Agricultural insurance block right confirmation method based on satellite remote sensing image and GEOSAM model

Abstract

The application provides an agricultural insurance insurable plot right confirmation method based on satellite remote sensing images and a GEOSAM model, comprising the following steps: a front-end interactive interface is used to take a screenshot of a map area where user dotting coordinates are located, and four-to-coordinate of the map screenshot is obtained; the user dotting coordinates, the map screenshot and the four-to-coordinate are sent to a back-end service as front-end request parameters; after the back-end service receives the front-end request parameters, a GeoTIFF image is generated, and the GeoTIFF image and the user dotting coordinates are input into the GEOSAM model for plot recognition to obtain an initial plot contour vector; the initial plot contour vector is converted into vector surface data, a geometric simplification process is carried out on the recognized plot polygon based on a Douglas-Pok algorithm, and the vector surface data is returned to the front-end interactive interface after coordinate system conversion for drawing display for user confirmation; finally, the vector surface data confirmed by the user is superimposed and analyzed with built-in crop remote sensing interpretation results, and standardized agricultural insurance insurable plot data is output.

Description

Technical Field

[0001] This invention relates to the field of agricultural insurance technology innovation and technical services, specifically to a method for confirming the ownership of agricultural insurance insured land parcels based on satellite remote sensing imagery and the GEOSAM model. Background Technology

[0002] Agricultural insurance, as an important means of preventing agricultural production risks and protecting farmers' income, faces a core challenge in its underwriting process: the accuracy and efficiency of land ownership confirmation and area verification. Existing technological solutions are often disconnected from actual business processes, making direct application difficult. Currently, the mapping of insured agricultural land plots mainly relies on manual drawing or traditional remote sensing segmentation methods, which have numerous technical shortcomings. Furthermore, the lack of a robust front-end and back-end collaborative architecture makes it difficult to meet the demands of large-scale, precise agricultural insurance operations.

[0003] Manual delineation relies on on-site surveys by staff or manual outlining based on low-resolution remote sensing images. This method is not only extremely inefficient but also susceptible to human error, resulting in poor accuracy of plot boundaries and making it difficult to meet the underwriting needs of large, scattered farmlands. Furthermore, manual delineation cannot accurately distinguish between planted and non-planted areas (such as roads, ditches, bare land, and wasteland) within a plot, which can easily lead to the inclusion of non-planted areas in the insured plot, thus artificially inflating the insured area. This can trigger risks such as fraudulent insurance applications and increase the operating costs of insurance companies.

[0004] Traditional remote sensing segmentation methods (such as threshold segmentation, edge detection, and traditional machine learning segmentation) are affected by factors such as remote sensing image resolution, spectral interference, and terrain differences. As a result, the accuracy of plot segmentation is low, it is difficult to accurately extract the subtle boundaries of farmland plots, and it is impossible to achieve linkage matching between plot segmentation and crop planting range. Manual intervention is still required for cropping and correction, which fails to fundamentally solve the dual problems of efficiency and accuracy.

[0005] CN121789068A provides an interactive method for extracting farmland boundaries for farmers based on satellite remote sensing and artificial intelligence. It addresses the need for front-end and back-end collaboration in practical operations by employing a front-end interaction model that calls AI algorithms to the back-end model. Ordinary staff can achieve professional results in a field setting through quick and simple interactive operations, similar to traditional methods. However, this approach requires not only satellite remote sensing images but also elevation and soil data, and even requires farmers to collect simple on-site photos and location data on mobile devices. Furthermore, it is disconnected from business processes and difficult to directly implement in agricultural insurance operations.

[0006] In order to solve the above problems, people have been seeking an ideal technological solution. Summary of the Invention

[0007] Therefore, it is necessary to provide a method for confirming the ownership of agricultural insurance insured land plots based on satellite remote sensing imagery and the GEOSAM model to address the aforementioned technical issues. By constructing a complete business chain of front-end interaction, back-end processing, model recognition, manual confirmation, and precise cropping, this method can meet the practical needs of accurate underwriting and efficient risk control in agricultural insurance.

[0008] To achieve the above objectives, the first aspect of the present invention provides a method for confirming the ownership of agricultural insurance-covered land parcels based on satellite remote sensing imagery and a GEOSAM model, comprising the following steps:

[0009] The front-end interactive interface loads and displays satellite remote sensing image data. When it detects that the user has made a checkpoint operation in the target insured area, it obtains the user's checkpoint coordinates and takes a screenshot of the map area where the user's checkpoint coordinates are located. At the same time, it obtains the four boundary coordinates of the map screenshot and sends the user's checkpoint coordinates, map screenshot and four boundary coordinates as front-end request parameters to the back-end service.

[0010] After receiving the request parameters from the front end, the backend service converts the map screenshot and the four boundary coordinates from the WGS84 geographic coordinate system to the Web Mercator projection coordinate system to obtain the GeoTIFF image. The GeoTIFF image and the user's marked coordinates are then input into the GEOSAM model for land parcel recognition to obtain the initial land parcel outline vector.

[0011] The backend service converts the initial plot outline vector into vector surface data, dynamically calculates the Douglas-Puk algorithm simplification tolerance by weighted fusion of global image scale and individual plot scale, and uses this tolerance to perform geometric simplification processing on the identified plot polygons, removes redundant boundary vertices, and fully preserves the overall shape of the plot, key boundary inflection points and core spatial features.

[0012] The backend service converts the vector surface coordinates of the geometrically simplified land polygons from the Web Mercator projection coordinate system to the WGS84 geographic coordinate system and returns them to the frontend interface.

[0013] The front-end interactive interface draws vector surface data on the map interface and displays it after overlaying it with satellite remote sensing image data for user confirmation;

[0014] The front-end interactive interface also overlays and analyzes the vector surface data confirmed by the user with the built-in remote sensing interpretation results of crops, retaining only the part of the vector surface data that overlaps with the corresponding crop interpretation area, and outputs standardized agricultural insurance insured land data for use in agricultural insurance filing and area calculation.

[0015] The above solution combines the front-end interactive interface, back-end services, GEOSAM model with multi-source remote sensing imagery and crop remote sensing interpretation to construct a complete business chain of front-end interaction, back-end processing, model recognition, manual confirmation, and precise cropping, which can meet the actual implementation needs of precise underwriting and efficient risk control in agricultural insurance.

[0016] To achieve the above objectives, a second aspect of the present invention provides a land title confirmation system for agricultural insurance coverage based on satellite remote sensing imagery and a GEOSAM model, comprising:

[0017] The front-end interactive interface is used to load and display satellite remote sensing image data. When it detects a user's marking operation in the target insured area, it obtains the user's marking coordinates, takes a screenshot of the map area where the user's marking coordinates are located, and obtains the four boundary coordinates of the map screenshot. The user's marking coordinates, map screenshot, and four boundary coordinates are sent as front-end request parameters to the back-end service. It also draws the received vector surface data on the map interface and displays it after overlaying it with the satellite remote sensing image data. It is also used to overlay and analyze the vector surface data with the built-in crop remote sensing interpretation results, retaining only the part of the vector surface data that overlaps with the corresponding crop interpretation area, and outputting standardized agricultural insurance insured plot data for agricultural insurance filing and area calculation.

[0018] The backend service, upon receiving request parameters from the frontend, converts the map screenshot and boundary coordinates from the WGS84 geographic coordinate system to the Web Mercator projection coordinate system to obtain GeoTIFF imagery. It then inputs the GeoTIFF imagery and user-marked coordinates into the GEOSAM model for land parcel identification, obtaining initial land parcel outline vectors. Furthermore, it converts the initial land parcel outline vectors into vector polygon data, dynamically calculates the Douglas-Puk algorithm simplification tolerance using a weighted fusion method combining global image scale and individual land parcel scale, and utilizes this tolerance to perform geometric simplification processing on the identified land parcel polygons, removing redundant boundary vertices and fully preserving the overall shape, key boundary inflection points, and core spatial features of the land parcels. Finally, it converts the vector polygon coordinates of the geometrically simplified land parcels from the Web Mercator projection coordinate system to the WGS84 geographic coordinate system and returns them to the frontend interactive interface for display.

[0019] To achieve the above objectives, a third aspect of the present invention provides a terminal device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; the memory is used to store computer programs; and the processor is used to execute the program stored in the memory to implement the steps of the method for confirming the ownership of agricultural insurance land parcels based on satellite remote sensing images and the GEOSAM model as described in the first aspect.

[0020] To achieve the above objectives, a fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the method for confirming the ownership of agricultural insurance-covered land parcels based on satellite remote sensing imagery and the GEOSAM model as described in the first aspect.

[0021] To achieve the above objectives, the fifth aspect of the present invention provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the method for confirming the ownership of agricultural insurance-covered land parcels based on satellite remote sensing imagery and the GEOSAM model as described in the first aspect.

[0022] The beneficial effects of this invention are as follows:

[0023] This invention is fully aligned with actual workflows, built upon a front-end architecture and back-end services. The front-end loads multi-source satellite remote sensing images, and users trigger screenshots by marking points and send processing requests to the back-end. The back-end then performs coordinate transformation, GEOSAM model recognition, and vector filtering, displaying the recognition results on the front-end for user confirmation. Finally, it combines the built-in crop interpretation results to complete cropping and output standardized insured land data, improving the efficiency and accuracy of agricultural insurance land filing, preventing the risk of false insurance, and achieving seamless integration between technical solutions and actual business operations. Attached Figure Description

[0024] Figure 1 This is a flowchart illustrating the method for confirming the ownership of agricultural insurance-covered land parcels as described in this invention. Detailed Implementation

[0025] This embodiment uses the confirmation of agricultural insurance insured land parcels in a wheat-growing area of ​​a certain county as an application scenario, and discloses a method for confirming agricultural insurance insured land parcels based on satellite remote sensing imagery and a GEOSAM model. It fully conforms to the actual workflow, and is built on a Vue+Leaflet front-end architecture and a Python+Flask back-end service. The front-end loads multi-source satellite remote sensing imagery, and the user triggers a screenshot by marking points and sends a processing request to the back-end. The back-end is called to complete coordinate transformation, GEOSAM model recognition, and vector filtering. After that, the recognition results are displayed on the front-end for the user to confirm. Finally, the cropping is completed by combining the built-in crop interpretation results, and standardized insured land parcel data is output, which improves the efficiency and accuracy of agricultural insurance land parcel filing, prevents the risk of false insurance, and achieves seamless integration between technical solutions and actual business.

[0026] The technical solution of the present invention will be further described in detail below through specific embodiments.

[0027] Example 1

[0028] This embodiment provides a method for confirming the ownership of agricultural insurance-covered land parcels based on satellite remote sensing imagery and the GEOSAM model, such as... Figure 1As shown, it includes the following steps:

[0029] Step 1: The front-end interactive interface loads and displays satellite remote sensing image data. When the user's point-marking operation is detected in the target insured area, the coordinates of the user's point-marking are obtained, and a screenshot of the map area where the user's point-marking coordinates are located is taken. At the same time, the four-way coordinates of the map screenshot are obtained. The user's point-marking coordinates, map screenshot, and four-way coordinates are sent to the back-end service as front-end request parameters.

[0030] Specifically, in this embodiment, the front-end page is built based on the Vue framework, integrating the Leaflet map component. A multi-source satellite imagery data source is configured through the Leaflet plugin. This multi-source satellite imagery data source includes at least one of high-resolution satellite remote sensing imagery and UAV remote sensing imagery. For example, the spatial resolution of the Gaofen-2 satellite remote sensing imagery is 1m; the spatial resolution of the UAV remote sensing imagery is 0.1m. Users can select different satellite imagery data sources according to their needs using the image switching button on the front-end page and load them into the map interface.

[0031] In the map interface, users locate the agricultural insurance coverage target plot they want to draw and manually mark a point within the plot. Note that at least one point is required. After marking the point, the Leaflet component automatically triggers the screenshot function, capturing a preset area around the user's marked point, such as a 500m x 500m area. The screenshot area can be adjusted according to actual needs. Note that the screenshot resolution matches the current map zoom level. For example, at zoom level 18, the screenshot resolution is 0.5m / pixel.

[0032] At the same time, the Leaflet component automatically obtains the four-boundary coordinates of the screenshot. The four-boundary coordinates are the extreme values ​​of latitude and longitude in the WGS84 geographic coordinate system, in the format [minLng, minLat, maxLng, maxLat], for example [116.3215, 39.9218, 116.3275, 39.9278].

[0033] The front-end interactive interface initiates HTTP protocol front-end requests through the Axios plugin to call the back-end service interface. The request parameters are encapsulated in JSON format, where the user's point coordinates are latitude and longitude coordinates in the WGS84 geographic coordinate system, and the map screenshot format is PNG or JPG.

[0034] The following is an example of a frontend request:

[0035] User-marked coordinates (WGS84 geographic coordinate system [{"lng":116.3245, "lat":39.9248}]), map screenshot (PNG format, Base64 encoded), map screenshot boundary coordinates ([116.3215, 39.9218, 116.3275,39.9278]).

[0036] Step 2: After receiving the request parameters from the front end, the backend service converts the map screenshot and the four boundary coordinates from the WGS84 geographic coordinate system to the Web Mercator projection coordinate system to obtain the GeoTIFF image. The GeoTIFF image and the user-marked coordinates are then input into the GEOSAM model for land parcel recognition to obtain the initial land parcel outline vector.

[0037] Specifically, the backend service is developed based on Python and Flask and deployed on a cloud server to listen to the corresponding interface port in order to receive frontend requests and parse parameters.

[0038] Specifically, the map screenshot and its boundary coordinates are converted from the WGS84 geographic coordinate system to the Web Mercator projection coordinate system, including:

[0039] The spatial reference of the image is determined by reading Base64 encoded PNG screenshots using the GDAL library and combining them with boundary coordinates.

[0040] The PyProj library is used to convert the boundary coordinates from the WGS84 geographic coordinate system to the Web Mercator projection coordinate system, thus obtaining the screenshot boundary coordinates in the Web Mercator projection coordinate system.

[0041] The map screenshot is converted into an image matrix. The GDAL library is used to give the image matrix spatial reference information in the Web Mercator projection coordinate system. Combined with the screenshot boundary coordinates, a GeoTIFF format image is generated to ensure that the spatial position of the image is consistent with the actual plot, which is used for subsequent GEOSAM model recognition.

[0042] Understandably, after obtaining GeoTIFF format imagery, the backend service uses a Python script to call the GEOSAM model interface to convert the generated Web Mercator projection coordinate system GeoTIFF imagery into an input format supported by the model, such as a 512×512 pixel tensor. At the same time, the WGS84 geographic coordinate system coordinates of user-marked points are converted into pixel coordinates in the Web Mercator projection coordinate system using the PyProj library, and then fed into the GEOSAM model as positive samples to guide the model to focus on the target plot for identification.

[0043] Specifically, the GEOSAM model analyzes the spectral and texture features of GeoTIFF images, combines the location information of positive samples, automatically identifies the edges of target plots, performs instance segmentation, and outputs plot identification results. The identification results are raster masks (in TIFF format), where the pixel value of the target plot area is 1 and the pixel value of the non-target plot area is 0.

[0044] Step 3: The backend service converts the initial plot outline vector into vector surface data. It dynamically calculates the Douglas-Puk algorithm simplification tolerance by weighted fusion of global image scale and individual plot scale. The tolerance is then used to perform geometric simplification on the identified plot polygons, eliminating redundant boundary vertices and fully preserving the overall shape of the plot, key boundary inflection points, and core spatial features.

[0045] Specifically, the backend service uses the GDAL library to convert the raster mask output by the GEOSAM model into vector data in GeoJSON format. The vector data contains multiple vector faces, each corresponding to an identified land parcel.

[0046] It is understandable that after the backend service converts the initial plot outline vector into vector surface data, it also performs joint preprocessing through geometric topology and spatial filtering to remove small noise patches, triangles and polygons with self-intersecting topological anomalies with an area smaller than the preset area threshold, and only retains vector surface features that contain the user's point coordinates within the spatial range.

[0047] Specifically, the joint preprocessing steps are as follows: First, calculate the area (in acres) of each vector surface. Assuming a preset area threshold of 0.03 acres, vector areas with an area less than 0.03 acres are considered non-insured plots or noise areas and are removed. Then, remove triangles, polygons, and polygons with self-intersection topological anomalies. Finally, perform spatial inclusion analysis on the retained vector surfaces to filter out vector surfaces containing at least one positive sample from a user's marker, obtaining the initially filtered vector surface data.

[0048] It should be noted that the preset area threshold can be adjusted according to the type of insured crop, with a threshold range of 0.01-0.05 mu.

[0049] In actual operation, self-intersecting polygons are first eliminated through topological validity checks, then triangular polygons are eliminated through shape feature filtering, and finally spatial inclusion analysis is performed on the remaining valid polygons to determine whether the land parcel is within the coverage area.

[0050] Specifically, spatial containment analysis is a fundamental spatial relationship query in Geographic Information Systems (GIS) used to determine whether a geometric object (such as a point, line, or polygon) is completely contained within another geometric object. In this embodiment, it is mainly used to determine whether a user-marked point is contained within these vector polygons. If it is contained, it is retained; otherwise, it is discarded.

[0051] Then, a weighted fusion method combining global image scale and individual plot scale is used to dynamically calculate the simplified tolerance of the Douglas-Peucker algorithm. Specifically, the formula for calculating the simplified tolerance of the Douglas-Peucker algorithm is as follows:

[0052]

[0053] In the formula, ε is the Douglas-Puk algorithm simplification tolerance, W is the image width in the Web Mercator projection coordinate system, H is the image height in the Web Mercator projection coordinate system, L is the length of the long side of the bounding rectangle of a single plot, and ∂ is a user-defined simplification adjustment coefficient used to manually control the overall simplification intensity.

[0054] In this embodiment, the simplification intensity is adaptively adjusted to meet the dual constraints of the overall size of the simplified tolerance integrated image and the actual size of each individual plot. This approach balances global scale consistency with the detailed representation of individual plots, effectively avoiding the problems of oversimplification of small plots and undersimplification of large plots, and improving the spatial regularity of vector polygons and the lightweight effect of data.

[0055] Finally, the tolerance is used to perform geometric simplification on the identified plot polygons. By recursively eliminating intermediate points that deviate from the line connecting the beginning and end by less than the tolerance, i.e., redundant boundary vertices, the overall shape of the plot, key boundary inflection points such as corners and protrusions, and other feature points and core spatial features are fully preserved.

[0056] As is understandable, the Douglas-Peucker algorithm is a classic curve simplification algorithm used to significantly reduce the number of vertices that make up a curve while preserving the overall shape of the curve. Its core idea is to approximate the original curve with a polyline with as few points as possible, and the approximation error does not exceed a preset tolerance.

[0057] Step 4: The backend service converts the vector surface coordinates of the geometrically simplified land polygon from the Web Mercator projection coordinate system to the WGS84 geographic coordinate system and returns them to the front-end interactive interface.

[0058] Specifically, the backend service uses the PyProj library to convert the initially filtered Web Mercator projection coordinate system vector polygon coordinates into WGS84 geographic coordinate system latitude and longitude coordinates using projection transformation formulas. The conversion accuracy error is controlled within 0.5 pixels to ensure that the vector polygon position accurately matches the front-end map interface. After the conversion is complete, the vector polygon data is encapsulated in GeoJSON format and returned to the front end via an HTTP response.

[0059] Step 5: The front-end interactive interface draws the vector surface data on the map interface and displays it after overlaying it with the satellite remote sensing image data for user confirmation.

[0060] The front-end receives WGS84 geographic coordinate system vector polygon data returned from the back-end and uses the Leaflet component to draw the vector polygons onto the map interface, overlaying them on the original satellite imagery. Furthermore, different colors can be used to mark identified areas for easier user viewing.

[0061] Users compare the identified land parcel boundaries with satellite imagery to check their accuracy and completeness. If there are any deviations, users can adjust the boundaries by dragging the vector surface vertices using the manual correction function provided on the front end. If the identification results are correct, users click the confirmation button to submit the confirmation.

[0062] It should be noted that if the recognition result is significantly off, the user can click the "Re-recognize" button to return to step 1 and re-mark the points.

[0063] Step 6: The front-end interactive interface will also overlay and analyze the vector surface data confirmed by the user with the built-in remote sensing interpretation results of crops, retain only the part of the vector surface data that overlaps with the corresponding crop interpretation area, and output standardized agricultural insurance insured land data for use in agricultural insurance filing and area calculation.

[0064] After the user confirms the recognition result, the front end overlays and analyzes the confirmed vector surface data (WGS84 geographic coordinate system) with the built-in crop remote sensing interpretation results. The built-in crop remote sensing interpretation results are pre-generated vector data (WGS84 geographic coordinate system, GeoJSON format) of wheat planting areas based on time-series remote sensing data. The front end uses the Leaflet spatial overlay plugin to perform intersection operations, retaining only the parts of the vector surface that overlap with the wheat interpretation area, and automatically removing non-wheat planting areas (such as roads, ditches, bare land, etc.).

[0065] After overlay and cropping, standardized agricultural insurance insured land data is output in SHP or GeoJSON format, including land parcel number, boundary coordinates, planting area, crop type, and other information. This data can be directly imported into the agricultural insurance business system for insurance filing, insured amount calculation, and risk control verification.

[0066] To verify the accuracy of the method for confirming the ownership of agricultural insurance-covered land parcels described in this embodiment, 100 insured land parcels were randomly selected, and their actual boundaries and planted areas were obtained through on-site investigation as the true values. The land parcel boundaries and planted areas output by the method of this invention were compared with the true values, and the accuracy index was calculated using a confusion matrix.

[0067] The accuracy of land parcel boundary extraction is calculated as (correctly extracted boundary length / actual boundary length) × 100%, resulting in an accuracy of 93.5%.

[0068] Trimming accuracy = (actual planting area after trimming / output planting area) × 100%, and the calculated trimming accuracy is 92.9%;

[0069] Both accuracy values ​​are above 90%, meeting the accuracy requirements for agricultural insurance coverage plots, confirming the validity of the output data. If the accuracy is below 90%, return to step 4 to adjust the recognition parameters of the GEOSAM model, or return to step 5 to adjust the area screening threshold, and re-execute subsequent steps until the accuracy meets the standard.

[0070] It is understood that this embodiment uses the method of the present invention, which is fully consistent with the actual workflow of Vue+Leaflet front-end and Python+Flask back-end. It completes the drawing and trimming of more than 5,000 insured plots in the wheat planting area of ​​the county. Compared with the traditional manual drawing method, the plot drawing efficiency is improved by more than 85%, and the labor cost is reduced by 75%. The plot boundary extraction accuracy reaches 93.5%, and the trimming accuracy reaches 92.9%, effectively eliminating non-planting areas within the plots, avoiding the problem of inflating the insured area, and significantly reducing the risk of false insurance and insurance fraud.

[0071] Meanwhile, this method achieves full-link automation of "front-end interaction - back-end processing - model recognition - manual confirmation - precise cropping". It is simple for users to operate, requiring no professional technical background. Moreover, the output standardized land data can be directly connected with the agricultural insurance business system to realize the integration of land rights confirmation, area calculation and risk control verification. It solves the problem of the disconnect between existing technology and actual business processes, and greatly improves the intelligence, accuracy and practicality of agricultural insurance underwriting.

[0072] Specifically, in some modified embodiments, the front-end framework can be replaced with Vue or React, and the map component can be replaced with OpenLayers or Mapbox without affecting the front-end interaction, screenshot, and coordinate acquisition functions; the back-end framework can be replaced with Django, and the GEOSAM model can be called via API interface or local deployment, both of which can achieve the land parcel recognition function; the area filtering threshold in step 5 can be adjusted according to the planting characteristics of different crops (such as corn and rice), and the threshold range can be flexibly set between 0.01 and 0.05 acres; the built-in crop remote sensing interpretation results can be replaced with real-time inversion-generated crop distribution data to improve the timeliness of cropping; the coordinate transformation can use ArcGIS Engine components instead of the PyProj library, which can also achieve accurate conversion between the Web Mercator projection coordinate system and the WGS84 geographic coordinate system without affecting the technical effect of the present invention.

[0073] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0074] Example 2

[0075] Based on the same inventive concept, this application also provides a land ownership confirmation system for agricultural insurance based on satellite remote sensing imagery and the GEOSAM model.

[0076] The agricultural insurance land ownership confirmation system includes:

[0077] The front-end interactive interface is used to load and display satellite remote sensing image data. When it detects a user's marking operation in the target insured area, it obtains the user's marking coordinates, takes a screenshot of the map area where the user's marking coordinates are located, and obtains the four boundary coordinates of the map screenshot. The user's marking coordinates, map screenshot, and four boundary coordinates are sent as front-end request parameters to the back-end service. It also draws the received vector surface data on the map interface and displays it after overlaying it with the satellite remote sensing image data. It is also used to overlay and analyze the vector surface data with the built-in crop remote sensing interpretation results, retaining only the part of the vector surface data that overlaps with the corresponding crop interpretation area, and outputting standardized agricultural insurance insured plot data for agricultural insurance filing and area calculation.

[0078] The backend service, upon receiving request parameters from the frontend, converts a map screenshot and its boundary coordinates from the WGS84 geographic coordinate system to the Web Mercator projection coordinate system to obtain GeoTIFF imagery. It then inputs the GeoTIFF imagery and user-marked coordinates into the GEOSAM model for land parcel identification, obtaining initial land parcel outline vectors. Furthermore, it converts these initial outline vectors into vector polygon data, dynamically calculating the Douglas-Pock algorithm simplification tolerance using a weighted fusion of global image scale and individual land parcel scale. This tolerance is then used to perform geometric simplification on the identified land parcel polygons, removing redundant boundary vertices while fully preserving the overall land parcel shape, key boundary inflection points, and core spatial features. Finally, it converts the vector polygon coordinates of the geometrically simplified land parcels from the Web Mercator projection coordinate system to the WGS84 geographic coordinate system and returns them to the frontend interface for display.

[0079] Example 3

[0080] Based on the above embodiments, this embodiment provides a terminal device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0081] Memory, used to store computer programs;

[0082] When the processor executes the program stored in the memory, it implements the steps of the method for confirming the ownership of agricultural insurance land parcels based on satellite remote sensing imagery and the GEOSAM model as described in Example 1.

[0083] Example 4

[0084] Based on the above embodiments, this embodiment provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the method for confirming the ownership of agricultural insurance-covered land parcels based on satellite remote sensing imagery and the GEOSAM model described in Embodiment 1.

[0085] Example 5

[0086] Based on the above embodiments, this embodiment provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the method for confirming the ownership of agricultural insurance-covered land parcels based on satellite remote sensing imagery and the GEOSAM model described in Embodiment 1.

[0087] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them; although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications can still be made to the specific implementation of the present invention or equivalent substitutions can be made to some technical features without departing from the spirit of the technical solutions of the present invention, and all such modifications and substitutions should be covered within the scope of the technical solutions claimed in the present invention.

Claims

1. A method for confirming the ownership of agricultural insurance-covered land parcels based on satellite remote sensing imagery and the GEOSAM model, characterized in that, Includes the following steps: The front-end interactive interface loads and displays satellite remote sensing image data. When it detects that the user has made a checkpoint operation in the target insured area, it obtains the user's checkpoint coordinates and takes a screenshot of the map area where the user's checkpoint coordinates are located. At the same time, it obtains the four boundary coordinates of the map screenshot and sends the user's checkpoint coordinates, map screenshot and four boundary coordinates as front-end request parameters to the back-end service. After receiving the request parameters from the front end, the backend service converts the map screenshot and the four boundary coordinates from the WGS84 geographic coordinate system to the Web Mercator projection coordinate system to obtain the GeoTIFF image. The GeoTIFF image and the user's marked coordinates are then input into the GEOSAM model for land parcel recognition to obtain the initial land parcel outline vector. The backend service converts the initial plot outline vector into vector surface data, dynamically calculates the Douglas-Puk algorithm simplification tolerance by weighted fusion of global image scale and individual plot scale, and uses this tolerance to perform geometric simplification processing on the identified plot polygons, removes redundant boundary vertices, and fully preserves the overall shape of the plot, key boundary inflection points and core spatial features. The backend service converts the vector surface coordinates of the geometrically simplified land polygons from the Web Mercator projection coordinate system to the WGS84 geographic coordinate system and returns them to the frontend interface. The front-end interactive interface draws vector surface data on the map interface and displays it after overlaying it with satellite remote sensing image data for user confirmation; The front-end interactive interface also overlays and analyzes the vector surface data confirmed by the user with the built-in remote sensing interpretation results of crops, retaining only the part of the vector surface data that overlaps with the corresponding crop interpretation area, and outputs standardized agricultural insurance insured land data for use in agricultural insurance filing and area calculation.

2. The method for confirming the ownership of agricultural insurance-covered land parcels based on satellite remote sensing imagery and the GEOSAM model according to claim 1, characterized in that, Convert the map screenshot and its boundary coordinates from the WGS84 geographic coordinate system to the Web Mercator projection coordinate system, including: The spatial reference of the image is determined by reading map screenshots from the GDAL library and combining them with the four boundary coordinates. The PyProj library is used to convert the boundary coordinates from the WGS84 geographic coordinate system to the Web Mercator projection coordinate system, thus obtaining the screenshot boundary coordinates in the Web Mercator projection coordinate system. The map screenshot is converted into an image matrix. The GDAL library is used to give the image matrix spatial reference information in the Web Mercator projection coordinate system. Combined with the screenshot boundary coordinates, a GeoTIFF format image is generated.

3. A method for confirming the ownership of agricultural insurance-covered land parcels based on satellite remote sensing imagery and a GEOSAM model, as described in claim 1 or 2, characterized in that... After the backend service converts the initial plot outline vector into vector surface data, it performs joint preprocessing through geometric topology and spatial filtering to remove fragmented noise patches, triangles and polygons with self-intersecting topological anomalies with an area smaller than a preset area threshold, and retain only vector surface features that contain the user's point coordinates within the spatial range.

4. The method for confirming the ownership of agricultural insurance-covered land parcels based on satellite remote sensing imagery and the GEOSAM model according to claim 3, characterized in that, The simplified tolerance calculation formula of the Douglas-Pock algorithm is as follows: In the formula, ε is the Douglas-Puk algorithm simplification tolerance, W is the image width in the Web Mercator projection coordinate system, H is the image height in the Web Mercator projection coordinate system, L is the length of the long side of the bounding rectangle of a single plot, and ∂ is a user-defined simplification adjustment coefficient used to manually control the overall simplification intensity.

5. The method for confirming the ownership of agricultural insurance-covered land parcels based on satellite remote sensing imagery and the GEOSAM model according to claim 1, characterized in that, The front-end interactive interface displays vector surface data while providing a manual correction function, allowing users to manually adjust the vector surface boundaries and confirm them. If the user does not confirm, return to the re-marking operation.

6. A system for confirming the ownership of agricultural insurance-covered land parcels based on satellite remote sensing imagery and a GEOSAM model, characterized in that, include: The front-end interactive interface is used to load and display satellite remote sensing image data. When it detects a user's marking operation in the target insured area, it obtains the user's marking coordinates, takes a screenshot of the map area where the user's marking coordinates are located, and obtains the four boundary coordinates of the map screenshot. The user's marking coordinates, map screenshot, and four boundary coordinates are sent as front-end request parameters to the back-end service. It also draws the received vector surface data on the map interface and displays it after overlaying it with the satellite remote sensing image data. It is also used to overlay and analyze the vector surface data with the built-in crop remote sensing interpretation results, retaining only the part of the vector surface data that overlaps with the corresponding crop interpretation area, and outputting standardized agricultural insurance insured plot data for agricultural insurance filing and area calculation. The backend service, upon receiving request parameters from the frontend, converts map screenshots and boundary coordinates from the WGS84 geographic coordinate system to the Web Mercator projection coordinate system to obtain GeoTIFF imagery. It then inputs the GeoTIFF imagery and user-marked coordinates into the GEOSAM model for land parcel identification, obtaining initial land parcel outline vectors. Furthermore, it converts these initial outline vectors into vector polygon data, dynamically calculates the Douglas-Puk algorithm simplification tolerance using a weighted fusion of global image scale and individual land parcel scale, and utilizes this tolerance to perform geometric simplification processing on the identified land parcel polygons, eliminating redundant boundary vertices and fully preserving the overall shape of the land parcels, key boundary inflection points, and core spatial features. It also transforms the vector surface coordinates of the geometrically simplified land polygons from the Web Mercator projection coordinate system to the WGS84 geographic coordinate system and returns them to the front-end interactive interface for display.

7. A terminal device, characterized in that: It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; The processor, when executing a program stored in memory, implements the steps of the method for confirming the ownership of agricultural insurance-covered land parcels based on satellite remote sensing imagery and the GEOSAM model as described in any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for confirming the ownership of agricultural insurance-covered land parcels based on satellite remote sensing imagery and the GEOSAM model as described in any one of claims 1 to 5.

9. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the steps of the method for confirming the ownership of agricultural insurance-covered land parcels based on satellite remote sensing imagery and the GEOSAM model as described in any one of claims 1 to 5.

Images