A method and system for collecting geographic information for national land space planning

Abstract

The present invention discloses a geographic information collection method and system for national land space planning. The present invention proposes obtaining original geographic space image information of a national land space to be planned, wherein the original geographic space image information includes data to be optimized; calling a trained optimization processing model to optimize the original geographic space image information to obtain optimized geographic space image information, wherein the optimization processing model includes an optimization unit and a verification unit; the trained optimization processing model is obtained by training the optimization processing model through adversarial learning between the optimization unit and the verification unit; data extraction is performed on the optimized geographic space image information to obtain target geographic information. In the present application, accurate geographic information can be accurately obtained according to real-time conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of national land space planning, and in particular to a method and system for collecting geographic information for national land space planning. Background Art

[0002] With the rapid development of society and the acceleration of urbanization, the importance of national land space planning has become increasingly prominent. Reasonable national land space planning can make full use of land resources, optimize spatial layout, and promote sustainable economic and social development. As the basic work of national land space planning, the accuracy and efficiency of geographic information collection have an important impact on the quality of planning.

[0003] Traditional geographic information collection methods mainly rely on manual field surveys and measurements. Although this method can obtain relatively accurate data, due to the full participation of manual labor, the work efficiency is low and the quality of the data is difficult to guarantee.

[0004] Therefore, researching and developing an efficient and accurate method for collecting geographic information has important practical significance and application value. Summary of the invention

[0005] The main purpose of the invention is to provide a method and system for collecting geographic information for national land space planning.

[0006] To achieve the above-mentioned purpose, the present invention proposes a method for collecting geographic information for national land space planning, comprising:

[0007] Acquire original geospatial image information of the area to be planned in the national territory, wherein the original geospatial image information includes data to be optimized;

[0008] Calling the trained optimization processing model to optimize the original geospatial image information to obtain optimized geospatial image information, wherein the optimization processing model includes an optimization unit and a verification unit; the trained optimization processing model is obtained by training the optimization processing model through adversarial learning between the optimization unit and the verification unit;

[0009] Data extraction is performed on the optimized geographic spatial image information to obtain target geographic information.

[0010] In an independent implementation embodiment, the trained optimization processing model is called to optimize the original geospatial image information to obtain optimized geospatial image information, including

[0011] Inputting the original geospatial image information into the optimization unit to optimize the data to be optimized, and obtaining intermediate geospatial image information;

[0012] Integrating the intermediate geospatial image information with the original geospatial image information to obtain optimized geospatial image information;

[0013] The data to be optimized includes any of the following: boundary data between planned areas and unplanned areas, and boundary data of different geographic information.

[0014] In an independently implemented embodiment, before obtaining the original geospatial image information of the area to be planned, the method further includes:

[0015] The optimization processing model is trained by adversarial learning between the optimization unit and the verification unit to obtain the trained optimization processing model;

[0016] In an independently implemented embodiment, the training of the optimization processing model through adversarial learning between the optimization unit and the verification unit to obtain the trained optimization processing model includes:

[0017] Acquire sample geospatial image information for training the optimization process, wherein the sample geospatial image information includes sample geospatial image information containing data to be optimized and sample geospatial image information not containing data to be optimized;

[0018] Inputting the sample geospatial image information into the optimization unit for optimization processing to obtain optimized sample geospatial image information;

[0019] Inputting the optimized sample geospatial image information and a sample geospatial image information that does not contain data to be optimized and is included in the sample geospatial image information into the inspection unit for inspection and processing to obtain an inspection result;

[0020] The optimization unit and the verification unit are trained according to the verification result.

[0021] In an independently implemented embodiment, obtaining sample geospatial image information for training an optimization processing model includes:

[0022] Acquire a data set for training the optimization processing model, wherein the data set for training the optimization processing model includes at least one sample geospatial image information containing data to be optimized and at least one sample geospatial image information without optimized data, wherein the sample geospatial image information containing data to be optimized and the sample geospatial image information without optimized data correspond to each other one by one;

[0023] Any sample geospatial image information containing data to be optimized and its corresponding sample geospatial image information not containing optimized data are selected to form the sample geospatial image information.

[0024] In an independently implemented embodiment, training the optimization unit and the verification unit according to the discrimination result includes:

[0025] Obtaining a training plan for the optimization processing model, and selecting a suitable training method according to the discrimination result;

[0026] The inspection unit is trained to improve the accuracy of discrimination; the optimization unit is trained to improve the accuracy of geospatial image information optimization.

[0027] In an independently implemented embodiment, training the inspection unit to improve the accuracy of discrimination includes:

[0028] Obtaining a global loss function of the inspection unit and current network parameters of the inspection unit;

[0029] The current network parameters of the verification unit are adjusted to reduce the value of the global loss function of the verification unit to train the verification unit.

[0030] In an independently implemented embodiment, training the optimization unit to optimize the accuracy of geospatial image information includes:

[0031] Obtaining a global loss function of the optimization unit and current network parameters of the optimization unit;

[0032] According to the principle of reducing the value of the global loss function of the optimization unit, the discrimination result is back-propagated to adjust the current network parameters of the optimization unit to train the optimization unit.

[0033] In an independently implemented embodiment, extracting data from the optimized geospatial image information to obtain target geographic information includes:

[0034] Segmenting the optimized geospatial image information to obtain a plurality of segmented geospatial image information;

[0035] The multiple segmented geographic space image information are classified to obtain target geographic information.

[0036] In a second aspect, a geographic information collection system for national land space planning includes a processor and a memory that communicate with each other, and the processor is used to read and execute a computer program from the memory to implement the method described in any one of claims 1-9.

[0037] The technical solution of the present invention adopts the method of obtaining the original geospatial image information of the area to be planned in the national land space, wherein the original geospatial image information includes the data to be optimized; calling the trained optimization processing model to optimize the original geospatial image information to obtain the optimized geospatial image information, wherein the optimization processing model includes an optimization unit and a verification unit; the trained optimization processing model is obtained by training the optimization processing model through adversarial learning between the optimization unit and the verification unit; extracting data from the optimized geospatial image information to obtain the target geographic information. In the present application, accurate geographic information can be accurately obtained according to the real-time situation. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a flowchart of a method for collecting geographic information for national land space planning;

[0039] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0040] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] like Figure 1 As shown, Figure 1 The figure is a flow chart of a method for collecting geographic information for national land space planning according to the present invention. Figure 1 As shown, a method for collecting geographic information for land space planning in this embodiment includes the following steps:

[0042] S101: Obtain original geospatial image information of the area to be planned in the national land space, wherein the original geospatial image information includes data to be optimized.

[0043] The original geospatial image information of the area to be planned can be obtained through modern technical means such as remote sensing technology and drone aerial photography, and this application does not limit this. Due to the limitations of camera distance, weather, camera pixels, etc. when drones or remote sensing technologies are used to shoot the area to be planned, some areas of the original geospatial image information are unclear. When the original geospatial image information is subsequently processed, relevant information cannot be accurately identified in this area. It is understandable that as long as the data can affect the subsequent acquisition of geographic information of the area to be planned through geospatial image information, it is referred to as data to be optimized in this application.

[0044] S102: Calling the trained network model to optimize the original geospatial image information to obtain optimized geospatial image information.

[0045] In the application, the network model is a generative adversarial network model, which includes a generative network and a discriminative network. Through adversarial learning between the generative network and the discriminative network, the optimized data can be well output. The generative network is used to optimize the original geospatial image information, that is, to optimize the data to be optimized in the original geospatial image information, and then obtain the optimized geospatial image information. The discriminative network is used to judge the geospatial image information generated by the generative network, to judge whether the original geospatial image information has been well optimized, that is, to judge whether the data to be optimized will affect the geographic information data of the area to be planned that will be obtained through the subsequent geospatial image information.

[0046] S103: extracting data from the optimized geographic space image information to obtain target geographic information.

[0047] It can be understood that the original geospatial image information of the area to be planned in the national land space is obtained, and the original geospatial image information includes the data to be optimized; the trained optimization processing model is called to optimize the original geospatial image information to obtain the optimized geospatial image information, wherein the optimization processing model includes an optimization unit and a verification unit; the trained optimization processing model is obtained by training the optimization processing model through adversarial learning between the optimization unit and the verification unit; data is extracted from the optimized geospatial image information to obtain the target geographic information. In this application, accurate geographic information can be accurately obtained according to the real-time situation.

[0048] Step S102: calling the trained network model to optimize the original geospatial image information to obtain optimized geospatial image information, including inputting the original geospatial image information into the generation network to optimize the data to be optimized to obtain intermediate geospatial image information; integrating the intermediate geospatial image information with the original geospatial image information to obtain optimized geospatial image information; wherein the data to be optimized includes any of the following: boundary data of planned areas and unplanned areas, boundary data of different geographic information. In this solution, the trained network model is called to optimize the original geospatial image information, that is, the generation network is called to optimize the data to be optimized to form intermediate geospatial image information, and then the intermediate geospatial image information is integrated with the original geospatial image information to obtain optimized geospatial image information. That is, in the intermediate geospatial image information, the data to be optimized is mainly processed, but in the processing process, it is likely to process other areas, such as data areas that do not need to be optimized. Therefore, after obtaining the intermediate geospatial image information, this solution merges the intermediate geospatial image information with the original geospatial image information, with the purpose of replacing the optimized data to be optimized in the original data with the optimized data to be optimized. The optimized geospatial image information thus formed, on the one hand, ensures that a large area of ​​the geospatial image information is the original geospatial image information, avoiding distortion of the geospatial image information due to optimization processing, and on the other hand, facilitates the extraction of useful information from the subsequent processing of the geospatial image information. In the present application, the data to be optimized often exist in the boundary areas of the data, such as the boundary areas of different geographic information, and the boundary data of planned areas and unplanned areas. The geographic information can be swamp area information, gravel area information, highway area information, vegetation area information, etc. The boundary area of ​​different geographic information can be the boundary between a swamp area and a gravel area, or the boundary between a gravel area and a vegetation area, and the present application does not impose any restrictions on this. In addition, since there may be a small part of the area to be planned that has been developed, for example, there is a sub-area in the area to be planned that has been planned, and there are houses, public facilities, etc.

[0049] Before step S102: obtaining the original geospatial image information of the area to be planned in the national land space, the method further includes: training the network model through adversarial learning between the generation network and the discrimination network to obtain the trained network model;

[0050] Wherein, the generating network is a shallow neural network.

[0051] In some embodiments, the training method may be: obtaining sample geospatial image information for training a network model, the sample geospatial image information including a sample geospatial image information containing data to be optimized and a sample geospatial image information not containing data to be optimized; inputting the sample geospatial image information into the generation network for optimization processing to obtain optimized sample geospatial image information; inputting the optimized sample geospatial image information and a sample geospatial image information not containing data to be optimized included in the sample geospatial image information into the discrimination network for discrimination processing to obtain a discrimination result; and training the generation network and the discrimination network according to the discrimination result.

[0052] In an independent implementation embodiment, the trained optimization processing model is called to optimize the original geospatial image information to obtain optimized geospatial image information, including

[0053] Inputting the original geospatial image information into the optimization unit to optimize the data to be optimized, and obtaining intermediate geospatial image information;

[0054] Integrating the intermediate geospatial image information with the original geospatial image information to obtain optimized geospatial image information;

[0055] The data to be optimized includes any of the following: boundary data between planned areas and unplanned areas, and boundary data of different geographic information.

[0056] In an independently implemented embodiment, before obtaining the original geospatial image information of the area to be planned, the method further includes:

[0057] The optimization processing model is trained by adversarial learning between the optimization unit and the verification unit to obtain the trained optimization processing model;

[0058] In an independently implemented embodiment, the training of the optimization processing model through adversarial learning between the optimization unit and the verification unit to obtain the trained optimization processing model includes:

[0059] Acquire sample geospatial image information for training the optimization process, wherein the sample geospatial image information includes sample geospatial image information containing data to be optimized and sample geospatial image information not containing data to be optimized;

[0060] Inputting the sample geospatial image information into the optimization unit for optimization processing to obtain optimized sample geospatial image information;

[0061] Inputting the optimized sample geospatial image information and a sample geospatial image information that does not contain data to be optimized and is included in the sample geospatial image information into the inspection unit for inspection and processing to obtain an inspection result;

[0062] The optimization unit and the verification unit are trained according to the verification result.

[0063] In an independently implemented embodiment, obtaining sample geospatial image information for training an optimization processing model includes:

[0064] Acquire a data set for training the optimization processing model, wherein the data set for training the optimization processing model includes at least one sample geospatial image information containing data to be optimized and at least one sample geospatial image information without optimized data, wherein the sample geospatial image information containing data to be optimized and the sample geospatial image information without optimized data correspond to each other one by one;

[0065] Any sample geospatial image information containing data to be optimized and its corresponding sample geospatial image information not containing optimized data are selected to form the sample geospatial image information.

[0066] In an independently implemented embodiment, training the optimization unit and the verification unit according to the discrimination result includes:

[0067] Obtaining a training plan for the optimization processing model, and selecting a suitable training method according to the discrimination result;

[0068] The inspection unit is trained to improve the accuracy of discrimination; the optimization unit is trained to improve the accuracy of geospatial image information optimization.

[0069] In an independently implemented embodiment, training the inspection unit to improve the accuracy of discrimination includes:

[0070] Obtaining a global loss function of the inspection unit and current network parameters of the inspection unit;

[0071] The current network parameters of the verification unit are adjusted to reduce the value of the global loss function of the verification unit to train the verification unit.

[0072] In an independently implemented embodiment, training the optimization unit to optimize the accuracy of geospatial image information includes:

[0073] Obtaining a global loss function of the optimization unit and current network parameters of the optimization unit;

[0074] According to the principle of reducing the value of the global loss function of the optimization unit, the discrimination result is back-propagated to adjust the current network parameters of the optimization unit to train the optimization unit.

[0075] In an independently implemented embodiment, extracting data from the optimized geospatial image information to obtain target geographic information includes:

[0076] Segmenting the optimized geospatial image information to obtain a plurality of segmented geospatial image information;

[0077] The multiple segmented geographic space image information are classified to obtain target geographic information.

[0078] The present invention also proposes a subject two, which includes A and subject one. The specific structure of subject one refers to the above-mentioned embodiment. Since this subject two adopts all the technical solutions of all the above-mentioned embodiments, it has at least all the beneficial effects brought by the technical solutions of the above-mentioned embodiments, which will not be described one by one here.

[0079] The above descriptions are only optional embodiments of the present invention, and are not intended to limit the patent scope of the present invention. All equivalent structural changes made using the contents of the present invention's specification and drawings, or directly / indirectly applied in other related technical fields, are included in the patent protection scope of the present invention.

Claims

1. A method for collecting geographic information for national land space planning, characterized in that: include: Acquire original geospatial image information of the area to be planned in the national territory, wherein the original geospatial image information includes data to be optimized; Calling the trained optimization processing model to optimize the original geospatial image information to obtain optimized geospatial image information, wherein the optimization processing model includes an optimization unit and a verification unit; the trained optimization processing model is obtained by training the optimization processing model through adversarial learning between the optimization unit and the verification unit; Extract data from optimized geospatial image information to obtain target geographic information; The trained optimization processing model is called to optimize the original geospatial image information to obtain optimized geospatial image information, including Inputting the original geospatial image information into the optimization unit to optimize the data to be optimized, and obtaining intermediate geospatial image information; Integrating the intermediate geospatial image information with the original geospatial image information to obtain optimized geospatial image information; The data to be optimized includes any of the following: boundary data of planned areas and unplanned areas, boundary data of different geographic information; Wherein, before obtaining the original geospatial image information of the area to be planned, the method further includes: The optimization processing model is trained by adversarial learning between the optimization unit and the verification unit to obtain the trained optimization processing model; The step of training the optimization processing model through adversarial learning between the optimization unit and the verification unit to obtain the trained optimization processing model includes: Acquire sample geospatial image information for training the optimization process, wherein the sample geospatial image information includes sample geospatial image information containing data to be optimized and sample geospatial image information not containing data to be optimized; Inputting the sample geospatial image information into the optimization unit for optimization processing to obtain optimized sample geospatial image information; Inputting the optimized sample geospatial image information and a sample geospatial image information that does not contain data to be optimized and is included in the sample geospatial image information into the inspection unit for inspection and processing to obtain an inspection result; training the optimization unit and the inspection unit according to the inspection result; The sample geospatial image information used to train the optimization processing model is obtained, including: Acquire a data set for training the optimization processing model, wherein the data set for training the optimization processing model includes at least one sample geospatial image information containing data to be optimized and at least one sample geospatial image information without optimized data, wherein the sample geospatial image information containing data to be optimized and the sample geospatial image information without optimized data correspond to each other one by one; Select any sample geospatial image information containing data to be optimized and its corresponding sample geospatial image information without optimized data to form the sample geospatial image information; The step of training the optimization unit and the inspection unit according to the discrimination result includes: Obtaining a training plan for the optimization processing model, and selecting a suitable training method according to the discrimination result; The inspection unit is trained to improve the accuracy of discrimination; the optimization unit is trained to improve the accuracy of geospatial image information optimization; The testing unit is trained to improve the accuracy of the discrimination, including: Obtaining a global loss function of the inspection unit and current network parameters of the inspection unit; Adjusting the current network parameters of the verification unit to reduce the value of the global loss function of the verification unit to train the verification unit; The optimization unit is trained to optimize the accuracy of geospatial image information, including: Obtaining a global loss function of the optimization unit and current network parameters of the optimization unit; According to the principle of reducing the value of the global loss function of the optimization unit, the discrimination result is back-propagated to adjust the current network parameters of the optimization unit to train the optimization unit; The step of extracting data from the optimized geographic space image information to obtain target geographic information includes: Segmenting the optimized geospatial image information to obtain a plurality of segmented geospatial image information; The multiple segmented geographic space image information are classified to obtain target geographic information.

2. A geographic information collection system for national land space planning, characterized in that: The invention comprises a processor and a memory communicating with each other, wherein the processor is used to read a computer program from the memory and execute the computer program to implement the method according to claim 1.

Images