Clean land identification method and system based on multiple space elements

Abstract

The invention discloses a net land identification method and system based on multiple space elements, and relates to the field of land supervision, and the method comprises the steps: obtaining remote sensing image data, vehicle passing data and land parcel ownership data of a to-be-researched region; based on a convolutional neural network, performing block identification on the remote sensing image data to obtain a remote sensing feature distribution diagram of the to-be-studied area; performing thermodynamic diagram analysis on the vehicle passing data to obtain a road network characteristic distribution diagram of the to-be-researched area, and performing spatial distribution analysis on the land parcel ownership data to obtain a land parcel ownership distribution table of the to-be-researched area; according to the remote sensing feature distribution map, the road network feature distribution map and the land parcel ownership distribution table, constructing a spatio-temporal data expression model of the to-be-researched area; and according to the spatio-temporal data expression model, in combination with a preset net land delivery identification model, identifying a to-be-identified land parcel in the to-be-studied area.

Description

Technical Field

[0001] The present application relates to the field of land supervision, and in particular to a method and system for clear land identification based on multiple spatial elements. Background Art

[0002] Clean land refers to land that meets the following five requirements during the land transfer process: clear land rights; resettlement compensation is in place; there are no legal and economic disputes; the planning conditions such as the location, use nature, and volume ratio of the land are clear; and other basic conditions necessary for construction and development are met. Before the land is transferred, it is beneficial to the rational use of land resources and avoid idle land waste. To this end, it is necessary to identify and monitor the transfer of clean land.

[0003] The existing net land transfer identification and monitoring supervision mainly relies on the relevant materials provided by various departments, combined with field visits, to manually verify the transfer rights and surface conditions of the land involved. This method lacks flexibility, has low efficiency and accuracy, and manual identification and verification wastes manpower and material resources. Therefore, how to improve the efficiency and accuracy of net land transfer identification and monitoring supervision is still a problem that needs to be solved urgently in existing technologies. Summary of the invention

[0004] The present application provides a method and system for clear land identification based on multiple spatial elements to solve the technical problems of low efficiency and accuracy of existing clear land identification supervision.

[0005] According to a first aspect of the embodiments of the present application, a method for identifying a clear area based on multiple spatial elements is provided, comprising:

[0006] Obtain remote sensing image data, vehicle traffic data and land ownership data of the area to be studied;

[0007] Based on a convolutional neural network, the remote sensing image data is segmented and recognized to obtain a remote sensing feature distribution map of the area to be studied;

[0008] Performing a heat map analysis on the vehicle traffic data to obtain a road network characteristic distribution map of the area to be studied, and performing a spatial distribution analysis on the land ownership data to obtain a land ownership distribution table of the area to be studied;

[0009] Constructing a spatiotemporal data expression model for the area to be studied based on the remote sensing feature distribution map, the road network feature distribution map and the land parcel ownership distribution table;

[0010] According to the spatiotemporal data expression model, combined with a preset net land transfer identification model, the land parcels to be identified in the study area are identified; wherein the net land transfer identification model is trained based on historical net land transfer data.

[0011] In this application, first, the remote sensing image data of the area to be studied is block-recognized based on a convolutional neural network to obtain a remote sensing feature distribution map. Then, heat map analysis is performed on the vehicle traffic data of the area to be studied to obtain a road network feature distribution map, and spatial distribution analysis is performed on the land parcel ownership data of the area to be studied to obtain a land parcel ownership distribution table. It can accurately identify and obtain the remote sensing features, road network features, and spatial distribution of land parcel ownership in the area to be studied through convolutional neural network, heat map analysis, and spatial distribution analysis, thereby improving the efficiency of identifying the clean land of the land parcel and further improving the efficiency of clean land identification supervision. Then, a spatio-temporal data expression model is constructed by combining the remote sensing feature distribution map, the road network feature distribution map, and the land parcel ownership distribution table, which can comprehensively describe the information of each land parcel in the area to be studied with three spatial elements, improve the comprehensiveness of land parcel information, and improve the accuracy of clean land identification. Then, by combining the spatio-temporal data expression model and the clean land transfer identification model trained based on historical clean land transfer data to identify the land parcels to be identified, the advantages of the two models can be integrated, the accuracy of clean land identification can be improved, and thus the accuracy of clean land identification supervision can be improved.

[0012] In some embodiments of this application, the block recognition of the remote sensing image data based on the convolutional neural network to obtain a remote sensing feature distribution map of the area to be studied specifically includes:

[0013] Perform image enhancement on the remote sensing image data to obtain a remotely sensed enhanced image;

[0014] Taking the preset land parcel length as the recognition unit length, divide the remotely sensed enhanced image into land parcels to obtain a remotely sensed block image;

[0015] Based on the convolutional neural network, perform block recognition on the remotely sensed block image to obtain a remote sensing feature distribution map.

[0016] This application first performs image enhancement on the remote sensing image data to obtain a remotely sensed enhanced image, and then divides the land parcels with the preset land parcel length to obtain a remotely sensed block image, which can divide the feature recognition into each land parcel, thereby improving the accuracy of the remote sensing feature distribution map obtained by block recognition based on the convolutional neural network.

[0017] In some embodiments of this application, the heat map analysis of the vehicle traffic data to obtain a road network feature distribution map of the area to be studied specifically includes:

[0018] Taking the preset duration as the division unit time, divide the vehicle traffic data into multiple groups of vehicle traffic sub-data;

[0019] Perform heat map analysis on multiple groups of vehicle traffic sub-data respectively to obtain multiple groups of road network sub-feature distribution maps;

[0020] Overlay the multiple groups of road network sub-feature distribution maps to obtain a road network feature distribution map.

[0021] In this application, the vehicle passing data is first divided into multiple groups of vehicle passing sub-data according to time, and then heat map analysis is performed to obtain multiple groups of road network sub-feature distribution maps, and then superimposed to obtain a road network feature distribution map. By dividing the vehicle passing data and performing heat map analysis separately, the accuracy of the heat map analysis can be improved, thereby improving the accuracy of the obtained road network feature distribution map.

[0022] In some embodiments of this application, the spatial distribution analysis of the land parcel ownership data to obtain the land parcel ownership distribution table of the area to be studied specifically includes:

[0023] Based on the land parcels, the land parcel ownership data is divided into multiple groups of land parcel ownership distribution data;

[0024] Perform spatial distribution analysis on the multiple groups of land parcel ownership distribution data to obtain multiple groups of land parcel ownership distribution sub-tables;

[0025] Integrate the multiple groups of land parcel ownership distribution sub-tables to obtain the land parcel ownership distribution table.

[0026] In this application, the land parcel ownership data is first divided into multiple groups of land parcel ownership distribution data according to the land parcels, and then spatial distribution analysis is performed, and then the land parcel ownership distribution table is obtained. By first dividing each land parcel and then performing spatial distribution analysis on it, the ownership spatial distribution of each land parcel can be accurately determined, thus better meeting the current task requirements.

[0027] In some embodiments of this application, the construction of the spatio-temporal data expression model of the area to be studied according to the remote sensing feature distribution map, the road network feature distribution map and the land parcel ownership distribution table specifically includes:

[0028] Based on longitude and latitude, the remote sensing feature distribution map and the road network feature distribution map are fused to obtain a regional feature distribution map;

[0029] According to each land parcel in the land parcel ownership distribution table, the land parcel ownership distribution table is embedded into the regional feature distribution map to obtain the spatio-temporal data expression model.

[0030] In this application, the remote sensing feature distribution map and the road network feature distribution map are first fused based on longitude and latitude, and then the land parcel ownership distribution table is embedded to obtain the spatio-temporal data expression model, which can fuse the spatio-temporal data of the three, comprehensively describe the information of each land parcel in the area to be studied with three spatial elements, and better meet the current task requirements.

[0031] In some embodiments of this application, the identification of the land parcels to be identified in the area to be studied according to the spatio-temporal data expression model and in combination with a preset net land transfer identification model specifically includes:

[0032] Based on the spatio-temporal data expression model, search for the identification information matrix of the plot to be identified;

[0033] Input the identification information matrix into the clean land transfer identification model to obtain the clean land identification result of the plot to be identified.

[0034] This application first searches for the plot to be identified based on the spatio-temporal data expression matrix to obtain the identification information matrix, and then inputs it into the clean land transfer identification model to obtain the clean land identification result. It can integrate the advantages of the two models, identify through comprehensive plot information types, improve the accuracy of clean land identification, and thus improve the accuracy of clean land identification supervision.

[0035] According to the second aspect of the embodiment of the present application, a clean land identification system based on multiple spatial elements is provided, including a data acquisition module, a first processing module, a second processing module, a model construction module, and a plot identification module;

[0036] The data acquisition module is used to acquire remote sensing image data, vehicle traffic data, and plot ownership data of the area to be studied;

[0037] The first processing module is used to perform block identification on the remote sensing image data based on a convolutional neural network to obtain a remote sensing feature distribution map of the area to be studied;

[0038] The second processing module is used to perform heat map analysis on the vehicle traffic data to obtain a road network feature distribution map of the area to be studied, and perform spatial distribution analysis on the plot ownership data to obtain a plot ownership distribution table of the area to be studied;

[0039] The model construction module is used to construct a spatio-temporal data expression model of the area to be studied according to the remote sensing feature distribution map, the road network feature distribution map, and the plot ownership distribution table;

[0040] The plot identification module is used to identify the plot to be identified in the area to be studied according to the spatio-temporal data expression model, in combination with a preset clean land transfer identification model; wherein, the clean land transfer identification model is trained based on historical clean land transfer data.

[0041] In some embodiments of the present application, the first processing module includes an image enhancement unit, a plot division unit, and a block identification unit;

[0042] The image enhancement unit is used to perform image enhancement on the remote sensing image data to obtain a remotely sensed enhanced image;

[0043] The plot division unit is used to divide the remotely sensed enhanced image with a preset plot length as the identification unit length to obtain a remotely sensed segmented image;

[0044] The block recognition unit is configured to perform block recognition on the remote sensing block image based on a convolutional neural network to obtain a remote sensing feature distribution map.

[0045] In some embodiments of the present application, the second processing module includes a traffic division unit, a heat analysis unit, and a feature superposition unit;

[0046] The traffic division unit is configured to divide the vehicle traffic data into multiple groups of vehicle traffic sub-data with a preset duration as the division unit time;

[0047] The heat analysis unit is configured to perform heat map analysis on multiple groups of vehicle traffic sub-data respectively to obtain multiple groups of road network sub-feature distribution maps;

[0048] The feature superposition unit is configured to superpose the multiple groups of road network sub-feature distribution maps to obtain a road network feature distribution map.

[0049] In some embodiments of the present application, the second processing module includes an ownership division unit, a time series analysis unit, and a data integration unit;

[0050] The ownership division unit is configured to divide the land parcel ownership data into multiple groups of land parcel ownership distribution data based on land parcels;

[0051] The time series analysis unit is configured to perform spatial distribution analysis on the multiple groups of land parcel ownership distribution data to obtain multiple groups of land parcel ownership distribution sub-tables;

[0052] The data integration unit is configured to integrate the multiple groups of land parcel ownership distribution sub-tables to obtain a land parcel ownership distribution table.

[0053] In some embodiments of the present application, the model construction module includes a feature fusion unit and a model construction unit;

[0054] The feature fusion unit is configured to fuse the remote sensing feature distribution map and the road network feature distribution map based on longitude and latitude to obtain a regional feature distribution map;

[0055] The model construction unit is configured to embed the land parcel ownership distribution table into the regional feature distribution map according to each land parcel in the land parcel ownership distribution table to obtain a spatio-temporal data expression model.

[0056] In some embodiments of the present application, the land parcel recognition module includes a land parcel search unit and a clean land recognition unit;

[0057] The land parcel search unit is configured to search for an identification information matrix of the land parcel to be recognized based on the spatio-temporal data expression model;

[0058] The clean land identification unit is configured to input the identification information matrix into the clean land transfer identification model to obtain the clean land identification result of the to-be-identified plot.

[0059] In this application, first, the remote sensing image data of the area to be studied is block-identified based on a convolutional neural network to obtain a remote sensing feature distribution map. Then, a heat map analysis is performed on the vehicle traffic data of the area to be studied to obtain a road network feature distribution map, and a spatial distribution analysis is performed on the plot ownership data of the area to be studied to obtain a plot ownership distribution table. It can accurately identify and obtain the remote sensing features, road network features, and plot ownership spatial distribution of the area to be studied through convolutional neural network, heat map analysis, and spatial distribution analysis, thereby improving the clean land identification efficiency of plots and thus improving the efficiency of clean land identification supervision. Then, a spatio-temporal data expression model is constructed by combining the remote sensing feature distribution map, the road network feature distribution map, and the plot ownership distribution table, which can comprehensively describe the information of each plot in the area to be studied with three spatial elements, improve the comprehensiveness of plot information, and improve the accuracy of clean land identification. Then, the spatio-temporal data expression model and the clean land transfer identification model trained based on historical clean land transfer data are combined to identify the to-be-identified plot, which can integrate the advantages of the two models, improve the accuracy of clean land identification, and thus improve the accuracy of clean land identification supervision. Description of the Drawings

[0060] Figure 1 : A flowchart showing a method for identifying clean land based on multiple spatial elements according to some embodiments of the present application;

[0061] Figure 2 : A module structure diagram showing a system for identifying clean land based on multiple spatial elements according to some embodiments of the present application. Detailed Embodiments

[0062] The following details the embodiments of the present application. The examples of the embodiments are shown in the drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions throughout. The embodiments described below by combining the drawings are exemplary and are only used to explain some embodiments of the present application, and cannot be understood as a limitation to the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments shown in the present application without creative efforts shall fall within the protection scope of the present application.

[0063] In the description of the present application, it should be understood that the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include at least one such feature. In the description of the present application, unless otherwise specifically defined, the meaning of "a plurality" or "several" is two or more.

[0064] The existing identification and monitoring and supervision of clean land transfer mainly involve manually verifying the transfer ownership and surface conditions of the land involved by various departments providing corresponding materials and combining on-site visits. This method lacks flexibility, has low efficiency and accuracy, and there is a situation of wasting manpower and material resources in manual identification and verification. Therefore, how to improve the efficiency and accuracy of identifying and monitoring and supervising clean land transfer remains a problem to be solved urgently in the existing technology.

[0065] Based on the above technical background, please refer to Figure 1 , an embodiment of the present application provides a method for identifying clean land based on multiple spatial elements, including steps S101 to step S10a, and the specific steps are as follows:

[0066] Step S101: Obtain remote sensing image data, vehicle passage data, and land parcel ownership data of the area to be studied.

[0067] Step S102: Based on a convolutional neural network, perform block recognition on the remote sensing image data to obtain a remote sensing feature distribution map of the area to be studied.

[0068] In some embodiments of the present application, the performing block recognition on the remote sensing image data based on a convolutional neural network to obtain a remote sensing feature distribution map of the area to be studied specifically includes:

[0069] Perform image enhancement on the remote sensing image data to obtain a remotely sensed enhanced image;

[0070] Divide the remotely sensed enhanced image with a preset land parcel length as the recognition unit length to obtain a remotely sensed segmented image;

[0071] Based on a convolutional neural network, perform block recognition on the remotely sensed segmented image to obtain a remote sensing feature distribution map.

[0072] In some embodiments of the present application, the implementation methods of the image enhancement include, but are not limited to, histogram equalization, gray world algorithm, Retinex algorithm, and automatic color equalization (ACE) algorithm, and the preferred implementation method is the histogram equalization algorithm.

[0073] In some embodiments of the present application, the preferred value of the preset land parcel length is 100m.

[0074] In some embodiments of the present application, the embodiments of the convolutional neural network include, but are not limited to, the R-CNN model and its improved models, the R-FCN model, and the SPP-Net model. The preferred embodiment is the R-CNN model.

[0075] By using a convolutional neural network to identify remote sensing image data, it is possible to identify and obtain the remote sensing feature data of buildings, structures, ponds, and vegetation on the ground surface to form a remote sensing feature distribution map. Generally speaking, the buildings, structures, ponds, and vegetation on the ground surface are closely related to the "surface clearing" work of the corresponding plots, and whether it is a clean plot is also inseparable from the "surface clearing" work. Therefore, forming a remote sensing feature distribution map through the remote sensing feature data of buildings, structures, ponds, and vegetation on the ground surface can provide guarantee for the identification of clean plots.

[0076] In the present application, the remote sensing image data is first subjected to image enhancement to obtain a remote sensing enhanced image, and then the remote sensing image is divided into blocks with a preset plot length to obtain a remote sensing segmented image, which can divide the feature recognition into each plot, thereby improving the accuracy of the remote sensing feature distribution map obtained by block recognition based on the convolutional neural network.

[0077] Step S103: Perform a heat map analysis on the vehicle traffic data to obtain a road network feature distribution map of the area to be studied, and perform a spatial distribution analysis on the plot ownership data to obtain a plot ownership distribution table of the area to be studied.

[0078] In some embodiments of the present application, the performing a heat map analysis on the vehicle traffic data to obtain a road network feature distribution map of the area to be studied specifically includes:

[0079] Taking a preset time period as the unit time for division, dividing the vehicle traffic data into multiple groups of vehicle traffic sub-data;

[0080] Performing a heat map analysis on each group of vehicle traffic sub-data respectively to obtain multiple groups of road network sub-feature distribution maps;

[0081] Overlaying the multiple groups of road network sub-feature distribution maps to obtain a road network feature distribution map.

[0082] In some embodiments of the present application, the preferred value of the preset time period is 24h.

[0083] In the present application, the vehicle traffic data is first divided into multiple groups of vehicle traffic sub-data according to time, then a heat map analysis is performed to obtain multiple groups of road network sub-feature distribution maps, and then they are overlaid to obtain a road network feature distribution map. By dividing the vehicle traffic data and performing a heat map analysis separately, the accuracy of the heat map analysis can be improved, thereby improving the accuracy of the obtained road network feature distribution map.

[0084] By performing heat map analysis on the vehicle passage heat map, a road network feature distribution map can be obtained, which can acquire the access conditions of each plot. The access condition is an important identification factor for a clean plot. If the access condition is poor or there is no access at all, it indicates that the corresponding plot is not a clean plot. Therefore, by performing heat map analysis on the vehicle passage heat map to obtain the road network feature distribution map for studying the access conditions of plots, it can provide a guarantee for clean plot identification.

[0085] In some embodiments of the present application, the spatial distribution analysis of the plot ownership data to obtain the plot ownership distribution table of the area to be studied specifically includes:

[0086] Based on plots as the division basis, the plot ownership data is divided into multiple groups of plot ownership distribution data;

[0087] Perform spatial distribution analysis on the multiple groups of plot ownership distribution data to obtain multiple groups of plot ownership distribution sub-tables;

[0088] Integrate the multiple groups of plot ownership distribution sub-tables to obtain the plot ownership distribution table.

[0089] By performing spatial distribution analysis on the plot ownership data, the current ownership situation of each plot can be obtained. The ownership situation is another important identification factor for a clean plot. If there are ownership problems in the current plot, such as the existence of collective ownership or un-cancelled ownership such as state-owned and collective land use rights in the current plot, it indicates that the current plot is not a clean plot. Therefore, by performing spatial distribution analysis on the plot ownership data to obtain the plot ownership distribution table, it can provide a guarantee for clean plot identification.

[0090] In the present application, the plot ownership data is first divided into multiple groups of plot ownership distribution data according to plots, and then spatial distribution analysis is performed to obtain the plot ownership distribution table. By first dividing each plot and then performing spatial distribution analysis on it, the ownership spatial distribution of each plot can be accurately determined, thus better meeting the current task requirements.

[0091] Step S104: Construct a spatio-temporal data expression model of the area to be studied according to the remote sensing feature distribution map, the road network feature distribution map, and the plot ownership distribution table.

[0092] In some embodiments of the present application, the constructing a spatio-temporal data expression model of the area to be studied according to the remote sensing feature distribution map, the road network feature distribution map, and the plot ownership distribution table specifically includes:

[0093] Based on longitude and latitude, fuse the remote sensing feature distribution map and the road network feature distribution map to obtain a regional feature distribution map;

[0094] According to each plot in the plot ownership distribution table, embed the plot ownership distribution table into the regional feature distribution map to obtain a spatio-temporal data expression model.

[0095] Specifically, when fusing the remote sensing feature distribution map and the road network feature distribution map based on longitude and latitude to obtain the regional feature distribution map, first align each longitude and latitude point on the remote sensing feature distribution map and the road network feature distribution map according to longitude and latitude, and then superimpose their respective information to fuse and obtain the regional feature distribution map.

[0096] Specifically, when embedding the plot ownership distribution table into the regional feature distribution map according to each plot in the plot ownership distribution table to obtain the spatio-temporal data expression model, write the plot ownership distribution table of the plot into all the longitude and latitude points corresponding to a certain plot, so as to embed the plot ownership distribution table into the regional feature distribution map to obtain the spatio-temporal data expression model.

[0097] This application first fuses the remote sensing feature distribution map and the road network feature distribution map based on longitude and latitude, and then embeds the plot ownership distribution table to obtain the spatio-temporal data expression model, which can fuse the spatio-temporal data of the three, and comprehensively describe the information of each plot in the area to be studied with three spatial elements, which better meets the current task requirements.

[0098] Step S105: According to the spatio-temporal data expression model, combined with a preset net land transfer identification model, identify the plots to be identified in the area to be studied; wherein, the net land transfer identification model is trained based on historical net land transfer data.

[0099] In some embodiments of this application, the identifying the plots to be identified in the area to be studied according to the spatio-temporal data expression model, combined with a preset net land transfer identification model, specifically includes:

[0100] Based on the spatio-temporal data expression model, search for and obtain the identification information matrix of the plots to be identified;

[0101] Input the identification information matrix into the net land transfer identification model to obtain the net land identification result of the plots to be identified.

[0102] In some embodiments of this application, the identification information matrix includes the remote sensing features, road network features and plot ownership distribution table of the plots to be identified.

[0103] This application first searches for the plots to be identified based on the spatio-temporal data expression matrix to obtain the identification information matrix, and then inputs it into the net land transfer identification model to obtain the net land identification result, which can combine the advantages of the two models, identify through comprehensive plot information types, improve the accuracy of net land identification, and thus improve the accuracy of net land identification supervision.

[0104] Compared with the prior art, in this application, first, the remote sensing image data of the area to be studied is block-recognized based on a convolutional neural network to obtain a remote sensing feature distribution map. Then, heat map analysis is performed on the vehicle traffic data of the area to be studied to obtain a road network feature distribution map, and spatial distribution analysis is performed on the land parcel ownership data of the area to be studied to obtain a land parcel ownership distribution table. It can accurately identify and obtain the remote sensing features, road network features, and spatial distribution of land parcel ownership in the area to be studied through convolutional neural network, heat map analysis, and spatial distribution analysis, thereby improving the efficiency of identifying clean land for land parcels and further improving the efficiency of clean land identification supervision. Then, a spatio-temporal data expression model is constructed by combining the remote sensing feature distribution map, the road network feature distribution map, and the land parcel ownership distribution table, which can comprehensively describe the information of each land parcel in the area to be studied with three spatial elements, improve the comprehensiveness of land parcel information, and improve the accuracy of clean land identification. Then, by combining the spatio-temporal data expression model and the clean land transfer identification model trained based on historical clean land transfer data to identify the land parcels to be identified, the advantages of the two models can be integrated, the accuracy of clean land identification can be improved, and thus the accuracy of clean land identification supervision can be improved.

[0105] Corresponding to the foregoing method, please refer to Figure 2 , an embodiment of this application provides a clean land identification system based on multiple spatial elements, including a data acquisition module 210, a first processing module 220, a second processing module 230, a model construction module 240, and a land parcel identification module 250;

[0106] The data acquisition module 210 is configured to acquire remote sensing image data, vehicle traffic data, and land parcel ownership data of the area to be studied;

[0107] The first processing module 220 is configured to perform block recognition on the remote sensing image data based on a convolutional neural network to obtain a remote sensing feature distribution map of the area to be studied;

[0108] The second processing module 230 is configured to perform heat map analysis on the vehicle traffic data to obtain a road network feature distribution map of the area to be studied, and perform spatial distribution analysis on the land parcel ownership data to obtain a land parcel ownership distribution table of the area to be studied;

[0109] The model construction module 240 is configured to construct a spatio-temporal data expression model of the area to be studied according to the remote sensing feature distribution map, the road network feature distribution map, and the land parcel ownership distribution table;

[0110] The land parcel identification module 250 is configured to identify the land parcels to be identified in the area to be studied according to the spatio-temporal data expression model in combination with a preset clean land transfer identification model; wherein, the clean land transfer identification model is trained based on historical clean land transfer data.

[0111] In some embodiments of the present application, the first processing module 220 includes an image enhancement unit, a plot division unit, and a block recognition unit;

[0112] The image enhancement unit is configured to perform image enhancement on the remote sensing image data to obtain a remote sensing enhanced image;

[0113] The plot division unit is configured to divide the remote sensing enhanced image with a preset plot length as the recognition unit length to obtain a remote sensing segmented image;

[0114] The block recognition unit is configured to perform block recognition on the remote sensing segmented image based on a convolutional neural network to obtain a remote sensing feature distribution map.

[0115] In some embodiments of the present application, the second processing module 230 includes a traffic division unit, a heat map analysis unit, and a feature superposition unit;

[0116] The traffic division unit is configured to divide the vehicle traffic data into multiple groups of vehicle traffic sub-data with a preset duration as the division unit time;

[0117] The heat map analysis unit is configured to perform heat map analysis on multiple groups of vehicle traffic sub-data respectively to obtain multiple groups of road network sub-feature distribution maps;

[0118] The feature superposition unit is configured to superpose the multiple groups of road network sub-feature distribution maps to obtain a road network feature distribution map.

[0119] In some embodiments of the present application, the second processing module 230 includes a property right division unit, a temporal analysis unit, and a data integration unit;

[0120] The property right division unit is configured to divide the plot property right data into multiple groups of plot property right distribution data based on plots;

[0121] The temporal analysis unit is configured to perform spatial distribution analysis on the multiple groups of plot property right distribution data to obtain multiple groups of plot property right distribution sub-tables;

[0122] The data integration unit is configured to integrate the multiple groups of plot property right distribution sub-tables to obtain a plot property right distribution table.

[0123] In some embodiments of the present application, the model construction module 240 includes a feature fusion unit and a model construction unit;

[0124] The feature fusion unit is configured to fuse the remote sensing feature distribution map and the road network feature distribution map based on longitude and latitude to obtain a regional feature distribution map;

[0125] The model construction unit is configured to embed the land ownership distribution table into the regional feature distribution map according to each plot in the land ownership distribution table, so as to obtain a spatio-temporal data expression model.

[0126] In some embodiments of the present application, the plot identification module 250 includes a plot search unit and a clean land identification unit;

[0127] The plot search unit is configured to search for an identification information matrix of the plot to be identified based on the spatio-temporal data expression model;

[0128] The clean land identification unit is configured to input the identification information matrix into the clean land transfer identification model to obtain a clean land identification result of the plot to be identified.

[0129] In this application, first, the remote sensing image data of the area to be studied is block-identified based on a convolutional neural network to obtain a remote sensing feature distribution map. Then, the heat map analysis of the vehicle traffic data in the area to be studied is performed to obtain a road network feature distribution map, and the spatial distribution analysis of the land ownership data in the area to be studied is performed to obtain a land ownership distribution table. It can accurately identify and obtain the remote sensing features, road network features, and land ownership spatial distribution of the area to be studied through convolutional neural network, heat map analysis, and spatial distribution analysis, thereby improving the efficiency of clean land identification for plots and thus improving the efficiency of clean land identification supervision. Then, by combining the remote sensing feature distribution map, the road network feature distribution map, and the land ownership distribution table to construct a spatio-temporal data expression model, it can comprehensively describe the information of each plot in the area to be studied with three spatial elements, improve the comprehensiveness of plot information, and improve the accuracy of clean land identification. Then, by combining the spatio-temporal data expression model and the clean land transfer identification model trained based on historical clean land transfer data to identify the plot to be identified, it can integrate the advantages of the two models, improve the accuracy of clean land identification, and thus improve the accuracy of clean land identification supervision.

[0130] It should be understood that the system provided by the embodiments of the present application corresponds to the foregoing method. A clean land identification system based on multiple spatial elements provided by the embodiments of the present application can implement a clean land identification method based on multiple spatial elements provided by any embodiment of the present application.

[0131] Adaptively, the embodiments of the present application also provide a computer device and a computer-readable storage medium.

[0132] The computer device includes: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor;

[0133] Wherein, when the processor executes the computer program, it implements a clean land identification method based on multiple spatial elements of the present application.

[0134] The computer-readable storage medium stores multiple instructions, and the instructions are suitable for being loaded by a processor to execute a net land identification method based on multiple spatial elements of the present application.

[0135] The above is part of the embodiments of the present application, and the purpose, technical solutions and beneficial effects of the present application are further described in detail. It should be clear that the above part of the embodiments of the present application should not be construed as a limitation of the present application. In particular, for those skilled in the art, any changes, modifications, equivalent replacements and variations made within the spirit and principle of the present application shall be included within the protection scope of the present application.

Claims

1. A clear land identification method based on multiple spatial elements, characterized in that: include: Obtain remote sensing image data, vehicle traffic data and land ownership data of the area to be studied; Based on a convolutional neural network, the remote sensing image data is segmented and recognized to obtain a remote sensing feature distribution map of the area to be studied; Performing a heat map analysis on the vehicle traffic data to obtain a road network characteristic distribution map of the area to be studied, and performing a spatial distribution analysis on the land ownership data to obtain a land ownership distribution table of the area to be studied; Constructing a spatiotemporal data expression model for the area to be studied based on the remote sensing feature distribution map, the road network feature distribution map and the land parcel ownership distribution table; According to the spatiotemporal data expression model, combined with a preset net land transfer identification model, the land parcels to be identified in the study area are identified; wherein the net land transfer identification model is trained based on historical net land transfer data.

2. A method for clear land identification based on multiple spatial elements according to claim 1, characterized in that: The remote sensing image data is segmented and identified based on a convolutional neural network to obtain a remote sensing feature distribution map of the area to be studied, specifically including: Performing image enhancement on the remote sensing image data to obtain a remote sensing enhanced image; Taking a preset plot length as an identification unit length, dividing the remote sensing enhanced image into plots to obtain a remote sensing block image; Based on a convolutional neural network, the remote sensing block image is subjected to block recognition to obtain a remote sensing feature distribution map.

3. The method for clear land identification based on multiple spatial elements according to claim 1, characterized in that: The heat map analysis of the vehicle traffic data to obtain a road network characteristic distribution map of the area to be studied specifically includes: Taking a preset time length as a division unit time, dividing the vehicle traffic data into a plurality of groups of vehicle traffic sub-data; Conduct heat map analysis on multiple groups of vehicle traffic sub-data respectively to obtain multiple groups of road network sub-feature distribution maps; The multiple groups of road network sub-feature distribution maps are superimposed to obtain a road network feature distribution map.

4. The method for clear land identification based on multiple spatial elements according to claim 1, characterized in that: The spatial distribution analysis of the land ownership data is performed to obtain a land ownership distribution table of the area to be studied, specifically including: Based on the land parcels, the land parcel ownership data is divided to obtain multiple groups of land parcel ownership distribution data; Performing spatial distribution analysis on the multiple groups of land parcel ownership distribution data to obtain multiple groups of land parcel ownership distribution sub-tables; The multiple groups of land parcel ownership distribution sub-tables are integrated to obtain a land parcel ownership distribution table.

5. The method for clear land identification based on multiple spatial elements according to claim 1, characterized in that: The step of constructing a spatiotemporal data expression model for the area to be studied based on the remote sensing feature distribution map, the road network feature distribution map and the land parcel ownership distribution table specifically includes: Based on longitude and latitude, the remote sensing feature distribution map and the road network feature distribution map are merged to obtain a regional feature distribution map; According to each plot in the plot ownership distribution table, the plot ownership distribution table is embedded in the regional characteristic distribution map to obtain a spatiotemporal data expression model.

6. The method for clear land identification based on multiple spatial elements according to claim 1, characterized in that: The identifying of the land parcels to be identified in the study area according to the spatiotemporal data expression model and in combination with the preset net land transfer identification model specifically includes: Based on the spatiotemporal data expression model, searching and obtaining an identification information matrix of the land parcel to be identified; The identification information matrix is ​​input into the net land transfer identification model to obtain the net land identification result of the land parcel to be identified.

7. A clear land identification system based on multiple spatial elements, characterized in that: It includes a data acquisition module, a first processing module, a second processing module, a model building module and a land parcel identification module; The data acquisition module is used to acquire remote sensing image data, vehicle traffic data and land ownership data of the area to be studied; The first processing module is used to perform block recognition on the remote sensing image data based on a convolutional neural network to obtain a remote sensing feature distribution map of the area to be studied; The second processing module is used to perform a heat map analysis on the vehicle traffic data to obtain a road network characteristic distribution map of the area to be studied, and to perform a spatial distribution analysis on the land ownership data to obtain a land ownership distribution table of the area to be studied; The model building module is used to build a spatiotemporal data expression model of the area to be studied based on the remote sensing feature distribution map, the road network feature distribution map and the land ownership distribution table; The land parcel identification module is used to identify the land parcels to be identified in the study area according to the spatiotemporal data expression model and in combination with a preset net land transfer identification model; wherein the net land transfer identification model is trained based on historical net land transfer data.

8. The clear land identification system based on multiple spatial elements according to claim 7 is characterized in that: The first processing module includes an image enhancement unit, a land parcel division unit and a parcel identification unit; The image enhancement unit is used to perform image enhancement on the remote sensing image data to obtain a remote sensing enhanced image; The plot division unit is used to divide the remote sensing enhanced image into plots using a preset plot length as an identification unit length to obtain a remote sensing block image; The block recognition unit is used to perform block recognition on the remote sensing block image based on a convolutional neural network to obtain a remote sensing feature distribution map.

9. The clear land identification system based on multiple spatial elements according to claim 7, characterized in that: The model building module includes a feature fusion unit and a model building unit; The feature fusion unit is used to fuse the remote sensing feature distribution map and the road network feature distribution map based on longitude and latitude to obtain a regional feature distribution map; The model building unit is used to embed the land parcel ownership distribution table into the regional characteristic distribution map according to each land parcel in the land parcel ownership distribution table to obtain a spatiotemporal data expression model.

10. The clear land identification system based on multiple spatial elements according to claim 7, characterized in that: The land parcel identification module includes a land parcel search unit and a clear land identification unit; The land parcel search unit is used to search for an identification information matrix of a land parcel to be identified based on the spatiotemporal data expression model; The clean land identification unit is used to input the identification information matrix into the clean land transfer identification model to obtain the clean land identification result of the land parcel to be identified.

Images