An Object-Oriented Geospatial Big Data Aggregation Method

Abstract

The present invention discloses an object-oriented geospatial big data aggregation method. Based on technologies such as object-oriented image segmentation, spatial association, and mode filtering, it can greatly simplify the relatively cumbersome processing and parameter selection process of existing spatio-temporal aggregation methods for aggregating multi-source, multi-scale, and multi-format data, improve the aggregation efficiency of geospatial big data, avoid the problem of different users selecting different scales in different scenarios, construct the spatial topological relationship of multi-source data sets, and thus can effectively extract the effective information of multi-source and multi-scale geospatial big data element layers.

Description

Technical Field

[0001] The present invention belongs to the technical field of geographic information analysis, and in particular relates to the design of an object-oriented geospatial big data aggregation method. Background Art

[0002] Geospatial data is the operating object of the geographic information system (GIS). It is the substantial content of the real world expressed by GIS after model abstraction. Geospatial data essentially refers to data that describes the natural, social, and cultural and economic landscapes with the spatial position on the earth's surface as a reference, mainly including numbers, text, graphics, images, and tables. In recent years, the development of technologies such as aerial remote sensing, ground sensors, communications, and the Internet has derived more data related to geographic location, further enriching the types, content, and dimensions of geospatial data. Geospatial big data has significantly improved the service capabilities of geographic information, significantly improved the accuracy of geographic information, and shortened the response time of geographic services. It can provide better data support for urban intelligence, personal location services, land and resources assessment, and ecological environment monitoring, and will provide users with more application value and service capabilities.

[0003] Since the generation of geospatial big data is not for a consistent purpose, its data sources, in addition to earth observation data of different spatial resolutions provided by different types of satellites, are mostly spatial location information left by objects during movement and change. Therefore, geospatial big data has the characteristics of a wide variety, different standards, multi-scale, multi-granularity, and multi-modality (format). As the premise and foundation of geospatial big data research, geospatial big data aggregation can integrate multi-source information from different aspects when facing more complex geographical problems, and ultimately achieve research goals such as data mining and information extraction.

[0004] Spatiotemporal aggregation is currently the most commonly used method for aggregating geospatial big data. Its basic principle is to aggregate big data that meets a certain spatiotemporal relationship based on location. In recent years, key spatiotemporal aggregation technologies such as raster computing, scale conversion, location association, and topological association have made certain progress, but most current studies still use regular grids to divide the study area, and there is no unified specification and standard for the aggregation scale, which leads to the selection of different spatial unit sizes affecting the evaluation of geospatial information and patterns. Traditional analysis methods mainly rely on statistical analysis and cluster analysis, which are limited to short-range relationships and limited-order patterns. In order to better analyze and select scales, more patterns need to be introduced to describe the overall picture of the data. It can be seen that how to choose an appropriate research scenario scale has become a difficult problem that currently constrains the aggregation and application research of geospatial big data. Summary of the invention

[0005] The object of the present invention is to solve the problems of the existing spatio-temporal aggregation method, such as the non-uniform standards for multi-source, multi-scale, and multi-format geospatial data, the cumbersome processing process, and the low aggregation efficiency. An object-oriented geospatial big data aggregation method is proposed.

[0006] The technical solution of the present invention is as follows: An object-oriented geospatial big data aggregation method, comprising the following steps:

[0007] S1. Select the remotely sensed image with the most appropriate spatial scale from the GIS data according to the specific research scenario of the user.

[0008] S2. Use the object-oriented segmentation technology, select the aggregation scale through the scale selection algorithm, and generate the geographical monomer feature set.

[0009] S3. Perform object-oriented land cover type interpretation on the remotely sensed image to obtain the raster data of the natural surface type with clear geographical boundaries and categories.

[0010] S4. Integrate the multi-source, multi-scale, and multi-format GIS data, and construct the corresponding geospatial data set in combination with the raster data of the natural surface type.

[0011] S5. Use the aggregation method of spatial matching and spatial filtering technology to extract the effective information of different spatial scale feature layers of the geospatial data set, and combine with the geographical monomer feature set to obtain the geospatial big data set.

[0012] Further, step S2 includes the following sub-steps:

[0013] S21. Test the road patches within the range of the remotely sensed image.

[0014] S22. When the interior of the road patch unit is uniform and there are differences between the road and the adjacent ground feature patch units, use the corresponding segmentation scale as the aggregation scale.

[0015] S23. Generate the geographical monomer feature set according to the aggregation scale.

[0016] Further, step S3 includes the following sub-steps:

[0017] S31. Use the vegetation index, water body index, building index, and point of interest as indicators, and adopt the decision tree classification method to divide the first-level classes.

[0018] S32. Adopt the maximum likelihood method based on samples to divide the second-level classes, and ensure that the number of samples for each second-level class is greater than 30.

[0019] S33. Interpret the remotely sensed image through the first-level classes and the second-level classes to obtain the raster data of the natural surface type with clear geographical boundaries and categories.

[0020] Furthermore, step S4 includes the following sub-steps:

[0021] S41. Perform spatialization processing on the digital statistical GIS data according to the spatial position attributes of the data and their corresponding spatial distribution characteristics to obtain digital statistical vector data.

[0022] S42. Perform resampling interpolation processing on the digital statistical vector data, the thematic feature map, the zoning boundary, and the grading map in the GIS data to obtain resampled interpolation raster data.

[0023] S43. Perform projection conversion processing on the natural land surface type raster data and the resampled interpolation raster data, and use spatial association technology to spatially integrate and inherit various raster data in a hierarchical form to construct a geospatial data set.

[0024] Furthermore, step S5 includes the following sub-steps:

[0025] S51. Sort the element layers in the geospatial data set according to the spatial resolution size, and the spatial resolution of the data in each layer from top to bottom increases from small to large.

[0026] S52. Use the geographical monomer elements in the geographical monomer element set as the minimum processing unit, and adopt the object-oriented mode filtering method to perform information aggregation on the element layers in the geospatial data set to obtain a geospatial big data set.

[0027] The beneficial effects of the present invention are as follows:

[0028] (1) Based on technologies such as object-oriented image segmentation, spatial association, and mode filtering, the present invention can greatly simplify the relatively cumbersome processing and parameter selection processes of existing spatio-temporal aggregation methods for aggregating multi-source, multi-scale, and multi-format data, improve the aggregation efficiency of geospatial big data, avoid the problem of different users selecting different scales in different scenarios, construct the spatial topological relationship of the multi-source data set, and thus can effectively extract the effective information of the element layers of multi-source and multi-scale geospatial big data.

[0029] (2) The present invention can quickly confirm the appropriate scale for geospatial big data aggregation, and can intelligently adjust the appropriate information aggregation scale according to the characteristics of a specific research object.

[0030] (3) The present invention can form the smallest processing unit object with practical significance in the two-dimensional space, and use this as a carrier to quickly aggregate information. Description of the Drawings

[0031] Figure 1 The figure shows a flowchart of an object-oriented geospatial big data aggregation method provided by an embodiment of the present invention.

[0032] Figure 2 The following is a flowchart of an object - oriented geospatial big data aggregation method provided by an embodiment of the present invention.

[0033] Figure 3 The following is a schematic diagram of the principle of an object - oriented geospatial big data aggregation method provided by an embodiment of the present invention. Detailed implementation manners

[0034] Now, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. It should be understood that the embodiments shown and described in the drawings are merely exemplary, intended to illustrate the principles and spirit of the present invention, and not to limit the scope of the present invention.

[0035] An embodiment of the present invention provides an object - oriented geospatial big data aggregation method, as Figures 1 to 3 collectively shown, including the following steps S1 - S5:

[0036] S1. According to the specific research scenario of the user, select the remotely sensed image with the most appropriate spatial scale from the GIS data.

[0037] In the embodiment of the present invention, based on the standard that the recognition accuracy of ground objects by remotely sensed images is 2 * 3 pixels, the remotely sensed image with the most appropriate spatial scale is selected from the GIS data in the order of increasing spatial resolution.

[0038] S2. Use object - oriented segmentation technology, select the aggregation scale through the scale selection algorithm, and generate a set of geographical monomer elements.

[0039] Step S2 includes the following sub - steps S21 - S23:

[0040] S21. Test the road patches within the range of the remotely sensed image.

[0041] In the embodiment of the present invention, multi - scale segmentation can be performed according to indexes such as the weights of different bands, colors, shapes, compactness, and smoothness in the remotely sensed image, and parameters of scales such as 5, 10, 50, 100, 500, etc. are set to complete multiple groups of multi - scale segmentation experimental results. The larger its value, the larger the polygon area of the segmented object, and the fewer the number of patches. Conversely, the smaller the polygon area, the more the number of patches. Use the scale selection algorithm to analyze the optimal segmentation scale. When the segmentation scale approaches the optimal value, the pure objects in the segmentation result increase, the difference between objects increases, and the overall mean variance of the segmented objects at this scale increases. When the overall mean variance of the image objects is the largest, the spectral difference between objects is the largest.

[0042] S22. When the inside of the road patch unit is uniform and there are differences between the road and adjacent ground object patch units, use its corresponding segmentation scale as the aggregation scale.

[0043] S23. Generate a set of geographical monomer elements according to the aggregation scale.

[0044] In the embodiments of the present invention, the set of geographical monomer elements can be used as a carrier for the aggregation of two-dimensional geographical elements of geospatial big data. It is a natural surface boundary with practical significance and also the most basic set of objects that can express surface features.

[0045] S3. Perform object-oriented land cover type interpretation on the remote sensing image to obtain raster data of natural surface types with clear geographical boundaries and categories.

[0046] Step S3 includes the following sub-steps S31 to S33:

[0047] S31. Use vegetation index, water body index, building index, and point of interest (POI) as indicators, and adopt decision tree classification method to divide the first-level classes.

[0048] S32. Adopt the maximum likelihood method based on samples to divide the second-level classes, ensuring that the number of samples for each second-level class is greater than 30.

[0049] S33. Interpret the remote sensing image through the first-level classes and the second-level classes to obtain raster data of natural surface types with clear geographical boundaries and categories.

[0050] In the embodiments of the present invention, the selected samples should ensure uniform spatial distribution to ensure the purity and typicality of land type features.

[0051] S4. Integrate multi-source, multi-scale, and multi-format GIS data, and construct corresponding geospatial data sets in combination with raster data of natural surface types.

[0052] Step S4 includes the following sub-steps S41 to S43:

[0053] S41. According to the spatial location attributes of the data and their corresponding spatial distribution characteristics, perform spatialization processing on digital statistical GIS data to obtain digital statistical vector data.

[0054] In the embodiments of the present invention, digital statistical GIS data includes data such as society, economy, and social behavior.

[0055] In the embodiments of the present invention, the spatialization processing can spatially link the data according to types such as administrative regions, oasis regions, built-up areas, and construction land.

[0056] S42. Perform resampling interpolation processing on the digital statistical vector data, thematic element maps, zoning boundaries, and grading maps in the GIS data to obtain resampled interpolation raster data.

[0057] In the embodiments of the present invention, the grid size of the resampled and interpolated raster data is consistent with the remote sensing images used to produce each vector data.

[0058] S43. Perform projection conversion processing on the natural surface type raster data and the resampled and interpolated raster data, and use spatial association technology to spatially integrate and inherit various raster data in a hierarchical form to construct a geospatial data set.

[0059] In the embodiments of the present invention, the transverse Mercator projection is used for large areas, and the Gaussian Kriging projection is used for small areas.

[0060] In the embodiments of the present invention, the constructed geospatial data set has a unified data format, retains data information, has a free spatial scale, and is consistent in spatial projection.

[0061] S5. Use the aggregation method of spatial matching and spatial filtering technologies to extract the effective information of different spatial scale element layers of the geospatial data set, and combine it with the geographical single element set to obtain a geospatial big data set.

[0062] Step S5 includes the following sub-steps S51 to S52:

[0063] S51. Sort the element layers in the geospatial data set according to the spatial resolution size, and the spatial resolution of the data in each layer from top to bottom is arranged from small to large.

[0064] S52. Use the geographical single element in the geographical single element set as the minimum processing unit, and adopt the object-oriented mode filtering method to aggregate the information of each element layer in the geospatial data set to obtain a geospatial big data set.

[0065] In the embodiments of the present invention, after obtaining the geospatial big data set, relevant geospatial information analysis work such as information extraction, spatial analysis, statistical analysis, and visualization mapping can be carried out on this basis.

[0066] Those of ordinary skill in the art will realize that the embodiments described here are to help readers understand the principles of the present invention, and it should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those of ordinary skill in the art can make various other specific deformations and combinations that do not deviate from the essence of the present invention according to the technical revelations disclosed in the present invention, and these deformations and combinations are still within the protection scope of the present invention.

Claims

1. An object-oriented geospatial big data aggregation method, characterized in that, Including the following steps: S1. Select the remotely sensed image with the most appropriate spatial scale from GIS data according to the specific research scenario of the user; S2. Use object-oriented segmentation technology to select the aggregation scale through a scale selection algorithm and generate a set of geographical monomer elements; S3. Conduct object-oriented land cover type interpretation on the remotely sensed image to obtain raster data of natural surface types with clear geographical boundaries and categories; S4. Integrate GIS data of multiple sources, multiple scales, and multiple formats, and construct a corresponding geographical spatial data set in combination with the raster data of natural surface types; S5. Use the aggregation method of spatial matching and spatial filtering technologies to extract the effective information of different spatial scale element layers in the geographical spatial data set, and combine it with the set of geographical monomer elements to obtain a geographical spatial big data set; Among them, the step S2 includes: S21. Test the road patches within the range of the remotely sensed image; S22. When the interior of the road patch unit is uniform and there are differences between the road and adjacent feature patch units, use its corresponding segmentation scale as the aggregation scale; S23. Generate a set of geographical monomer elements according to the aggregation scale; The step S4 includes: S41. Perform spatialization processing on the digital statistical type GIS data according to the spatial position attributes of the data and their corresponding spatial distribution characteristics to obtain digital statistical type vector data; S42. Perform resampling interpolation processing on the digital statistical type vector data, the thematic element map, zoning boundary, and grading map in the GIS data to obtain resampled interpolation raster data; S43. Perform projection conversion processing on the raster data of natural surface types and the resampled interpolation raster data, and use spatial association technology to spatially integrate and inherit various raster data in a hierarchical form to construct a geographical spatial data set.

2. The geospatial big data aggregation method according to claim 1, wherein The step S3 includes the following sub-steps: S31. Use vegetation index, water body index, building index, and points of interest as indicators, and adopt the decision tree classification method to divide the first-level classes; S32. Adopt the maximum likelihood method based on samples to divide the second-level classes, ensuring that the number of samples for each second-level class is greater than 30; S33. Interpret the remotely sensed image through the first-level classes and second-level classes to obtain raster data of natural surface types with clear geographical boundaries and categories.

3. The geospatial big data aggregation method according to claim 1, wherein The step S5 includes the following sub-steps: S51. Sort the element layers in the geographical spatial data set according to the spatial resolution size, and the spatial resolution of the data layers from top to bottom increases from small to large; S52. Use the geographical monomer elements in the set of geographical monomer elements as the smallest processing unit, and adopt the object-oriented mode filtering method to aggregate the information of each element layer in the geographical spatial data set to obtain a geographical spatial big data set.

Images