A method and system for automatic generation of cadastral maps based on AI

Abstract

This invention discloses an AI-based method and system for automatically generating land parcel maps, specifically relating to the field of land parcel map generation technology. This invention uses a radio frequency transceiver and a drone to collect the coordinates of the inflection points of the boundary lines of a land parcel and the coordinates of the intersection points of the boundary lines between the land parcel and adjacent land parcels. The boundary lines are obtained by connecting the coordinates of each inflection point in sequence. Then, the boundary lines and land parcel area are corrected and judged. If the boundary line correction task is not triggered, the process directly proceeds to the land parcel area correction stage; otherwise, it proceeds to the land parcel boundary point correction stage. From numerous inflection points, points that best represent the direction of the land parcel are selected as boundary points, solving the problem of a large workload in boundary point processing. Simultaneously, the invention ensures that the land parcel area and boundary lines remain consistent with the actual land parcel area and boundary lines, and analyzes the four boundaries of the land parcel, thus guaranteeing the accuracy of the subsequent land parcel map to a certain extent.

Description

Technical Field

[0001] This invention relates to the field of parcel map generation technology, and more specifically, to an AI-based method and system for automatic parcel map generation. Background Technology

[0002] A cadastral map is a cadastral map that describes the location of a land parcel, the relationship between boundary points and lines, and the numbers of adjacent land parcels. When creating a cadastral map, it is necessary to accurately locate the boundary points, boundary lines, and the four boundaries of the land parcel. Unclear land ownership information will greatly affect the effectiveness of the cadastral map. Therefore, a precise and efficient method is needed to obtain land ownership information.

[0003] Existing methods for automatically generating land parcel maps use drones to take aerial photos of the land parcel and its surrounding areas to create a base map when obtaining land parcel ownership information. Then, the base map is used to conduct door-to-door surveys and identify land boundaries to determine the land parcel boundary points. Connecting these boundary points yields the land parcel boundary lines, which in turn provides information on the four boundaries of the land parcel. This method can, to a certain extent, ensure the accuracy of land parcel map production.

[0004] However, existing methods still have some problems: they require door-to-door surveys and boundary identification, and the resulting land parcel boundary points are mostly inflection points of the parcels. For parcels with many inflection points, thousands of boundary points are generated, which makes the description of boundary points and their directions quite complicated and greatly increases the workload of determining land parcel ownership information. Therefore, it is still necessary to further improve the efficiency of determining land parcel ownership information. Summary of the Invention

[0005] To overcome the aforementioned shortcomings of existing technologies, this invention provides an AI-based method and system for automatically generating land parcel maps. It uses a radio frequency transmitter to collect land parcel ownership information, uses a drone to acquire land parcel geographic data, and then integrates and corrects the land parcel ownership information. From numerous turning points, it selects points that better represent the direction of the land parcel as boundary points, thus solving the problem of the large workload in boundary point processing.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an AI-based method for automatically generating cadastral maps, comprising the following steps:

[0007] S1. Collect land parcel data: Use radio frequency transmitters to collect land parcel ownership information and use drones to acquire land parcel geographic data.

[0008] S2. Integrate and correct land ownership information: Correct the land boundary lines, land area and land boundary points in the collected land ownership information;

[0009] S3. Analyze the boundaries of the land parcel: The boundary line is divided into four parts, namely east, south, west and north, using the method of right rectangular side projection. The boundary information of the land parcel is obtained based on the set boundary determination rules.

[0010] S4. AI-based identification and annotation of land parcel geographic data: AI algorithms are used to identify and annotate the collected land parcel geographic data.

[0011] S5. Create a parcel map template: Use ArcGIS to create a parcel map template;

[0012] S6. Automated batch generation of parcel maps: Run a Python script in the ArcMap command prompt. The script automatically reads the parcel attribute data package and generates a parcel map for each parcel based on the MXD template.

[0013] S7. Parcel Map Inspection and Correction: Check whether the parcel boundaries are accurate, whether the text labels are correct, and whether the map elements are complete. If any one or more of the above conditions are not met, the parcel map needs to be corrected. After the correction is completed, export the parcel map as a preset image format.

[0014] S8. Data collection for parcel map generation: Collect parcel ownership information correction data, parcel map verification and correction data, and parcel geographic data identification data.

[0015] S9. Evaluation of Parcel Map Generation Quality: The quality of parcel map generation is evaluated after processing the collected parcel ownership information correction data, parcel map verification and correction data, and parcel geographic data identification data.

[0016] To achieve the above objectives, the present invention provides the following technical solution: an AI-based automatic cadastral map generation system, which implements the above-mentioned AI-based automatic cadastral map generation method, including:

[0017] Land parcel data acquisition module: Uses radio frequency transmitters to collect land parcel ownership information and uses drones to acquire land parcel geographic data;

[0018] Land ownership information correction module: used to correct the land boundary lines, land area and land boundary points in the collected land ownership information;

[0019] Parcel Boundary Analysis Module: The boundary line is divided into four parts (east, south, west, and north) using the method of rectangular edge projection, and the parcel boundary information is obtained based on the set boundary determination rules;

[0020] Parcel geographic data identification module: Based on AI algorithms, it identifies and labels the collected parcel geographic data;

[0021] Categorical Map Template Creation Module: Creates categorical map templates using ArcGIS;

[0022] Parcel map auto-generation module: Run a Python script in the ArcMap command prompt. The script automatically reads the parcel attribute data package and generates a parcel map for each parcel based on the MXD template.

[0023] Parcel Map Self-Check and Correction Module: Checks whether the parcel boundaries are accurate, whether the text annotations are correct, and whether the map elements are complete. If any one or more of the above conditions are not met, the parcel map needs to be corrected. After the correction is completed, the parcel map is exported as a preset image format.

[0024] Parcel map generation data acquisition module: used to collect parcel ownership information correction data, parcel map verification and correction data, and parcel geographic data identification data;

[0025] The parcel map generation quality evaluation module evaluates the quality of parcel map generation after processing the collected parcel ownership information correction data, parcel map verification and correction data, and parcel geographic data identification data.

[0026] The technical effects and advantages of this invention are as follows:

[0027] This invention provides a number of radio frequency (RF) transmitters to land parcel owners or responsible persons. Upon receiving the RF transmitters, the land parcel owner or responsible person inputs the parcel number, the name of the land parcel owner or responsible person, the parcel area, and the RF transmitter number into the transmitter. The RF transmitters are then sequentially positioned according to their numbers at the turning points of the parcel boundary lines and at the intersections with adjacent parcel boundary lines. The input information from the RF transmitters is transmitted as RF communication signals. This controls a drone to acquire orthophoto real-view image data of the parcel area using configured high-definition camera equipment, following a pre-set flight path. During this process, the RF communication signals transmitted by each RF transmitter are received by an RF communication signal receiving device mounted on the drone, thus obtaining... The system includes the land parcel number, the name of the land parcel owner or the land parcel ownership unit, the land parcel area, the coordinates of the turning points of the land parcel's boundary lines, and the coordinates of the intersection points of the land parcel's boundary lines with those of adjacent land parcels. This avoids the workload of on-site boundary demarcation. The system then corrects and determines the land parcel boundary lines and area. If the land parcel boundary line correction task is not triggered, the system directly proceeds to the land parcel area correction stage. If the land parcel area correction task is not triggered, the system directly proceeds to the land parcel boundary point correction stage. From numerous turning points, points that best represent the direction of the land parcel are selected as boundary points, solving the problem of a large workload in boundary point processing. Simultaneously, the system ensures that the land parcel area and boundary lines remain consistent with the actual land parcel area and boundary lines. Furthermore, the system analyzes the four boundaries of the land parcel, which to a certain extent guarantees the accuracy of subsequent land parcel maps. Attached Figure Description

[0028] Figure 1 This is a diagram illustrating the method steps of the present invention.

[0029] Figure 2 This is a system structure block diagram of the present invention. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] like Figure 1 The embodiment shown provides an AI-based method for automatically generating cadastral maps, including the following steps:

[0032] S1. Collect land parcel data: Use radio frequency transmitters to collect land parcel ownership information and use drones to acquire land parcel geographic data.

[0033] Furthermore, the land parcel ownership information includes the parcel number, parcel boundary line, first boundary point, parcel area, and name of the parcel owner or parcel ownership unit. The parcel geographic data is the parcel ground orthophoto real scene image data, and the parcel boundary line is the ownership boundary line between the parcel and adjacent parcels.

[0034] Furthermore, the specific steps for collecting land parcel data are as follows:

[0035] A1. To issue a certain number of radio frequency transmitters to the land parcel owner or the person in charge of the land parcel owner unit, wherein the radio frequency transmitters are equipped with the functions of transmitting radio frequency communication signals and inputting information;

[0036] A2. After receiving the radio frequency transmitter, the land parcel owner or the person in charge of the land parcel ownership unit shall input the land parcel number, the name of the land parcel owner or the land parcel ownership unit, the land parcel area, and the radio frequency transmitter number into the radio frequency transmitter. The radio frequency transmitter shall be arranged in the order of the numbers at the turning point of the land parcel boundary line and at the intersection with the boundary line of the adjacent land parcel. The radio frequency transmitters arranged at the turning point of the land parcel boundary line shall be numbered a1, a2, ..., an, and the radio frequency transmitters arranged at the intersection with the boundary line of the adjacent land parcel shall be numbered b1, b2, ..., bn. The radio frequency transmitters located at both the turning point of the land parcel boundary line and the intersection with the boundary line of the adjacent land parcel shall have both numbers.

[0037] In this embodiment, it should be specifically noted that if the 9th inflection point on the land parcel boundary line is also the 1st intersection point with the boundary line of an adjacent land parcel, then the radio frequency transmitter placed at that point is numbered a9b1.

[0038] A3. The input information of the radio frequency transmitter is transmitted as a radio frequency communication signal. The UAV is controlled to use the configured high-definition camera equipment to obtain orthophoto real-view image data of the land parcel area according to the set flight route. During this process, the radio frequency communication signal receiving equipment deployed on the UAV receives the radio frequency communication signals sent by each radio frequency transmitter and obtains the land parcel number, land parcel owner or land parcel owner unit name, land parcel area, radio frequency transmitter number information and the geographical coordinates of the radio frequency transmitter input from the received radio frequency transmitter.

[0039] In this embodiment, it should be specifically noted that the drone is equipped with high-definition camera equipment and radio frequency communication signal receiving equipment. Common high-definition camera equipment includes high-resolution CCD digital cameras, laser scanners, and multispectral cameras. High-resolution CCD digital cameras have high resolution and can capture fine textures and details of the ground or target area. Laser scanners measure distances and generate 3D models by emitting laser beams and receiving reflected signals. Multispectral cameras can capture spectral information of different bands and can be used to analyze vegetation cover and soil type. In this embodiment, a high-resolution CCD digital camera is selected on the drone to obtain high-quality orthophoto images of the land parcel, which is beneficial to improving the accuracy of subsequent identification of the components within the land parcel. Common radio frequency communication signal receiving equipment includes radio frequency receivers and GPS receivers. Radio frequency receivers can ensure the stability of receiving radio frequency communication signals emitted by the radio frequency transmitter on the land parcel. GPS receivers can determine the precise location data of the radio frequency transmitter used. In this embodiment, a radio frequency receiver is selected on the drone to ensure the stability of the received signal.

[0040] A4. Connect the corresponding geographical coordinates of the radio frequency transmitters located at the inflection points of the land parcel boundary lines in sequence according to their numbers. The resulting closed lines are the land parcel boundary lines. The boundary lines a1a2, a2a3, ..., ana1 are numbered x1, x2, ..., xn in sequence. Mark the geographical coordinates and numbers of the radio frequency transmitters located at the inflection points of the land parcel boundary lines as the third boundary point and corresponding number. Mark the geographical coordinates and numbers of the radio frequency transmitters located at the intersections with the boundary lines of adjacent land parcels as the first boundary point and corresponding number.

[0041] S2. Integrate and correct land ownership information: Correct the land boundary lines, land area and land boundary points in the collected land ownership information;

[0042] Furthermore, the specific steps for integrating and correcting land parcel ownership information are as follows:

[0043] B1. Retrieve the boundary lines of the parcel and the adjacent parcels, and observe whether there are any overlapping areas between the corresponding parcel area and the adjacent parcel area besides the boundary lines. If there are, the parcel boundary line correction task is triggered. If not, the parcel boundary line correction task is not triggered, and the parcel area correction process is directly initiated.

[0044] In this embodiment, it should be specifically noted that when the land parcel boundary line correction task is triggered, the first execution plan is selected first, that is, the land ownership data of the land parcel and the adjacent land parcel are retrieved and the boundary lines are directly corrected. If the boundary lines in the land ownership source data are unclear or the boundary lines are inconsistent with the actual situation, the second execution plan is selected, that is, the land parcel owner or the person in charge of the land parcel owner unit is notified to conduct on-site boundary demarcation with the land parcel owner or the person in charge of the adjacent land parcel owner unit and the investigators. After the land parcel boundary line correction is completed, the geographical coordinates and numbers of the first boundary point and the third boundary point, as well as the land parcel boundary line number, are updated.

[0045] B2. When entering the land parcel area correction stage, check the coordinates and corresponding numbers of each RF transmitter placement point to ensure that the coordinates and numbers of the RF transmitters are arranged clockwise or counterclockwise according to the boundary line. Then, substitute the RF transmitter coordinates into the Gaussian area formula to calculate the land parcel area Ax. Based on the calculated land parcel area and the received land parcel area, calculate the area deviation coefficient αm. The area deviation coefficient is the ratio between the absolute value of the difference between the calculated land parcel area Ax and the received land parcel area As and the smaller of the two values. The specific formula is as follows: When the calculated land area Ax is the same as the received land area As, αm=0. The calculated area deviation coefficient is compared with the preset area deviation upper limit. If the calculated area deviation coefficient is greater than the area deviation upper limit, the land area correction task is triggered. Otherwise, the land area correction task is not triggered, and the land boundary point correction stage is entered.

[0046] In this embodiment, it should be specifically noted that the Gaussian area formula is as follows: After calculation, let Ax represent the calculated land area, (xi,yi) be the coordinates of the i-th RF transmitter placement point of the land polygon, and n be the number of RF transmitters. Here is a set of RF transmitter placement point coordinates to demonstrate the process of calculating the land area using the Gaussian area formula: RF transmitter placement point coordinates are pa1(30m,50m), pa2(30m,150m), pa3(150m,240m), pa4(150m,50m), then the land area is... .

[0047] In this embodiment, it is specifically noted that when the land parcel area correction task is triggered, the location of the land parcel radio frequency transmitter is confirmed one by one, and the location of the radio frequency transmitter with incorrect position is corrected. If the location of the radio frequency transmitter does not need to be corrected, the received land parcel area is directly replaced with the calculated land parcel area.

[0048] In this embodiment, it should be specifically noted that after the correction of the land parcel boundary line, area, or land parcel area, if the turning point of the land parcel boundary line or the intersection point of the boundary line between the land parcel and the adjacent land parcel changes, the corresponding boundary point number will be updated sequentially.

[0049] B3. Extract the first boundary point to form a point set. Taking the first boundary point numbered b1 as the starting point, calculate the straight-line distance D1 between the first boundary point numbered b1 and the first boundary point numbered b2 based on the point coordinates and the Euclidean distance formula. At the same time, obtain the boundary line distance L1 between b1 and b2. Calculate L1 / D1 and compare the calculation result with the preset value. If the calculated value is greater than the preset value, it is determined that a new boundary point needs to be inserted between b1 and b2. Otherwise, it is not necessary to insert a new boundary point. Continue to determine the insertion of a new boundary point between the first boundary point numbered b2 and the first boundary point numbered b3.

[0050] B4. When it is determined that a new boundary point needs to be inserted between b1 and b2, find the point with the farthest distance between the boundary lines of b1 and b2 and the straight line connecting b1 and b2, and insert it between b1 and b2 as the new boundary point. Mark the new boundary point as the first boundary point and number it b2. The original first boundary point numbers b2, b3, ..., bn are automatically shifted one position to the right and changed to b3, b4, ..., bn+1.

[0051] B5. Repeat steps B3 and B4 until the ratio of the boundary line distance to the straight line distance of any adjacent first boundary point is less than or equal to the preset value. Then the point set correction of the first boundary point is completed.

[0052] S3. Analyze the boundaries of the land parcel: The boundary line is divided into four parts, namely east, south, west and north, using the method of right rectangular side projection. The boundary information of the land parcel is obtained based on the set boundary determination rules.

[0053] Furthermore, the specific steps for analyzing the boundaries of the land parcel are as follows:

[0054] C1. First, divide the boundary line into four parts: east, south, west, and north using the method of right rectangular side projection;

[0055] C2. For all neighboring parcels of this parcel, if a neighboring parcel has a common edge with this parcel in any of the four directions of this parcel (east, south, west, north), then the neighboring parcel is included in the boundary information of the corresponding direction of this parcel. If a neighboring parcel has a common edge with multiple parts of this parcel, then the neighboring parcel is included in the boundary information of the corresponding multiple directions of this parcel.

[0056] In this embodiment, it is necessary to specifically explain the process of parcel boundary analysis using the following examples: Parcel F1 is designated as F1, and neighboring parcels are F2, F3, F4, F5, F6, F7, F8, and F9. Neighboring parcels F4 and F8 share boundary points only with parcel F1. F2 and F1 share a common boundary on both the west and north sides. F3 shares a common boundary on the north side only with F1. F5 and F1 share a common boundary on both the north and east sides. F6 shares a common boundary on the east side only with F1. F7 and F1 share a common boundary on both the south and west sides. There are common boundaries. F9 only shares a common boundary with F1 on the south side. Therefore, the east boundary of this parcel F1 is F5 and F6, the west boundary is F2 and F7, the south boundary is F7 and F9, and the north boundary is F2, F3, and F5. For enclaves (island parcels), since there are no adjacent boundary lines, it is impossible to obtain their four boundaries using the above method. The solution is to determine whether this parcel is completely contained by another parcel. If so, then the four boundaries of this parcel are all of the other parcel. That is, if F1 is completely within the area of ​​parcel F10, then the four boundaries of F1 are all of F10.

[0057] S4. AI-based identification and annotation of land parcel geographic data: AI algorithms are used to identify and annotate the collected land parcel geographic data.

[0058] Furthermore, the specific operations for AI to identify and label land parcel geographic data are as follows: Identify the types of elements within the land parcel, highlight the edge outlines of buildings, structures, roads, and topographic features within the land parcel, highlight boundary lines and boundary points, label buildings within the land parcel with building names as data labels, label structures within the land parcel with structure names as data labels, label topographic features within the land parcel with topographic feature names as data labels, label each segment of boundary line with boundary line number and boundary line side length as data labels, and label boundary points with boundary point number and coordinates.

[0059] S5. Create a parcel map template: Use ArcGIS to create a parcel map template;

[0060] Furthermore, the specific steps for creating the parcel map template are as follows:

[0061] D1. Open ArcMap to create a layout, set the size, and then add a title, legend, and north arrow to the layout;

[0062] D2. Add the parcel map layer, MappingIndex layer, boundary point layer, boundary line layer, annotation layer, building and structure layer, adjacent parcel map layer, and line vector map layer converted from QSDW parcel strata, and set the display order, symbols, and styles of each layer;

[0063] In this embodiment, it is specifically noted that the parcel map layer is the core layer, used to display the boundaries, shape, and extent of the parcels. The parcel map layer is typically created based on parcel polygon data in a geodatabase, ensuring that each parcel has a clear boundary and extent definition. Boundary points are key nodes on the parcel boundaries, and boundary lines are line segments connecting these nodes, together forming the parcel boundaries. These two layers are used to accurately represent the boundary locations of the parcels, ensuring the accuracy of ownership boundaries. The MappingIndex layer serves as the index layer for the automatic mapping code, used to traverse each parcel and extract relevant information from the attribute table. The annotation layer is used to add necessary text annotations, such as parcel number, land category number, etc. Information such as area, rights holder, and address helps users quickly identify and understand relevant information about the land parcel; the building layer displays information such as the location, level, and structure of buildings or structures within the land parcel, which helps users understand the layout and nature of buildings or structures within the land parcel; the adjacent land parcel layer displays information about other land parcels adjacent to the current land parcel, including land parcel number, boundary lines, etc., which helps users understand the relationship between the current land parcel and surrounding land parcels; the line vector layer converted from the QSDW land parcel layer is used to display the boundaries of the land parcel map, and may need to be clipped and processed through program code to meet specific display needs, such as displaying only a small line segment as a "spike".

[0064] D3. Use ArcPy to write an automatic cartographic script that includes reading layer attributes, drawing map features, updating text elements, intelligently adjusting the map display scale, and intelligently placing annotations.

[0065] D4. Save as an MXD template file.

[0066] S6. Automated batch generation of parcel maps: Run a Python script in the ArcMap command prompt. The script automatically reads the parcel attribute data package and generates a parcel map for each parcel based on the MXD template.

[0067] In this embodiment, it should be specifically noted that the parcel attribute data package stores the parcel number, the name of the parcel owner or parcel owner unit, the parcel area, the parcel boundary line number, the parcel boundary line side length, the first boundary point set, the parcel boundary information, and the parcel geographic data after identification and annotation. The data format in the data package is an ArcGIS-recognizable format.

[0068] S7. Parcel Map Inspection and Correction: Check whether the parcel boundaries are accurate, whether the text labels are correct, and whether the map elements are complete. If any one or more of the above conditions are not met, the parcel map needs to be corrected. After the correction is completed, export the parcel map as a preset image format.

[0069] S8. Data collection for parcel map generation: Collect parcel ownership information correction data, parcel map verification and correction data, and parcel geographic data identification data.

[0070] Furthermore, the land parcel ownership information correction data includes the area deviation coefficient of each parcel, the corrected coordinates of the first boundary point, and the true coordinates of the first boundary point; the parcel map verification and correction data includes the number of correct text annotations, the number of corrected text annotations, the number of correct map elements, the number of corrected map elements, the correct parcel boundary length, and the corrected map boundary length; the parcel geographic data identification data includes the number of correctly identified and labeled parcel features and the total number of parcel features.

[0071] S9. Evaluation of Parcel Map Generation Quality: The quality of parcel map generation is evaluated after processing the collected parcel ownership information correction data, parcel map verification and correction data, and parcel geographic data identification data.

[0072] Furthermore, the specific steps for evaluating the quality of the parcel map generation are as follows:

[0073] E1. Compare the corrected coordinates Pai(xai, yai) of the i-th first boundary point with the corresponding true coordinates Pbi(xbi, ybi) of the i-th first boundary point, and calculate the position deviation coefficient αwi of the i-th first boundary point. The specific formula is as follows: The average position deviation coefficient αwe is obtained by averaging the calculated position deviation coefficients.

[0074] E2. The parcel map production accuracy coefficient βe can be obtained by calculating the ratio of the correct text label quantity nax to the sum of the correct text label quantity nay, the correct map feature quantity nbx to the sum of the correct map feature quantity nby, and the correct parcel boundary length nxx to the sum of the correct parcel boundary length ncy. The specific formula is as follows: ;

[0075] E3. The parcel geographic data identification accuracy coefficient γe can be obtained by calculating the ratio of the number of correctly identified and labeled parcel features (nex) to the total number of parcel features (nz). The specific formula is as follows: ;

[0076] E4. The higher the parcel map production accuracy coefficient and parcel geographic data recognition accuracy coefficient, the higher the parcel map generation quality. The lower the average location deviation coefficient and area deviation coefficient, the higher the parcel map generation quality. Calculate the product of the parcel map production accuracy coefficient and the parcel geographic data recognition accuracy coefficient, and calculate the product of the average location deviation coefficient compensation processing result and the area deviation coefficient compensation processing result. The ratio of these two is the parcel map generation quality index Rz. The specific formula is as follows: .

[0077] In this embodiment, it should be noted that the preset values ​​used are selected based on actual needs, and no specific value limit is imposed here.

[0078] like Figure 2 This embodiment provides an AI-based automatic parcel map generation system, including a parcel data acquisition module, a parcel ownership information correction module, a parcel boundary analysis module, a parcel geographic data identification module, a parcel map template creation module, a parcel map automatic generation module, a parcel map self-checking and correction module, a parcel map generation data acquisition module, a parcel map generation quality evaluation module, and a database.

[0079] The land parcel data acquisition module is connected to the land parcel ownership information correction module. The land parcel ownership information correction module is connected to the land parcel boundary analysis module and the land parcel geographic data identification module. The land parcel ownership information correction module, the land parcel boundary analysis module, the land parcel geographic data identification module, the land parcel map template creation module, and the land parcel map automatic generation module are connected. The land parcel map automatic generation module, the land parcel map self-checking and correction module, the land parcel map generation data acquisition module, and the land parcel map generation quality evaluation module are connected sequentially. All modules in the system are connected to the database.

[0080] The land parcel data acquisition module uses a radio frequency transmitter to collect land parcel ownership information and uses a drone to acquire land parcel geographic data.

[0081] The land ownership information correction module is used to correct the land boundary lines, land area, and land boundary points in the collected land ownership information.

[0082] The land parcel boundary analysis module uses the rectangular side projection method to divide the boundary line into four parts: east, south, west, and north, and obtains the land parcel boundary information based on the set boundary determination rules.

[0083] The parcel geographic data identification module uses AI algorithms to identify and label the collected parcel geographic data;

[0084] The parcel map template creation module uses ArcGIS to create parcel map templates;

[0085] The automatic parcel map generation module runs a Python script in the ArcMap command prompt. After automatically reading the parcel attribute data package, the script generates a parcel map for each parcel based on the MXD template.

[0086] The parcel map self-checking and correction module checks whether the parcel boundaries are accurate, whether the text annotations are correct, and whether the map elements are complete. If any one or more of the above conditions are not met, the parcel map needs to be corrected. After the correction is completed, the parcel map is exported as a preset image format.

[0087] The parcel map generation data acquisition module is used to collect parcel ownership information correction data, parcel map verification and correction data, and parcel geographic data identification data.

[0088] The parcel map generation quality evaluation module processes the collected parcel ownership information correction data, parcel map verification and correction data, and parcel geographic data identification data to evaluate the parcel map generation quality.

[0089] The database is used to store data information for all modules in the system.

[0090] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An AI-based method for automatically generating cadastral maps, characterized in that: Includes the following steps: S1. Collect land parcel data: Use radio frequency transmitters to collect land parcel ownership information and use drones to acquire land parcel geographic data. S2. Integrate and correct land ownership information: Correct the land boundary lines, land area and land boundary points in the collected land ownership information; S3. Analyze the boundaries of the land parcel: The boundary line is divided into four parts, namely east, south, west and north, using the method of right rectangular side projection. The boundary information of the land parcel is obtained based on the set boundary determination rules. S4. AI-based identification and annotation of land parcel geographic data: AI algorithms are used to identify and annotate the collected land parcel geographic data. The specific operations of AI identification and annotation of land parcel geographic data are as follows: Identify the object elements within the land parcel, highlight the edge outlines of buildings, structures, roads and topographic features within the land parcel, highlight the boundary lines and boundary points, use the building name as a data label to mark the buildings within the land parcel, use the structure name as a data label to mark the structures within the land parcel, use the topographic feature name as a data label to mark the topographic features within the land parcel, use the boundary line number and boundary line side length as data labels to mark each segment of the boundary line, and use the boundary point number and coordinates to mark the boundary points; S5. Create a parcel map template: Use ArcGIS to create a parcel map template; S6. Automated batch generation of parcel maps: Run a Python script in the ArcMap command prompt. The script automatically reads the parcel attribute data package and generates a parcel map for each parcel based on the MDX template. S7. Parcel Map Inspection and Correction: Check whether the parcel boundaries are accurate, whether the text labels are correct, and whether the map elements are complete. If any one or more of the above conditions are not met, the parcel map needs to be corrected. After the correction is completed, export the parcel map as a preset image format. S8. Data collection for parcel map generation: Collect parcel ownership information correction data, parcel map verification and correction data, and parcel geographic data identification data. The land parcel ownership information correction data includes the area deviation coefficient of each parcel, the corrected coordinates of the first boundary point, and the true coordinates of the first boundary point; the parcel map verification and correction data includes the number of correct text annotations, the number of corrected text annotations, the number of correct map elements, the number of corrected map elements, the correct parcel boundary length, and the corrected map boundary length; the parcel geographic data identification data includes the number of correctly identified and labeled parcel features and the total number of parcel features. S9. Evaluation of Parcel Map Generation Quality: The quality of parcel map generation is evaluated after processing the collected parcel ownership information correction data, parcel map verification and correction data, and parcel geographic data identification data.

2. The method for automatically generating cadastral maps based on AI according to claim 1, characterized in that: The land parcel ownership information includes the parcel number, parcel boundary line, first boundary point, parcel area, and name of the parcel owner or parcel ownership unit. The parcel geographic data is the parcel ground orthophoto real scene image data, and the parcel boundary line is the ownership boundary line between the parcel and adjacent parcels.

3. The method for automatically generating cadastral maps based on AI according to claim 1, characterized in that: The specific steps for collecting land parcel data are as follows: A1. To issue a certain number of radio frequency transmitters to the land parcel owner or the person in charge of the land parcel owner unit, wherein the radio frequency transmitters are equipped with the functions of transmitting radio frequency communication signals and inputting information; A2. After receiving the radio frequency transmitter, the land parcel owner or the person in charge of the land parcel ownership unit shall input the land parcel number, the name of the land parcel owner or the land parcel ownership unit, the land parcel area, and the radio frequency transmitter number into the radio frequency transmitter. The radio frequency transmitter shall be arranged in the order of the numbers at the turning point of the land parcel boundary line and at the intersection with the boundary line of the adjacent land parcel. The radio frequency transmitters arranged at the turning point of the land parcel boundary line shall be numbered a1, a2, ..., an, and the radio frequency transmitters arranged at the intersection with the boundary line of the adjacent land parcel shall be numbered b1, b2, ..., bn. The radio frequency transmitters located at both the turning point of the land parcel boundary line and the intersection with the boundary line of the adjacent land parcel shall have both numbers. A3. The input information of the radio frequency transmitter is transmitted as a radio frequency communication signal. The UAV is controlled to use the configured high-definition camera equipment to obtain orthophoto real-view image data of the land parcel area according to the set flight route. During this process, the radio frequency communication signal receiving equipment deployed on the UAV receives the radio frequency communication signals sent by each radio frequency transmitter and obtains the land parcel number, land parcel owner or land parcel owner unit name, land parcel area, radio frequency transmitter number information and the geographical coordinates of the radio frequency transmitter input from the received radio frequency transmitter. A4. Connect the corresponding geographical coordinates of the radio frequency transmitters located at the inflection points of the land parcel boundary lines in sequence according to their numbers. The resulting closed lines are the land parcel boundary lines. The boundary lines a1a2, a2a3, ..., ana1 are numbered x1, x2, ..., xn in sequence. Mark the geographical coordinates and numbers of the radio frequency transmitters located at the inflection points of the land parcel boundary lines as the third boundary point and corresponding number. Mark the geographical coordinates and numbers of the radio frequency transmitters located at the intersections with the boundary lines of adjacent land parcels as the first boundary point and corresponding number.

4. The method for automatically generating cadastral maps based on AI according to claim 1, characterized in that: The specific steps for integrating and correcting land ownership information are as follows: B1. Retrieve the boundary lines of the parcel and the adjacent parcels, and observe whether there are any overlapping areas between the corresponding parcel area and the adjacent parcel area besides the boundary lines. If there are, the parcel boundary line correction task is triggered. If not, the parcel boundary line correction task is not triggered, and the parcel area correction process is directly initiated. B2. If the land parcel has a regular shape, the area is calculated directly based on the geographical coordinates of the inflection points after matching the corresponding geometric formula. If the land parcel has an irregular shape, the land parcel is divided into multiple regularly shaped sub-regions, and the area is calculated based on the geographical coordinates of the inflection points after matching the corresponding geometric formula. An area deviation coefficient αm is calculated based on the calculated area and the received area. The area deviation coefficient is the ratio between the absolute value of the difference between the calculated area Ax and the received area As and the smaller of the two. The specific formula is as follows: When the calculated land area Ax is the same as the received land area As, αm=0. The calculated area deviation coefficient is compared with the preset area deviation upper limit. If the calculated area deviation coefficient is greater than the area deviation upper limit, the land area correction task is triggered. Otherwise, the land area correction task is not triggered, and the land boundary point correction stage is entered. B3. Extract the first boundary point to form a point set. Taking the first boundary point numbered b1 as the starting point, calculate the straight-line distance D1 between the first boundary point numbered b1 and the first boundary point numbered b2 based on the point coordinates and the Euclidean distance formula. At the same time, obtain the boundary line distance L1 between b1 and b2. Calculate L1 / D1 and compare the calculation result with the preset value. If the calculated value is greater than the preset value, it is determined that a new boundary point needs to be inserted between b1 and b2. Otherwise, it is not necessary to insert a new boundary point. Continue to determine the insertion of a new boundary point between the first boundary point numbered b2 and the first boundary point numbered b3. B4. When it is determined that a new boundary point needs to be inserted between b1 and b2, find the point with the farthest distance between the boundary lines of b1 and b2 and the straight line connecting b1 and b2, and insert it between b1 and b2 as the new boundary point. Mark the new boundary point as the first boundary point and number it b2. The original first boundary point numbers b2, b3, ..., bn are automatically shifted one position to the right and changed to b3, b4, ..., bn+1. B5. Repeat steps B3 and B4 until the ratio of the boundary line distance to the straight line distance of any adjacent first boundary point is less than or equal to the preset value. Then the point set correction of the first boundary point is completed.

5. The method for automatically generating cadastral maps based on AI according to claim 1, characterized in that: The specific steps for analyzing the boundaries of the land parcel are as follows: C1. First, divide the boundary line into four parts: east, south, west, and north using the method of right rectangular side projection; C2. For all neighboring parcels of this parcel, if a neighboring parcel has a common edge with this parcel in any of the four directions of this parcel (east, south, west, north), then the neighboring parcel is included in the boundary information of the corresponding direction of this parcel. If a neighboring parcel has a common edge with multiple parts of this parcel, then the neighboring parcel is included in the boundary information of the corresponding multiple directions of this parcel.

6. The method for automatically generating cadastral maps based on AI according to claim 1, characterized in that: The specific steps for evaluating the quality of the parcel map generation are as follows: E1. Compare the corrected coordinates Pai(xai, yai) of the i-th first boundary point with the corresponding true coordinates Pbi(xbi, ybi) of the i-th first boundary point, and calculate the position deviation coefficient αwi of the i-th first boundary point. The specific formula is as follows: The average position deviation coefficient αwe is obtained by averaging the calculated position deviation coefficients. E2. The parcel map production accuracy coefficient βe can be obtained by calculating the ratio of the correct text label quantity nax to the sum of the correct text label quantity nay, the correct map feature quantity nbx to the sum of the correct map feature quantity nby, and the correct parcel boundary length nxx to the sum of the correct parcel boundary length ncy. The specific formula is as follows: ; E3. The accuracy coefficient γe for parcel geographic data identification and labeling can be obtained by calculating the ratio of the number of correctly identified and labeled parcel features (nex) to the total number of parcel features (nz). The specific formula is as follows: ; E4. The higher the parcel map production accuracy coefficient and the parcel geographic data identification and annotation accuracy coefficient, the higher the parcel map generation quality. The lower the average location deviation coefficient and the area deviation coefficient, the higher the parcel map generation quality. Calculate the product of the parcel map production accuracy coefficient and the parcel geographic data identification and annotation accuracy coefficient, and calculate the product of the average location deviation coefficient compensation processing result and the area deviation coefficient compensation processing result. The ratio of these two is the parcel map generation quality index Rz. The specific formula is as follows: .

7. An AI-based automatic cadastral map generation system, implementing the AI-based automatic cadastral map generation method as described in any one of claims 1-6, characterized in that: include: Land parcel data acquisition module: Uses radio frequency transmitters to collect land parcel ownership information and uses drones to acquire land parcel geographic data; Land ownership information correction module: used to correct the land boundary lines, land area and land boundary points in the collected land ownership information; Parcel Boundary Analysis Module: The boundary line is divided into four parts (east, south, west, and north) using the method of rectangular edge projection, and the parcel boundary information is obtained based on the set boundary determination rules; Parcel geographic data identification module: Based on AI algorithms, it identifies and labels the collected parcel geographic data; Categorical Map Template Creation Module: Creates categorical map templates using ArcGIS; Parcel map auto-generation module: Run a Python script in the ArcMap command prompt. The script automatically reads the parcel attribute data package and generates a parcel map for each parcel based on the MDX template. Parcel Map Self-Check and Correction Module: Checks whether the parcel boundaries are accurate, whether the text annotations are correct, and whether the map elements are complete. If any one or more of the above conditions are not met, the parcel map needs to be corrected. After the correction is completed, the parcel map is exported as a preset image format. Parcel map generation data acquisition module: used to collect parcel ownership information correction data, parcel map verification and correction data, and parcel geographic data identification data; The parcel map generation quality evaluation module evaluates the quality of parcel map generation after processing the collected parcel ownership information correction data, parcel map verification and correction data, and parcel geographic data identification data.

Images