponding vector data of field blocks are generated for each study area. Vector data is a spatial data format that uses geometric elements such as points, lines, and surfaces to represent geographical entities, and it can accurately describe the shape, boundary, etc. of field blocks.

[0036] S14. Input the satellite meteorological data into the pest and disease meteorological index evaluation model, and calculate the decadal pest and disease incidence index of the target area through the pest and disease meteorological index evaluation model.

[0037] In the embodiment of the present invention, it is first necessary to construct a pest and disease meteorological index evaluation model. Specifically, it includes: obtaining the third satellite image data, the fourth satellite image data and the fifth satellite image data with the target time resolution of the target area. The third satellite image data and the fourth satellite image data are remote sensing data obtained by multi-source satellites with different resolutions; extracting the first pest and disease meteorological data from the fifth satellite image data, and interpolating the meteorological data into the target pest and disease meteorological data with the target resolution. The target pest and disease meteorological data includes air temperature, air pressure, specific humidity, precipitation and wind speed; constructing a pest and disease meteorological index evaluation model based on the target pest and disease meteorological data, the third satellite image data and the fourth satellite image data. Among them, the first pest and disease meteorological data includes, but is not limited to, key pest and disease meteorological data such as air temperature, air pressure, specific humidity, precipitation and wind speed.

[0038] Furthermore, extract the second pest and disease meteorological data from the satellite meteorological data; calculate the daily contribution value of the second pest and disease meteorological data to the crop infection within the period; calculate the decadal contribution mean value based on the daily contribution value; calculate the decadal pest and disease incidence index of the target area based on the decadal contribution mean value by using the meteorological data comprehensive evaluation method. Among them, the second pest and disease meteorological data includes, but is not limited to, key pest and disease meteorological data such as air temperature, air pressure, specific humidity, precipitation and wind speed. Here, the first and the second are only used to distinguish the data used in model training from the data input into the model.

[0039] S15. Perform a masking operation on the decadal pest and disease incidence index based on the plot vector data to obtain the target pest and disease incidence index of each plot for each decade.

[0040] Set a masking layer based on the plot vector data; perform a masking operation on the decadal pest and disease incidence index and the masking layer to obtain the target pest and disease incidence index of each plot for each decade. The purpose of the masking operation is to correspond the pest and disease incidence index with specific fields, so as to analyze the pest and disease incidence of each field in different time periods. For example, take each field area in the plot vector data as a mask, and crop and extract the pest and disease incidence index data, so that each field corresponds to an independent pest and disease incidence index value. The target pest and disease incidence index of each plot for each decade can be used to analyze the pest and disease occurrence risks of different fields in different time periods, providing a scientific basis for pest and disease monitoring and control. For example, if the pest and disease incidence index of a certain field in a certain decade is high, it indicates that the risk of pest and disease occurrence in this field is relatively large during this period, and timely control measures need to be taken.

[0041] The crop pest and disease monitoring method that integrates multi-source satellite remote sensing data provided by the present invention obtains the first satellite image data, the second satellite image data, field investigation data, and satellite meteorological data of the target area. The first satellite image data and the second satellite image data are remote sensing data obtained by multi-source satellites with different resolutions, and the resolution of the first satellite image data is less than that of the second satellite image data; according to the first satellite image data and the field investigation data, the target crop distribution data of the target area is extracted; through the edge detection method, the field block segmentation is performed according to the second satellite image data and the target crop distribution data to obtain the plot vector data of multiple study areas corresponding to each crop; the satellite meteorological data is input into the pest and disease meteorological index evaluation model, and the ten-day pest and disease incidence index of the target area is calculated through the pest and disease meteorological index evaluation model; based on the plot vector data, a masking operation is performed on the ten-day pest and disease incidence index to obtain the target pest and disease incidence index of each plot in each ten-day period. Compared with the limitations existing in the current crop pest and disease classification methods, especially the defects of the traditional ground survey method such as high cost, low efficiency, poor real-time performance, incomplete information and insufficient accuracy of a single satellite data source, by this method, large-scale and continuous crop pest and disease information is obtained through satellite remote sensing technology, reducing the frequency and scope of field investigations, improving the economic benefits of monitoring. At the same time, the automated data processing and classification process also further reduces the dependence on professional personnel and improves the efficiency and reliability of monitoring.

[0042] Figure 2 It is the second flow schematic diagram of the crop pest and disease monitoring method that integrates multi-source satellite remote sensing data provided by the present invention. As Figure 2 shown, this method includes the following: S21. Obtain the first satellite image data, the second satellite image data, field investigation data, and satellite meteorological data of the target area.

[0043] In the embodiment of the present invention, the first satellite image data and the second satellite image data are remote sensing data obtained by multi-source satellites with different resolutions.

[0044] As Figure 3 shown, the remote sensing image data of the target area can be obtained by a high-resolution satellite and a low-resolution satellite respectively. The high-resolution satellite image can clearly display the details of ground objects, such as the planting boundaries of crops, the shapes of fields, etc., providing basic information for the subsequent extraction of crop distribution data. For example, satellite images with a resolution of 0.5 meters or higher can more accurately distinguish the planting areas of different crops. Although the low-resolution satellite image is not as detailed as the high-resolution image, its coverage range is wider and it can provide macroscopic information of a larger range. For example, satellite images with a resolution of about 16 meters can be used to analyze the land use types and vegetation coverage in the area to assist in field block segmentation.

[0045] For example, the first satellite image data is the 16-meter data image of the Gaofen-1 (GF1) satellite. The Gaofen-1 satellite is an important part of the major project of China's high-resolution earth observation system. The image data it obtains has a relatively high spatial resolution. A 16-meter resolution means that each pixel on the image represents a range of 16 meters × 16 meters on the ground. This resolution data can clearly distinguish larger-scale ground object features such as land use types and crop planting areas, providing basic remote sensing data support for the extraction of crop distribution.

[0046] Obtain on-site survey data of the target area through manual on-site inspections. On-site survey personnel can go to the site of the target area to conduct on-site observations, records, and samplings of the crop planting situation. For example, record information such as the types of crops planted, growth conditions, and planting densities in different plots. These data can provide accurate reference bases for the extraction of crop distribution data, making up for the deficiencies of satellite image data in crop identification. It is also possible to conduct detailed investigations on some representative plots and even collect soil samples, plant samples, etc. For example, during on-site surveys, survey personnel can record the coordinates of different crop planting areas through GPS positioning devices. On-site survey data can provide accurate reference information for the interpretation of remote sensing images, helping to improve the accuracy of crop distribution extraction.

[0047] Obtain satellite meteorological data of the target area. These data usually include meteorological elements such as temperature, precipitation, humidity, and wind speed. These meteorological data have an important impact on the occurrence and spread of pests and diseases and are important input data for the meteorological index evaluation model of pests and diseases. For example, the combined conditions of temperature and humidity may affect the hatching and growth rates of certain pests and diseases.

[0048] S22. Input the first satellite image data into the supervised classification model, and output the distribution results of each crop in the target area through the supervised classification model.

[0049] The basic principle of supervised classification is to use known training samples (i.e., the crop distribution areas determined in on-site survey data) to train the classifier, enabling the classifier to learn information such as the spectral characteristics of different crops on remote sensing images. The specific steps are as follows: 1) Sample selection: According to on-site survey data, select typical areas of different crops on the remote sensing image as training samples. The types and distributions of crops in these sample areas are known, and they should cover as many crop types and different growth conditions and environmental conditions in the target area as possible.

[0050] 2) Feature extraction: Extract features related to crops from remote sensing images, such as spectral features (reflectance in different bands, etc.), texture features (texture patterns of ground objects in the image), etc. These features are the basis for the classifier to determine which crop a pixel belongs to.

[0051] 3) Model training: Use the selected training samples and the extracted features to train the classifier. Common classifiers include the maximum likelihood classifier, support vector machine classifier, etc. The classifier will learn the feature patterns of different crops based on the training samples and establish a classification model.

[0052] 4) Model application: Apply the trained classification model to the entire remote sensing image, classify each pixel in the image, and determine which crop it belongs to. Finally, a classification result map of crop distribution is obtained.

[0053] S23. Correct and supplement the distribution results of each crop in the target area based on the field survey data to obtain the target crop distribution data of the target area.

[0054] Combine the 16-meter data image of the GF-1 satellite and the field survey data. Remote sensing images provide information on the distribution of crops over a large area, but there may be certain errors; although the field survey data is accurate, the coverage is limited. Through data fusion, the field survey data can be used to correct and supplement the interpretation results of remote sensing images, improving the accuracy and reliability of crop distribution extraction. Finally, the target crop distribution data of the target area is obtained.

[0055] S24. Perform data fusion on the second satellite image data and the target crop distribution data to obtain fusion data.

[0056] Perform data fusion on the second satellite image data and the target crop distribution data. For example, overlay the crop distribution data on the satellite image so that the crop distribution information can be used as a reference during image analysis.

[0057] S25. Based on the data characteristics of the second satellite image data, identify the field boundary information in the fusion data through edge detection methods.

[0058] In the embodiments of the present invention, the satellite image can also be preprocessed, including radiometric correction, geometric correction, atmospheric correction, etc., to improve the quality and usability of the image. Furthermore, using information such as the spectral features and texture features of the satellite image, combined with the target crop distribution data, the boundary information of the field is identified.

[0059] For example, edge detection algorithms can be adopted: traditional methods such as Canny edge detection and Sobel operator, or edge detection methods based on deep learning (such as HED), to extract the boundary lines of the fields from satellite images. Another example is the segmentation algorithm: region-based segmentation methods (such as region growing and watershed algorithm) or graph theory-based segmentation methods (such as normalized cut) to divide the image into different fields.

[0060] S26. Match the field boundary information with the distribution data of each type of crop to obtain the field segmentation results of multiple study areas corresponding to each type of crop.

[0061] According to the distribution data of each type of crop, match the identified field boundaries with the crop distribution areas. For example, if a certain field area is marked as a certain type of crop in the target crop distribution data, then divide this field into the plot corresponding to this crop.

[0062] S27. Convert the field segmentation results into vector data to obtain the plot vector data of multiple study areas corresponding to each type of crop.

[0063] After the field segmentation is completed, convert the segmentation results into vector data format. Vector data represents the location and shape of geographical entities in the form of points, lines, and surfaces, with higher precision and operability. The fields corresponding to each type of crop are represented as vectorized polygons, and each polygon contains geographical coordinate information (such as longitude and latitude), crop type attributes, etc. These vector data can be stored in common geographical information system (GIS) data formats, such as Shapefile, GeoJSON, etc., for subsequent spatial analysis and applications.

[0064] S28. Extract the second pest and disease meteorological data from the satellite meteorological data.

[0065] In the embodiments of the present invention, first, a pest and disease meteorological index evaluation model needs to be constructed. Specifically, it includes: obtaining the third satellite image data, the fourth satellite image data, and the fifth satellite image data with a target time resolution of the target area. The third satellite image data and the fourth satellite image data are remote sensing data obtained by multi-source satellites with different resolutions; extracting the first pest and disease meteorological data from the fifth satellite image data and interpolating the meteorological data into the target pest and disease meteorological data with the target resolution. The target pest and disease meteorological data includes air temperature, air pressure, specific humidity, precipitation, and wind speed; constructing a pest and disease meteorological index evaluation model based on the target pest and disease meteorological data, the third satellite image data, and the fourth satellite image data. Among them, the first pest and disease meteorological data includes but is not limited to key pest and disease meteorological data such as air temperature, air pressure, specific humidity, precipitation, and wind speed.

[0066] Furthermore, extract the second pest meteorological data from satellite meteorological data. Among them, the second pest meteorological data includes, but is not limited to, key pest meteorological data such as air temperature, air pressure, specific humidity, precipitation, and wind speed.

[0067] For example, using SMAP and TRMM satellite image data with high temporal resolution (3 hours), extract key pest meteorological data such as air temperature, air pressure, specific humidity, precipitation, and wind speed in the study area near real-time, and interpolate the key meteorological data to a 10-meter resolution using the Krige interpolation method.

[0068] High temporal resolution means that the revisit period of satellite image data is short, and image data of the same area can be obtained within a short time (such as every day or every 3 hours).

[0069] SMAP (Soil Moisture Active Passive) satellite: Covers the globe once every 2 - 3 days, providing soil moisture and meteorological data.

[0070] TRMM (Tropical Rainfall Measuring Mission) satellite: Provides precipitation data every 3 hours.

[0071] Function: High temporal resolution data can capture the rapid changes in meteorological conditions and provide near real-time dynamic information for pest monitoring.

[0072] High temporal resolution satellite image data usually contains direct or indirect meteorological information, but certain data processing is required to extract key pest meteorological data (such as air temperature, air pressure, specific humidity, precipitation, and wind speed).

[0073] Data source and extraction method: SMAP satellite: Direct data: Soil moisture, surface temperature.

[0074] Indirect data: Through model inversion, meteorological parameters such as air temperature and specific humidity can be extracted.

[0075] TRMM satellite: Direct data: Precipitation rate.

[0076] Indirect data: Through fusion with other data, parameters such as wind speed and air pressure can be extracted.

[0077] Data processing steps: Data download: Download high temporal resolution SMAP and TRMM data from the satellite data platform.

[0078] Data preprocessing: Cloud removal processing: Remove the interference of clouds on meteorological data.

[0079] Resampling: Unify the data to the same spatial resolution.

[0080] Meteorological parameters extraction: Air temperature: extracted from SMAP surface temperature data or inverted through thermal infrared bands.

[0081] Air pressure: obtained by fusing meteorological models (such as ERA5) with satellite data.

[0082] Specific humidity: Inverted from soil moisture data from SMAP or calculated by meteorological models.

[0083] Precipitation: Extracted directly from TRMM precipitation rate data.

[0084] Interpolation to 10-meter resolution: Convert low-resolution satellite meteorological data (such as SMAP's 36 kilometers and TRMM's 5 kilometers) to 10-meter resolution data through spatial interpolation methods to match it with high-resolution plot segmentation data (such as 0.5 meters). Improve the spatial accuracy of meteorological data so that it can be integrated with high-resolution crop distribution data and plot segmentation data to support refined pest and disease monitoring.

[0085] Kriging interpolation is an interpolation method based on spatial autocorrelation, which can take into account the spatial distribution characteristics of data and generate high-precision interpolation results.

[0086] S29. Calculate the daily contribution value of the second pest and disease meteorological data to crop infection within a period based on the second pest and disease meteorological data.

[0087] The key meteorological data are substituted into the meteorological index evaluation model for pests and diseases, and the pest and disease incidence index is generated for each ten-day period. The meteorological index evaluation model is a percentage model, which specifically includes substituting the key meteorological data into the formula for the daily contribution value of meteorological data caused by disease, which is established using mathematical statistics methods, during the period when crops are susceptible to disease, with a ten-day period as a cycle, to calculate the daily contribution value of the key meteorological data to crop disease within the period.

[0088] The formula for the contribution value of meteorological data to disease-causing days is used to quantify the contribution of daily key meteorological data (such as temperature, humidity, precipitation, etc.) to crop disease infection.

[0089] Input: Daily meteorological data (such as temperature, humidity, precipitation, etc.).

[0090] Output: Contribution value of daily meteorological data to crop disease infection (usually a value between 0 and 1, indicating the degree of contribution).

[0091] The formula for the contribution of meteorological data to disease-causing days is usually established based on mathematical statistics methods (such as regression analysis, logistic regression, exponential function, etc.). For example, Formula 1: Formula 1 Wherein, is the contribution value of meteorological data to the disease-causing day on the d th day; is the temperature on the d th day; is the humidity on the d th day; is the precipitation on the d th day; is the conversion function of meteorological data (such as exponential function, piecewise function), which is used to map meteorological data to the contribution value; is the weight coefficient, indicating the relative importance of each meteorological factor to the crop disease.

[0092] S210. Calculate the dekad contribution mean value based on the daily contribution value.

[0093] After summarizing the daily contribution values of the pest and disease meteorological data to the crop disease, calculate the average value within each dekad (10 days) to obtain the dekad contribution mean value.

[0094] S211. Calculate the dekad pest and disease incidence index of the target area based on the dekad contribution mean value using the comprehensive meteorological data evaluation method.

[0095] Use the comprehensive meteorological data evaluation formula to generate the pest and disease incidence index within the period, and then calculate the average value of the pest and disease occurrence index within the monitoring period based on the pest and disease incidence indices of each period.

[0096] The comprehensive meteorological data evaluation formula is used to comprehensively evaluate the daily contribution value of meteorological data to the disease-causing day and generate the pest and disease incidence index within the period (such as each dekad).

[0097] Input: The daily contribution value of meteorological data to the disease-causing day within the period.

[0098] Output: The pest and disease incidence index within the period (usually a value between 0 and 100, indicating the disease risk).

[0099] The comprehensive meteorological data evaluation formula is usually calculated based on the mean method or the weighted average method. Such as Formula 2: Formula 2 Wherein, is the pest and disease incidence index within the p period; is the contribution value of meteorological data to the disease-causing day on the d th day; n is the number of days within the period (such as 10 days per dekad).

[0100] S212. Set a mask layer based on the vector data of the plot.

[0101] S213. Perform a masking operation on the disease and pest incidence index for each ten-day period and the masking layer to obtain the target disease and pest incidence index for each plot in each ten-day period.

[0102] The masking operation is an important data processing step in the embodiments of the present invention, which is used to extract relevant information of specific regions or specific crops from global data. By using a binary or multi-valued masking layer, the target data is screened or filtered to extract the data of the region of interest (such as the planting area of specific crops). The masking layer is usually a raster or vector data with the same spatial resolution as the target data, where the pixel values of the region of interest are 1 (or True), and the other regions are 0 (or False). The result of the masking operation only retains the valid data of the crop planting area and excludes the interference of non-farmland areas.

[0103] The specific steps are as follows: 1) Prepare the masking layer: Input data: Crop distribution data (usually in raster or vector format).

[0104] Generate the masking layer: If it is raster data, set the pixel values of the crop planting area to 1 and the other areas to 0; if it is vector data, convert it to raster data with the same resolution as the target data (such as the disease and pest incidence index).

[0105] 2) Perform the masking operation; Input data: Disease and pest incidence index data (usually in raster format).

[0106] Masking operation: Multiply the disease and pest incidence index data and the masking layer pixel by pixel.

[0107] At the pixel positions where the value in the masking layer is 1, retain the original value of the disease and pest incidence index.

[0108] At the pixel positions where the value in the masking layer is 0, set the disease and pest incidence index to an invalid value (such as NaN).

[0109] 3) Output the result: Output data: The disease and pest incidence index data after the masking operation, which only contains the valid values in the crop planting area.

[0110] Calculate the target disease and pest incidence index: After the masking operation, the target disease and pest incidence index for each plot in each ten-day period is obtained. These incidence indices can be used to analyze the disease and pest occurrence risks of different fields at different time periods, providing a scientific basis for the monitoring and control of diseases and pests. For example, if the disease and pest incidence index of a certain field in a certain ten-day period is high, it indicates that the risk of disease and pest occurrence in this field during this time period is relatively large, and timely control measures need to be taken.

[0111] Furthermore, the target disease and insect pest incidence index of each plot in each decade is attached as an attribute to the attribute table of the corresponding plot vector data to obtain the disease and insect pest incidence index of each plot in each decade, so as to meet the refined management and monitoring of plots proposed by modern agriculture.

[0112] The embodiment of the present invention ensures real-time, accuracy and spatial continuity through multi-scale remote sensing satellite image data extraction; SMAP and TRMM satellite data are free, and there is no need to purchase related instruments to obtain meteorological data, thereby reducing costs.

[0113] The following is an example of winter wheat planting in Hanchuan City to illustrate this method: Using the 16-meter image data of the Gaofen-1 (GF1) satellite and field survey data, the supervised classification method was used to extract the winter wheat planting distribution data of Hanchuan City. Then, the 0.5-meter data of Google Images was used to extract the field segmentation data of Yanglingou Town, Hanchuan City by relying on the edge detection method of deep learning. The high-temporal resolution (3-hour) SMAP and TRMM satellite image data were used to extract the key meteorological element data of pests and diseases in Hanchuan City, such as temperature, air pressure, specific humidity, precipitation and wind speed, in near real time. The key meteorological data were interpolated to a resolution of 10 meters using the Kriging interpolation method. Then, the key meteorological elements were substituted into the pest and disease meteorological index evaluation model to generate the winter wheat stripe rust occurrence index map of Hanchuan City for each decade. The winter wheat planting distribution data of Hanchuan City was used to perform masking operation on the winter wheat stripe rust occurrence index of each decade for feature matching. The winter wheat stripe rust occurrence index of each decade after the masking operation was attached as an attribute to the plot vector data attribute table to realize the refined management and monitoring of farmland.

[0114] Compared with the prior art, the present invention has the following main advantages: 1. Improving classification accuracy: by fusing multi-source satellite remote sensing data and combining deep learning feature extraction and classification algorithms, the accuracy of crop disease and insect pest classification is significantly improved. 2. Enhancing generalization ability: the constructed disease and insect pest classification model has strong generalization ability and can be applied to the classification of diseases and insect pests in different regions and crops. 3. Improving monitoring efficiency: realizing real-time monitoring and early warning of crop diseases and insect pests, improving monitoring efficiency and reducing monitoring costs. 4. Promoting the development of smart agriculture: providing important technical support for smart agriculture and promoting the intelligent and precise development of agricultural production.

[0115] The crop disease and pest monitoring device integrating multi-source satellite remote sensing data provided by the present invention is described below. The crop disease and pest monitoring device integrating multi-source satellite remote sensing data described below and the crop disease and pest monitoring method integrating multi-source satellite remote sensing data described above can be referenced to each other.

[0116] Figure 4This is a schematic structural diagram of a crop pest and disease monitoring device that integrates multi-source satellite remote sensing data, specifically including: An acquisition module 401, configured to acquire first satellite image data, second satellite image data, field survey data, and satellite meteorological data of a target area. The first satellite image data and the second satellite image data are remote sensing data obtained by multi-source satellites with different resolutions, and the resolution of the first satellite image data is less than that of the second satellite image data. For detailed descriptions, refer to the relevant descriptions corresponding to the above method embodiments, which will not be elaborated here.

[0117] An extraction module 402, configured to extract target crop distribution data of the target area according to the first satellite image data and the field survey data. For detailed descriptions, refer to the relevant descriptions corresponding to the above method embodiments, which will not be elaborated here.

[0118] A segmentation module 403, configured to perform field block segmentation according to the second satellite image data and the target crop distribution data by means of edge detection to obtain plot vector data of multiple study areas corresponding to each crop. For detailed descriptions, refer to the relevant descriptions corresponding to the above method embodiments, which will not be elaborated here.

[0119] A calculation module 404, configured to input the satellite meteorological data into a pest and disease meteorological index evaluation model, and calculate the decadal pest and disease incidence index of the target area through the pest and disease meteorological index evaluation model. For detailed descriptions, refer to the relevant descriptions corresponding to the above method embodiments, which will not be elaborated here.

[0120] A processing module 405, configured to perform a masking operation on the decadal pest and disease incidence index based on the plot vector data to obtain the target pest and disease incidence index of each plot for each decade. For detailed descriptions, refer to the relevant descriptions corresponding to the above method embodiments, which will not be elaborated here.

[0121] Figure 5 Illustrates a schematic physical structure diagram of an electronic device, such as Figure 5As shown in the figure, the electronic device may include: a processor 810, a communications interface 820, a memory 830, and a communication bus 840. Among them, the processor 810, the communications interface 820, and the memory 830 complete communication with each other through the communication bus 840. The processor 810 may call the logical instructions in the memory 830 to execute a method for monitoring crop pests and diseases by fusing multi-source satellite remote sensing data. The method includes: obtaining first satellite image data, second satellite image data, field survey data, and satellite meteorological data of a target area. The first satellite image data and the second satellite image data are remote sensing data obtained by multi-source satellites with different resolutions, and the resolution of the first satellite image data is less than that of the second satellite image data; extracting target crop distribution data of the target area according to the first satellite image data and the field survey data; performing field block segmentation according to the second satellite image data and the target crop distribution data by an edge detection method to obtain plot vector data of multiple study areas corresponding to each crop; inputting the satellite meteorological data into a pest and disease meteorological index evaluation model, and calculating the decadal pest and disease incidence index of the target area through the pest and disease meteorological index evaluation model; performing a masking operation on the decadal pest and disease incidence index based on the plot vector data to obtain the target pest and disease incidence index of each plot for each decade.

[0122] In addition, when the logical instructions in the above-mentioned memory 830 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0123] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the method for monitoring crop pests and diseases by fusing multi-source satellite remote sensing data provided by the above-mentioned various methods. The method includes: obtaining first satellite image data, second satellite image data, field survey data, and satellite meteorological data of a target area. The first satellite image data and the second satellite image data are remote sensing data obtained by multi-source satellites with different resolutions, and the resolution of the first satellite image data is less than that of the second satellite image data; extracting target crop distribution data of the target area according to the first satellite image data and the field survey data; performing field block segmentation according to the second satellite image data and the target crop distribution data by an edge detection method to obtain plot vector data of multiple study areas corresponding to each crop; inputting the satellite meteorological data into a pest and disease meteorological index evaluation model, and calculating the decadal pest and disease incidence indexes of the target area through the pest and disease meteorological index evaluation model; performing a masking operation on the decadal pest and disease incidence indexes based on the plot vector data to obtain the target pest and disease incidence indexes of each plot in each decade.

[0124] On another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it realizes the method for monitoring crop pests and diseases by fusing multi-source satellite remote sensing data provided by the above-mentioned various methods. The method includes: obtaining first satellite image data, second satellite image data, field survey data, and satellite meteorological data of a target area. The first satellite image data and the second satellite image data are remote sensing data obtained by multi-source satellites with different resolutions, and the resolution of the first satellite image data is less than that of the second satellite image data; extracting target crop distribution data of the target area according to the first satellite image data and the field survey data; performing field block segmentation according to the second satellite image data and the target crop distribution data by an edge detection method to obtain plot vector data of multiple study areas corresponding to each crop; inputting the satellite meteorological data into a pest and disease meteorological index evaluation model, and calculating the decadal pest and disease incidence indexes of the target area through the pest and disease meteorological index evaluation model; performing a masking operation on the decadal pest and disease incidence indexes based on the plot vector data to obtain the target pest and disease incidence indexes of each plot in each decade.

[0125] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative work.

[0126] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0127] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features. And these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A crop pest and disease monitoring method integrating multi-source satellite remote sensing data, characterized in that: include: Acquire first satellite image data, second satellite image data, field survey data and satellite meteorological data of a target area, wherein the first satellite image data and the second satellite image data are remote sensing data obtained by multi-source satellites with different resolutions, and the resolution of the first satellite image data is smaller than that of the second satellite image data; Extracting target crop distribution data of the target area according to the first satellite image data and the field survey data; By using an edge detection method, field segmentation is performed according to the second satellite image data and the target crop distribution data to obtain plot vector data of multiple research areas corresponding to each crop; Inputting the satellite meteorological data into a pest and disease meteorological index evaluation model, and calculating the pest and disease incidence index of the target area by ten days through the pest and disease meteorological index evaluation model; The ten-day pest and disease incidence index is subjected to masking operation based on the plot vector data to obtain the target pest and disease incidence index of each plot in each ten-day period.

2. The method according to claim 1, characterized in that: The step of extracting target crop distribution data of the target area according to the first satellite image data and the field survey data includes: Inputting the first satellite image data into a supervised classification model, and outputting the distribution result of each crop in the target area through the supervised classification model; Correcting and supplementing the distribution results of each crop in the target area based on the field survey data to obtain target crop distribution data of the target area; The supervised classification model is trained by the following steps: Obtaining sample survey data and sample satellite image data of the target area; Selecting a target area for each crop on the sample satellite image data as a training sample according to the sample survey data; extracting target features of each crop from the sample satellite image data; A supervised classification model is trained based on the target features and the training samples.

3. The method according to claim 2, characterized in that The edge detection method is used to segment the field according to the second satellite image data and the target crop distribution data to obtain the field vector data of multiple research areas corresponding to each crop, including: fusing the second satellite image data with the target crop distribution data to obtain fused data; Based on the data features of the second satellite image data, identifying the field boundary information in the fused data by an edge detection method; Matching the field boundary information with the distribution data of each crop to obtain field segmentation results of multiple research areas corresponding to each crop; The field segmentation result is converted into vector data to obtain the field vector data of multiple research areas corresponding to each crop.

4. The method according to claim 1, characterized in that The method further comprises: Acquire third satellite image data, fourth satellite image data and fifth satellite image data of the target area at a target time resolution, wherein the third satellite image data and the fourth satellite image data are remote sensing data obtained by multi-source satellites with different resolutions; Extracting first pest and disease meteorological data from the fifth satellite image data, and interpolating the meteorological data into target pest and disease meteorological data with a target resolution, wherein the target pest and disease meteorological data includes air temperature, air pressure, specific humidity, precipitation and wind speed; A pest and disease meteorological index evaluation model is constructed based on the target pest and disease meteorological data, the third satellite image data and the fourth satellite image data.

5. The method according to claim 3, characterized in that: The step of inputting the satellite meteorological data into a pest and disease meteorological index evaluation model, and calculating the pest and disease incidence index of the target area by ten days through the pest and disease meteorological index evaluation model, comprises: Extracting second pest and disease meteorological data from the satellite meteorological data; Calculate the daily contribution value of the second pest and disease meteorological data to crop infection within a period based on the second pest and disease meteorological data; The ten-day contribution average is calculated based on the daily contribution value; Based on the ten-day contribution mean, a comprehensive evaluation method of meteorological data is used to calculate the ten-day pest and disease incidence index of the target area.

6. The method according to claim 5, characterized in that The masking operation is performed on the pest and disease incidence index of each decade based on the plot vector data to obtain the target pest and disease incidence index of each plot in each decade, including: Setting a mask layer based on the plot vector data; The ten-day disease and insect pest incidence index and the mask layer are subjected to mask operation processing to obtain the target disease and insect pest incidence index of each plot in each ten-day period.

7. The method according to claim 6, characterized in that The method further comprises: The target disease and insect pest incidence index of each plot in each decade is attached as an attribute to the attribute table of the corresponding plot vector data.

8. A crop pest monitoring device integrating multi-source satellite remote sensing data, characterized in that: include: an acquisition module, configured to acquire first satellite image data, second satellite image data, field survey data, and satellite meteorological data of a target area, wherein the first satellite image data and the second satellite image data are remote sensing data obtained by multi-source satellites with different resolutions, and the resolution of the first satellite image data is smaller than that of the second satellite image data; An extraction module, used for extracting target crop distribution data of the target area according to the first satellite image data and the field survey data; A segmentation module, configured to segment the field according to the second satellite image data and the target crop distribution data by an edge detection method, and obtain field vector data of multiple research areas corresponding to each crop; A calculation module, used for inputting the satellite meteorological data into a pest and disease meteorological index evaluation model, and calculating the pest and disease incidence index of each decade in the target area through the pest and disease meteorological index evaluation model; A processing module is used to perform mask operation processing on the disease and insect pest incidence index of each decade based on the plot vector data to obtain the target disease and insect pest incidence index of each plot in each decade.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the crop disease and insect pest monitoring method integrating multi-source satellite remote sensing data as described in any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the crop disease and insect pest monitoring method integrating multi-source satellite remote sensing data as described in any one of claims 1 to 7 is implemented.

Images