A method for rapid interpretation and identification of rural houses using remote sensing

Abstract

In response to the limitations of the existing technology, the present invention proposes a method for rapid interpretation and identification of rural houses using remote sensing. The solution of the present invention adopts a two-step interpretation approach, screening effective images based on the location of residential areas and then interpreting them, eliminating most of the background, reducing prediction costs, and improving the accuracy and efficiency of building detection. It effectively reduces the workload of high-resolution remote sensing object identification at a large regional scale. The more dispersed the residential areas are, the more significantly the efficiency can be improved: the prediction amount in the eastern plains is about 1 / 2 of the full prediction; in the central and eastern mountainous and hilly areas, the prediction amount is about 1 / 5 of the full prediction; and in the western region, where residential areas are extremely dispersed, the prediction amount can be reduced to 1 / 100 of the full prediction.

Description

Technical Field

[0001] The present invention relates to the technical field of remote sensing image recognition, in particular to the recognition of buildings in rural areas using remote sensing images, and more particularly to a method for rapid remote sensing interpretation and recognition of farmhouses. Background Art

[0002] Since the development of remote sensing technology, building Figure 1 This has always been a key research direction. my country is experiencing rapid urbanization, and the overall urban-rural landscape is undergoing a rapid transformation. Rural housing provides a crucial breakthrough point for monitoring rural development. The establishment and updating of a rural housing geospatial database is crucial. Traditional field survey methods require significant labor and time, especially in remote areas. The development of high-resolution imagery and the application of deep learning in remote sensing have opened up new avenues for interpreting rural architecture, specifically rural housing.

[0003] The Chinese invention patent with an announcement date of June 19, 2013, is a method for generating multifunctional urban land spatial information using high-resolution remote sensing. The method introduces the theory of complex system hierarchy and proposes a spatial information classification system for urban land based on three scales: landscape type, functional area type, and land cover type, which are multifunctional and adaptive to urban planning management and environmental improvement. Based on the precise correction and registration of Landsat TM, Google Earth, and auxiliary map remote sensing images, Landsat TM is used to classify urban landscape types, and a three-level classification type merging and combination and information mining knowledge base is constructed. The classification information is then merged to form a first-level classification result for urban land. Under the constraints of the upper-level classification information, a second-level digital functional area classification and a third-level land cover classification are carried out.

[0004] However, these existing technologies are primarily designed for urban areas, while my country is vast and rural settlements are sparse. Compared to urban areas, rural settlements are more dispersed, encompassing vast areas of farmland and forestland, with construction occupying only a very small portion. Interpreting full-frame imagery is costly. Furthermore, the identification of rural housing targets is susceptible to interference from background features, such as farmland, forestland, and roads. Therefore, existing technologies still have limitations in interpreting rural housing nationwide. Summary of the Invention

[0005] In view of the limitations of the existing technology, the present invention proposes a method for rapid interpretation and identification of farmhouses using remote sensing data. The technical solution adopted by the present invention is:

[0006] A method for rapid interpretation and identification of farmhouses by remote sensing includes the following steps:

[0007] S1, obtain satellite images and rural settlement reference data;

[0008] S2, determining an image window where a rural settlement may exist from the satellite image according to the rural settlement reference data;

[0009] S3, obtaining an image slice to be predicted from the satellite image map using the image window;

[0010] S4, using a preset farmhouse interpretation model to identify farmhouses on the image slices.

[0011] Compared with the existing technology, the solution of the present invention adopts a two-step interpretation approach, screening effective images based on the location of settlements and then interpreting them, eliminating most of the background, reducing prediction costs, and improving the accuracy and efficiency of building detection; effectively reducing the workload of high-resolution remote sensing object identification at a large regional scale; and the more dispersed the settlements, the more significantly the efficiency can be improved: the prediction amount in the eastern plains is about 1 / 2 of the full prediction; in the central and eastern mountainous and hilly areas, the prediction amount is about 1 / 5 of the full prediction; the settlements in the western region are extremely dispersed, and the prediction amount can be reduced to 1 / 100 of the full prediction.

[0012] As a preferred solution, in step S2, the image window where rural settlements may be located is determined by:

[0013] A vector grid is constructed according to a preset spatial range, and the vector grid is used as a spatial index of the image slice. In combination with the rural settlement reference data, an image window where a rural settlement may exist is determined from the satellite image.

[0014] As a preferred solution, step S2 includes the following process:

[0015] S21, dividing the satellite image into a spatial vector grid Grid1 according to the pixel size requirement of the farmhouse interpretation model for the input image;

[0016] S22, generating a spatial vector grid Grid2 consistent with a preset spatial range of the image window based on the spatial vector grid Grid1;

[0017] S23, combining the spatial position relationship corresponding to the rural residential point reference data, screening out the spatial vector grid Grid3 where the rural residential points may be located from the spatial vector grid Grid2;

[0018] S24, obtaining an image window where rural settlements may exist according to the spatial range of the spatial vector grid Grid3.

[0019] Furthermore, the satellite imagery is an 18-level image with a resolution of 11651*7767 pixels and a scale of 1:10,000.

[0020] Under the condition that the farmhouse interpretation model requires the pixel size of the input image to be 512*512 pixels, in step S21, the spatial vector grid Grid1 of each satellite image includes 368 small windows of 23*16 pixels.

[0021] As a preferred solution, the rural settlement reference data includes impervious surface data and rural POI data.

[0022] As a preferred solution, the method further comprises the following steps:

[0023] S5, according to the spatial information of the satellite image, assigning spatial coordinates to the farmhouse mask identified in step S4 and performing vectorization.

[0024] Furthermore, the method further includes the following steps:

[0025] S6, merging the results obtained from step S5 to form a farmhouse space database.

[0026] The present invention also includes the following contents:

[0027] A rural house remote sensing rapid interpretation and identification system includes a data acquisition module, an image window determination module, an image slice acquisition module, and a model interpretation module; the data acquisition module is connected to the image window determination module and the image slice acquisition module; the image window determination module is connected to the image slice acquisition module; the image slice acquisition module is connected to the model interpretation module; wherein:

[0028] The data acquisition module is used to acquire satellite imagery and rural settlement reference data;

[0029] The image window determination module is used to determine an image window where a rural settlement may exist from the satellite image according to the rural settlement reference data;

[0030] The image slice acquisition module is used to obtain the image slice to be predicted from the satellite image map using the image window;

[0031] The model interpretation module is used to use a preset farmhouse interpretation model to identify farmhouses in the image slices.

[0032] Compared with the existing technology, the solution of the present invention adopts a two-step interpretation approach, screening effective images based on the location of settlements and then interpreting them, eliminating most of the background, reducing prediction costs, and improving the accuracy and efficiency of building detection; effectively reducing the workload of high-resolution remote sensing object identification at a large regional scale; and the more dispersed the settlements, the more significantly the efficiency can be improved: the prediction amount in the eastern plains is about 1 / 2 of the full prediction; in the central and eastern mountainous and hilly areas, the prediction amount is about 1 / 5 of the full prediction; the settlements in the western region are extremely dispersed, and the prediction amount can be reduced to 1 / 100 of the full prediction.

[0033] As a preferred solution, it further includes a regularized vectorization module, wherein the regularized vectorization module is connected to the model interpretation module;

[0034] The regularized vectorization module is used to assign spatial coordinates to the farmhouse mask identified by the model interpretation module according to the spatial information of the satellite image, and perform vectorization.

[0035] Furthermore, it also includes a database generation module, which is connected to the regularization vectorization module:

[0036] The database generation module is used to merge the results obtained by the regularization vectorization module to form a farmhouse space database.

[0037] A storage medium stores a computer program, which, when executed by a processor, implements the steps of the aforementioned method for rapid interpretation and identification of farmhouses via remote sensing.

[0038] A computer device includes a storage medium, a processor, and a computer program stored in the storage medium and executable by the processor. When the computer program is executed by the processor, the steps of the above-mentioned method for rapid interpretation and identification of rural houses by remote sensing are implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A schematic diagram of a flow chart of a method for rapid interpretation and identification of farmhouses using remote sensing technology provided in Example 1 of the present invention;

[0040] Figure 2 Schematic diagram of the principle of the method for rapid interpretation and identification of farmhouses using remote sensing provided in Example 1 of the present invention;

[0041] Figure 3 This is a flow chart of step S2 of the method for rapid interpretation and identification of farmhouses using remote sensing provided in Example 1 of the present invention;

[0042] Figure 4 Schematic diagram of the principle of step S2 of the method for rapid interpretation and identification of farmhouses using remote sensing provided in Example 1 of the present invention;

[0043] Figure 5 This is the distribution map of impervious data in Yunfu City, Guangdong Province;

[0044] Figure 6 This is the distribution map of POI data in Yunfu City, Guangdong Province;

[0045] Figure 7 A schematic flow chart of another method for rapid interpretation and identification of farmhouses using remote sensing technology provided in Example 1 of the present invention;

[0046] Figure 8 A schematic diagram of a system for rapid interpretation and identification of farmhouses using remote sensing technology provided in Example 2 of the present invention;

[0047] Figure 9 This is a schematic diagram of another system for rapid interpretation and identification of farmhouses using remote sensing technology provided in Example 2 of the present invention. DETAILED DESCRIPTION

[0048] The accompanying drawings are for illustrative purposes only and are not to be construed as limiting this patent;

[0049] It should be clear that the embodiments described are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the embodiments of the present application.

[0050] The terms used in the embodiments of the present application are for the purpose of describing specific embodiments only and are not intended to limit the embodiments of the present application. The singular forms "a," "the," and "the" used in the embodiments of the present application and the appended claims are also intended to include plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0051] When the following description refers to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims. In the description of the present application, it should be understood that the terms "first", "second", "third", etc. are only used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence, nor can they be understood as indicating or implying relative importance. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to the specific circumstances.

[0052] In addition, in the description of this application, unless otherwise specified, "plurality" refers to two or more. "And / or" describes the relationship between associated objects, indicating that three relationships can exist. For example, "A and / or B" can mean: A exists alone, A and B exist simultaneously, and B exists alone. The character " / " generally indicates that the associated objects are in an "or" relationship. The present invention is further described below with reference to the accompanying drawings and examples.

[0053] In order to solve the limitations of the prior art, this embodiment provides a technical solution, which will be further described below in conjunction with the accompanying drawings and embodiments.

[0054] Example 1

[0055] A method for rapid interpretation and identification of farmhouses via remote sensing. Figure 1 as well as Figure 2 , including the following steps:

[0056] S1, obtain satellite images and rural settlement reference data;

[0057] S2, determining an image window where a rural settlement may exist from the satellite image according to the rural settlement reference data;

[0058] S3, obtaining an image slice to be predicted from the satellite image map using the image window;

[0059] S4, using a preset farmhouse interpretation model to identify farmhouses on the image slices.

[0060] Compared with the existing technology, this embodiment adopts a two-step interpretation approach, screening effective images based on the location of settlements and then interpreting them, eliminating most of the background, reducing prediction costs, and improving the accuracy and efficiency of building detection; effectively reducing the workload of high-resolution remote sensing object identification at a large regional scale; and the more dispersed the settlements, the more significantly the efficiency can be improved: the prediction amount in the eastern plains is about 1 / 2 of the full prediction; in the central and eastern mountainous and hilly areas, the prediction amount is about 1 / 5 of the full prediction; the settlements in the western region are extremely dispersed, and the prediction amount can be reduced to 1 / 100 of the full prediction.

[0061] Specifically, the farmhouse interpretation model is an existing research result. Using the Mask R-CNN deep learning framework and sub-meter remote sensing imagery, based on the humanistic and geographical characteristics of farmhouses in various provinces in China, a farmhouse remote sensing interpretation model suitable for local areas was trained by province, including a national model and a provincial model. For details, please refer to the paper of the same inventors as this case (Li Xuan, Xu Weipan, Huang Yaofu, Chen Haohui, Qin Xiaozhen, Li Ying, Deng Mingliang, Jiang Junhao, Qin Yawen. Analysis of the spatial distribution characteristics of farmhouses in China based on remote sensing interpretation [J]. Acta Geographica Sinica, 2022, 77(4): 835-851.). The farmhouse remote sensing rapid interpretation and identification method proposed in this case can be regarded as a method innovation based on this farmhouse interpretation model.

[0062] As a preferred embodiment, in step S2, the image window where the rural settlement may be located is determined by:

[0063] A vector grid is constructed according to a preset spatial range, and the vector grid is used as a spatial index of the image slice. In combination with the rural settlement reference data, an image window where a rural settlement may exist is determined from the satellite image.

[0064] As a preferred embodiment, in step S2, please refer to Figure 3 as well as Figure 4 , including the following processes:

[0065] S21, dividing the satellite image into a spatial vector grid Grid1 according to the pixel size requirement of the farmhouse interpretation model for the input image;

[0066] S22, generating a spatial vector grid Grid2 consistent with a preset spatial range of the image window based on the spatial vector grid Grid1;

[0067] S23, combining the spatial position relationship corresponding to the rural residential point reference data, screening out the spatial vector grid Grid3 where the rural residential points may be located from the spatial vector grid Grid2;

[0068] S24, obtaining an image window where rural settlements may exist according to the spatial range of the spatial vector grid Grid3.

[0069] Furthermore, the satellite imagery is an 18-level image with a resolution of 11651*7767 pixels and a scale of 1:10,000.

[0070] Under the condition that the farmhouse interpretation model requires the pixel size of the input image to be 512*512 pixels, in step S21, the spatial vector grid Grid1 of each satellite image includes 368 small windows of 23*16 pixels.

[0071] Specifically, satellite imagery Figure 1 It is generally stored in the form of slices, with a scale ranging from 1:1 million to 1:5000. Since the degree of concentration of settlements in each satellite image is different, there are approximately 0-200 image windows in each satellite image where rural settlements may exist. This embodiment uses 1:10,000 18-level images as the data source (11651*7767 pixels). According to the image input requirements of the rural house interpretation model (512*512 pixels), a 1:10,000 image will be divided into 368 (23*16) small windows. Based on the 1:10,000 spatial grid Grid1, a grid Grid2 is generated that is consistent with the spatial range of the image window, and each grid in Grid2 corresponds to a window of an image.

[0072] As an optional embodiment, Grid1, Grid2, and Grid3 are all stored in shp format.

[0073] As a preferred embodiment, the rural settlement reference data includes impervious surface data and rural POI data.

[0074] Specifically, impervious surface is a data indicator used to describe settlements. It reflects the surface covered by various impervious building materials, such as roofs, roads, and squares made of tiles, asphalt, cement concrete, etc. This example uses the global 10m land cover data released by Professor Gong Peng's team at Tsinghua University, which includes impervious surface data. This example uses the raster extraction tool to extract the DN=80 grid from the land cover data to obtain the impervious surface data. For an example, see Figure 5 , which is the distribution map of impervious data in Yunfu City, Guangdong Province. The white part in the figure is the impervious surface.

[0075] POI (Point of Interest) is a point of interest. In a geographic information system, a POI can be a house, a shop, a mailbox, a bus stop, etc. This example uses the AutoNavi API interface to obtain rural POIs coded as 190108 and 190109 nationwide; AutoNavi Map POIs contain geographic location and attribute category information. Based on the above category codes, a total of 3.646 million rural point data were screened nationwide. For examples, please refer to Figure 6 , which is a distribution map of POI data in Yunfu City, Guangdong Province, where the points in the map are rural POIs.

[0076] Therefore, by fusing multi-source data and using grid positioning and data retrieval in the remote sensing interpretation process, data can be better queried, screened, and retrieved.

[0077] As a preferred embodiment, please refer to Figure 7, further comprising the following steps:

[0078] S5, according to the spatial information of the satellite image, assigning spatial coordinates to the farmhouse mask identified in step S4 and performing vectorization.

[0079] Furthermore, the following steps are also included:

[0080] S6, merging the results obtained from step S5 to form a farmhouse space database.

[0081] Example 2

[0082] See also Figure 8 A rural house remote sensing rapid interpretation and identification system includes a data acquisition module 1, an image window determination module 2, an image slice acquisition module 3, and a model interpretation module 4; the data acquisition module 1 is connected to the image window determination module 2 and the image slice acquisition module 3; the image window determination module 2 is connected to the image slice acquisition module 3; the image slice acquisition module 3 is connected to the model interpretation module 4; wherein:

[0083] The data acquisition module 1 is used to acquire satellite images and rural settlement reference data;

[0084] The image window determination module 2 is used to determine the image window where the rural settlement may exist from the satellite image according to the rural settlement reference data;

[0085] The image slice acquisition module 3 is used to obtain the image slice to be predicted from the satellite image map using the image window;

[0086] The model interpretation module 4 is used to use a preset farmhouse interpretation model to identify farmhouses on the image slices.

[0087] Compared with the existing technology, this embodiment adopts a two-step interpretation approach, screening effective images based on the location of settlements and then interpreting them, eliminating most of the background, reducing prediction costs, and improving the accuracy and efficiency of building detection; effectively reducing the workload of high-resolution remote sensing object identification at a large regional scale; and the more dispersed the settlements, the more significantly the efficiency can be improved: the prediction amount in the eastern plains is about 1 / 2 of the full prediction; in the central and eastern mountainous and hilly areas, the prediction amount is about 1 / 5 of the full prediction; the settlements in the western region are extremely dispersed, and the prediction amount can be reduced to 1 / 100 of the full prediction.

[0088] As a preferred embodiment, it further comprises a regularized vectorization module 5, wherein the regularized vectorization module 5 is connected to the model interpretation module 4;

[0089] The regularized vectorization module 5 is used to assign spatial coordinates to the farmhouse mask identified by the model interpretation module 4 according to the spatial information of the satellite image, and perform vectorization.

[0090] For further information, see Figure 9 , further comprising a database generation module 6, wherein the database generation module 6 is connected to the regularization vectorization module 5:

[0091] The database generation module 6 is used to merge the results obtained by the regularization vectorization module 5 to form a farmhouse space database.

[0092] Example 3

[0093] A storage medium stores a computer program, which, when executed by a processor, implements the steps of the method for rapid interpretation and identification of rural houses using remote sensing as described in Example 1.

[0094] Example 4

[0095] A computer device includes a storage medium, a processor, and a computer program stored in the storage medium and executable by the processor. When the computer program is executed by the processor, the steps of the method for rapid interpretation and identification of rural houses via remote sensing are implemented as described in Example 1.

[0096] Obviously, the above embodiments of the present invention are merely examples for the purpose of clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. Those skilled in the art will appreciate that other variations or modifications can be made based on the above description. It is not necessary and impossible to enumerate all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the claims of the present invention.

Claims

1. A method for rapid interpretation and identification of farmhouses using remote sensing, characterized in that: The following steps are involved: S1, obtaining satellite imagery and rural settlement reference data; the satellite imagery is an 18-level image with 11651*7767 pixels and a scale of 1:10,000; the rural settlement reference data includes impervious surface data and rural POI data; S2, determining an image window where a rural settlement may exist from the satellite image based on the rural settlement reference data; determining the image window where a rural settlement may exist by: constructing a vector grid according to a preset spatial range, using the vector grid as a spatial index of an image slice, and combining the rural settlement reference data to determine an image window where a rural settlement may exist from the satellite image; S21, dividing the satellite image into a spatial vector grid Grid1 according to the pixel size requirement of the rural house interpretation model for the input image; under the condition that the pixel size requirement of the rural house interpretation model for the input image is 512*512 pixels, the spatial vector grid Grid1 of each satellite image includes 368 small windows of 23*16 pixels; S22, generating a spatial vector grid Grid2 consistent with a preset spatial range of the image window based on the spatial vector grid Grid1; S23, combining the spatial position relationship corresponding to the rural residential point reference data, screening out the spatial vector grid Grid3 where the rural residential points may be located from the spatial vector grid Grid2; S24, obtaining an image window where rural settlements may exist based on the spatial range of the spatial vector grid Grid3; S3, obtaining an image slice to be predicted from the satellite image map using the image window; S4, using a preset farmhouse interpretation model to identify farmhouses on the image slices; S5, according to the spatial information of the satellite image, assigning spatial coordinates to the farmhouse mask identified in step S4 and performing vectorization.

2. The method for rapid interpretation and identification of rural houses by remote sensing according to claim 1, characterized in that: The following steps are also included: S6, merging the results obtained from step S5 to form a farmhouse space database.

3. A rural house remote sensing rapid interpretation and identification system, characterized by: The system comprises a data acquisition module (1), an image window determination module (2), an image slice acquisition module (3) and a model interpretation module (4); the data acquisition module (1) is connected to the image window determination module (2) and the image slice acquisition module (3); the image window determination module (2) is connected to the image slice acquisition module (3); the image slice acquisition module (3) is connected to the model interpretation module (4); wherein: The data acquisition module (1) is used to acquire satellite imagery and rural settlement reference data; the satellite imagery uses an 18-level image with 11651*7767 pixels and a scale of 1:10,000; the rural settlement reference data includes impervious surface data and rural POI data; The image window determination module (2) is used to determine the image window where the rural residential area may exist from the satellite image map based on the rural residential area reference data; the image window where the rural residential area may exist is determined by the following method: constructing a vector grid based on a preset spatial range, using the vector grid as the spatial index of the image slice, and combining the rural residential area reference data to determine the image window where the rural residential area may exist from the satellite image map; dividing the satellite image map into a spatial vector grid Grid1 based on the pixel size requirement of the rural house interpretation model for the input image; under the condition that the pixel size requirement of the rural house interpretation model for the input image is 512*512 pixels, the spatial vector grid Grid1 of each satellite image map includes 368 small windows of 23*16 pixels respectively; generating a spatial vector grid Grid2 consistent with the preset spatial range of the image window based on the spatial vector grid Grid1; combining the spatial position relationship corresponding to the rural residential area reference data, screening out a spatial vector grid Grid3 where the rural residential area may exist from the spatial vector grid Grid2; and obtaining the image window where the rural residential area may exist based on the spatial range of the spatial vector grid Grid3; The image slice acquisition module (3) is used to obtain the image slice to be predicted from the satellite image map using the image window; The model interpretation module (4) is used to use a preset farmhouse interpretation model to identify farmhouses on the image slices, and to assign spatial coordinates to the identified farmhouse masks based on the spatial information of the satellite image, and to perform vectorization.

4. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for rapid interpretation and identification of rural houses via remote sensing are implemented as described in any one of claims 1 to 2.

5. A computer device, characterized in that: The method comprises a storage medium, a processor, and a computer program stored in the storage medium and executable by the processor. When the computer program is executed by the processor, the steps of the method for rapid interpretation and identification of rural houses via remote sensing are implemented.

Images