An automatic processing method for equally dividing strip-shaped element patches in complex landforms

Abstract

The present invention relates to an automatic processing method for equally dividing strip element spots in complex land type spots, belonging to the field of geographic information science and technology, including: scanning and analyzing complex land type spot data according to the spatial distribution characteristics and topological relationships of the strip elements to obtain classification and recognition results; extracting strip elements through the classification and recognition results, establishing an automatic recognition model for irregular strip elements based on a neural network decision tree structure, recognizing the strip elements to obtain the position and shape of the irregular strip element spots; analyzing the irregular strip element spots through a fishbone segmentation method to obtain a geographic center line; vertically segmenting the strip element spots according to the geographic center line, integrating the segmented irregular strip element spots with adjacent spots, and obtaining the land type attributes of the adjacent spots. The strip elements in the geographic information data are divided into two, and the boundaries are integrated and managed according to the attributes of the adjacent spots, thereby improving the accuracy of land type spot recognition and segmentation.

Description

Technical Field

[0001] The invention belongs to the field of geographic information science and technology, and specifically relates to an automatic processing method for equally dividing strip element patches in complex land type patches. Background Art

[0002] With the rapid development of geographic information systems (GIS) and remote sensing technology, the automatic identification and equalization of land type patches has become a challenging task. In particular, in complex land type patches, such as urban areas or agricultural areas, there are a large number of irregular strip element patches. When using these strip element data for research and application, it is often necessary to delineate the management boundary according to the geographic center line of the strip element. Strip element data are often irregular in shape and lack the complete expression of strip geographic entities, especially the expression of geographic entities is often disconnected at level crossings or overpasses, and the characteristics of linear extension are hidden at the intersection, resulting in reduced scalability of strip surface feature data in some application scenarios. Traditional algorithms are difficult to effectively handle this situation. When skeleton line extraction and analysis is performed based on roads, ditches, rivers, ridges and other patches, skeleton line results that do not conform to the actual situation will appear, which is manifested in the problem of skeleton line continuity and an undesirable skeleton line structure. This skeleton line structure will affect the further application of other research such as automatic configuration of surface feature annotations, delineation of management units, polygon splitting and merging, road structure identification, and river watershed identification. Summary of the invention

[0003] In order to solve the above problems existing in the prior art, the present invention provides an automatic identification method of strip element patches in evenly divided complex land-type patches;

[0004] The purpose of the present invention can be achieved through the following technical solutions:

[0005] An automatic processing method for equally dividing strip element patches in complex land type patches, comprising:

[0006] Acquire complex land type patch data, preset the spatial distribution characteristics of strip elements, scan and analyze the complex land type patch data according to the spatial distribution characteristics and topological relationship of the strip elements to obtain classification and recognition results, extract strip elements through the classification and recognition results, establish an automatic recognition model for irregular strip elements based on a neural network decision tree structure, and recognize the strip elements through the automatic recognition model to obtain the position and shape of the irregular strip element patch;

[0007] The irregular strip element patches are analyzed by the fishbone segmentation method to obtain the geographic center line corresponding to the strip element; the strip element patches are vertically segmented according to the geographic center line, the segmented irregular strip element patches are integrated with adjacent patches, and the land attributes of the adjacent patches are obtained.

[0008] Specifically, the spatial distribution characteristics of the strip elements include a direct facing mode, an oblique facing mode, and a mixed mode; the direct facing mode is that when the strip elements are adjacent to each other, the projections of the common arc segments at the same distance and in the opposite direction overlap; the oblique facing mode is that there is no overlapping part after the projection of the common arc segments, but they are parallel in the oblique direction and have the same distance; the mixed mode is the arc segment spatial distribution mode when the direct facing mode and the oblique facing mode appear at the same time.

[0009] Specifically, the scanning and analysis method is: according to semantic and topological relationships, identify and screen the strip surface element data and the common arcs between the data; according to the spatial distribution characteristics of the strip elements, classify and identify the common arcs according to the spatial distribution pattern to obtain the position and shape of the strip element patches.

[0010] Specifically, the method for obtaining the classification recognition result is as follows: obtaining common arc segments through the topological relationship and semantic relationship between the strip element data, and taking the common arc segments whose arc length is less than the maximum length threshold of the layer under the corresponding strip element data type as the pattern recognition candidate set; judging the category of the patches before and after the overlapping area according to the actual distribution of the objects, if the patch categories do not match, the corresponding common arc segments are removed from the pattern recognition candidate set, and the classification recognition result is obtained through the pattern recognition candidate set.

[0011] Specifically, the fishbone segmentation method includes: extracting the endpoints of the corresponding irregular polygon according to the irregular strip element pattern, extending the endpoints 45° along the normal position of the irregular polygon, counting the intersections of the extended ends, and obtaining a similar center line by connecting the intersections; extending the similar center line to the boundary of the irregular polygon to obtain the geographic center line.

[0012] Specifically, the vertical segmentation method is: cutting the irregular polygon through the geographic center line to obtain segmented polygons.

[0013] Specifically, the integration method is: assigning the image attributes of the plots on both sides of the geographic center line to the adjacent segmentation polygons.

[0014] Specifically, the automatic recognition model recognition method is:

[0015] Constructing a constrained triangulated network according to the strip element pattern, extracting an initial skeleton line based on the constrained triangulated network, and constructing a topological relationship according to the initial skeleton line to obtain a skeleton line;

[0016] Preset a width limit threshold, and prune the branch suspension lines of the skeleton line according to the width limit threshold;

[0017] A triangular face set is obtained based on the intersection of the skeleton lines through the constrained triangulated network identification, and a branch triangular face set is obtained by performing a subtraction operation on the triangular face set according to the in-plane triangles corresponding to the strip element pattern;

[0018] For each branch of the branch triangular face set, two adjacent points other than the nodes are extracted to form a branch vector, the angle of any branch vector is calculated, and the two vectors with the largest angles are selected as the skeleton line extension direction. If the skeleton line extension direction corresponding to the strip element pattern input by the automatic recognition model is not a quantitative constant, then the strip element pattern is an irregular strip element pattern.

[0019] The beneficial effects of the present invention are:

[0020] (1) By setting fishbone and vertical segmentation methods, irregular strip elements are reasonably incorporated into the management boundaries of adjacent patches, ensuring the even distribution of the boundaries of irregular strip elements. The segmented irregular strip elements are integrated with the adjacent patches, and the land type attributes of the adjacent patches are obtained. The extended boundaries of the adjacent land type patches are obtained, and the unit boundaries are defined using the extended boundaries. This improves the accuracy of land type patch recognition and segmentation, extends the patch data, and improves the continuity between irregular strip element patches.

[0021] (2) By setting up an automatic recognition model, the location and shape of irregular strip element patches can be accurately identified. The system has a high degree of automation and can quickly process large amounts of geographic information data. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0023] Figure 1 The present invention is a flow chart of an automatic processing method for equally dividing strip element patches in complex land-type patches. DETAILED DESCRIPTION

[0024] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation methods, structures, features and effects of the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments.

[0025] See also Figure 1 , an automatic processing method for equally dividing strip element patches in complex land-type patches, comprising:

[0026] Acquire complex land type patch data, preset the spatial distribution characteristics of strip elements, scan and analyze the complex land type patch data according to the spatial distribution characteristics and topological relationship of the strip elements to obtain classification and recognition results, extract strip elements through the classification and recognition results, establish an automatic recognition model for irregular strip elements based on a neural network decision tree structure, and recognize the strip elements through the automatic recognition model to obtain the position and shape of the irregular strip element patch;

[0027] In this embodiment, the geographic patch is a physical unit extracted from a high-resolution image, with clear "patch boundary morphology" (geometry) and "spectral physical characteristics" (algebra). Its connotation and extension include: the precise position, morphology and cluster structure of the patch in spatial distribution; the natural and social functional attributes of the patch; the quantitative indicators and changing process of the information content carried by the patch; the state difference of the patch in the environment (internal cause) and the changing trend after being affected by external disturbance (external cause).

[0028] The irregular strip element patches are analyzed by the fishbone segmentation method to obtain the geographic center line corresponding to the strip element; the strip element patches are vertically segmented according to the geographic center line, the segmented irregular strip element patches are integrated with adjacent patches, and the land attributes of the adjacent patches are obtained.

[0029] Specifically, the spatial distribution characteristics of the strip elements include a direct facing mode, an oblique facing mode, and a mixed mode; the direct facing mode is that when the strip elements are adjacent to each other, the projections of the common arc segments at the same distance and in the opposite direction overlap; the oblique facing mode is that there is no overlapping part after the projection of the common arc segments, but they are parallel in the oblique direction and have the same distance; the mixed mode is the arc segment spatial distribution mode when the direct facing mode and the oblique facing mode appear at the same time.

[0030] Specifically, the scanning and analysis method is: according to semantic and topological relationships, identify and screen the strip surface element data and the common arcs between the data; according to the spatial distribution characteristics of the strip elements, classify and identify the common arcs according to the spatial distribution pattern to obtain the position and shape of the strip element patches.

[0031] Specifically, the method for obtaining the classification recognition result is as follows: obtaining common arc segments through the topological relationship and semantic relationship between the strip element data, and taking the common arc segments whose arc length is less than the maximum length threshold of the layer under the corresponding strip element data type as the pattern recognition candidate set; judging the category of the patches before and after the overlapping area according to the actual distribution of the objects, if the patch categories do not match, the corresponding common arc segments are removed from the pattern recognition candidate set, and the classification recognition result is obtained through the pattern recognition candidate set.

[0032] In this embodiment, among the pattern recognition results after all common arc segment pattern matching, classification and identification are successful and screened, the pattern derivation method for arc segments with small arc length difference is to directly connect the head and tail end points of the two arc segments respectively. If an intersection occurs after connecting the head and tail end points of the two arc segments, the connection method of the head and tail points is changed until there is no intersection, and then the surface is constructed; for arc segments with large arc length difference, the projection area of ​​the shorter arc segment on the longer arc segment is used as the constructed surface; for the pattern recognition results obtained according to different pattern recognitions with arc length difference, different grayscale values ​​are used to distinguish the two derived surfaces; finally, attributes are assigned to the surface constructed by the corresponding pattern derivation scene, and the specific assignment rules are: the entity information corresponding to the arc segment with longer arc length is used as the entity information of the constructed surface.

[0033] Specifically, the fishbone segmentation method includes: extracting the endpoints of the corresponding irregular polygon according to the irregular strip element pattern, extending the endpoints 45° along the normal position of the irregular polygon, counting the intersections of the extended ends, and obtaining a similar center line by connecting the intersections; extending the similar center line to the boundary of the irregular polygon to obtain the geographic center line.

[0034] Specifically, the vertical segmentation method is: cutting the irregular polygon through the geographic center line to obtain segmented polygons.

[0035] Specifically, the integration method is: assigning the image attributes of the plots on both sides of the geographic center line to the adjacent segmentation polygons.

[0036] In this embodiment, the dilation-erosion transformation and the concave filling processing are introduced to realize the accurate calculation of the outer boundary contour; the skeleton line correction under the boundary constraint: the main skeleton line is corrected with the bridge surface boundary as the constraint to eliminate the defect that its end node does not fall on the boundary, and the neighboring line is extracted according to the optimized skeleton line; the neighboring image patch entity reconstruction: according to the principle of maximum semantic coverage or the principle of maximum area similarity, the entity information of the image patches within the neighboring area and near the outer boundary contour after the neighboring processing is annotated.

[0037] Specifically, the automatic recognition model recognition method is:

[0038] Constructing a constrained triangulated network according to the strip element pattern, extracting an initial skeleton line based on the constrained triangulated network, and constructing a topological relationship according to the initial skeleton line to obtain a skeleton line;

[0039] Preset a width limit threshold, and prune the branch suspension lines of the skeleton line according to the width limit threshold;

[0040] A triangular face set is obtained based on the intersection of the skeleton lines through the constrained triangulated network identification, and a branch triangular face set is obtained by performing a subtraction operation on the triangular face set according to the in-plane triangles corresponding to the strip element pattern;

[0041] For each branch of the branch triangular face set, two adjacent points other than the nodes are extracted to form a branch vector, the angle of any branch vector is calculated, and the two vectors with the largest angles are selected as the skeleton line extension direction. If the skeleton line extension direction corresponding to the strip element pattern input by the automatic recognition model is not a quantitative constant, then the strip element pattern is an irregular strip element pattern.

[0042] In this embodiment, the automatic identification algorithm of the spots based on the neural network decision tree sets the threshold and weight of the identification index for the identification of the spot type in the existing research. For the spatial pattern characteristic index of the surface element, a spatial pattern characteristic index system including the area of ​​the spatial object itself, the base-to-height ratio, the spatial relationship between the objects, the geometry (distance, direction, etc.) and the semantic features is established. With the above basic spatial pattern characteristic index system as the input layer, the spatial pattern is understood based on the neural network, and the parameters and thresholds are adaptively adjusted through sample training to realize the automatic identification of regular spots and irregular spots, extract structured objects from the spot data based on the semantic information, and identify their aggregation patterns according to the spatial structure description parameters; by introducing the Miter-type buffer transformation, the outer boundary contour of the typical aggregation pattern is extracted, and the surface elements inside the contour are typified; finally, the structured object contour is restored and the merged result is added to the spot data through the spatial difference operation. The typical characteristics of the structured objects can be effectively maintained locally, and the changes in the area of ​​various types of land can be balanced as a whole, and other visual conflicts are not easy to occur.

[0043] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. An automatic processing method for equally dividing strip element patches in complex land-type patches, characterized in that: include: Acquire complex land type patch data, preset the spatial distribution characteristics of strip elements, scan and analyze the complex land type patch data according to the spatial distribution characteristics and topological relationship of the strip elements to obtain classification and recognition results, extract strip elements through the classification and recognition results, establish an automatic recognition model for irregular strip elements based on a neural network decision tree structure, and recognize the strip elements through the automatic recognition model to obtain the position and shape of the irregular strip element patch; The irregular strip element patches are analyzed by the fishbone segmentation method to obtain the geographic center line corresponding to the strip element; the strip element patches are vertically segmented according to the geographic center line, the segmented irregular strip element patches are integrated with adjacent patches, and the land attributes of the adjacent patches are obtained.

2. The automatic processing method for equally dividing strip element patches in complex land-type patches according to claim 1 is characterized in that: The spatial distribution characteristics of the strip elements include a direct facing mode, an oblique facing mode, and a mixed mode; the direct facing mode refers to the overlapping parts of the projections of common arc segments at the same distance and in the opposite direction when the strip elements are adjacent to each other; the oblique facing mode refers to the absence of overlapping parts after the projection of the common arc segments, but they are parallel in the oblique direction and at the same distance; the mixed mode is the arc segment spatial distribution mode when the direct facing mode and the oblique facing mode appear at the same time.

3. The automatic processing method for equally dividing strip element patches in complex land-type patches according to claim 1 is characterized in that: The scanning analysis method is: According to the semantic and topological relationships, the strip surface element data and the common arcs between the data are identified and screened; according to the spatial distribution characteristics of the strip elements, the common arcs are classified and identified according to the spatial distribution patterns to obtain the position and shape of the strip element patches.

4. The automatic processing method for equally dividing strip element patches in complex land-type patches according to claim 1 is characterized in that: The classification recognition result is obtained by: Through the topological relationship and semantic relationship between the strip feature data, the common arc segments are obtained, and the common arc segments whose arc length is less than the maximum length threshold of the layer under the corresponding strip feature data type are used as the pattern recognition candidate set; according to the actual distribution of the ground objects, the categories of the patches before and after the overlapping area are judged. If the patch categories do not match, the corresponding common arc segments will be removed from the pattern recognition candidate set, and the classification recognition results will be obtained through the pattern recognition candidate set.

5. The automatic processing method for equally dividing strip element patches in complex land-type patches according to claim 1 is characterized in that: The fishbone segmentation method includes: extracting the endpoints of the corresponding irregular polygon according to the irregular strip element pattern, extending the irregular polygon at 45 degrees along the normal position of the irregular polygon at the endpoints, counting the intersections of the extended ends, and obtaining a similar center line by connecting the intersections; extending the similar center line to the boundary of the irregular polygon to obtain a geographic center line.

6. The automatic processing method for equally dividing the strip element patches in the complex land-type patches according to claim 5 is characterized in that: The vertical segmentation method is: cutting the irregular polygon through the geographic center line to obtain segmented polygons.

7. The automatic processing method for equally dividing strip element patches in complex land-type patches according to claim 6 is characterized in that: The integration method is: assigning the image attributes of the plots on both sides of the geographic center line to the adjacent segmentation polygons.

8. The automatic processing method for equally dividing strip element patches in complex land-type patches according to claim 1 is characterized in that: The automatic recognition model recognition method is: Constructing a constrained triangulated network according to the strip element pattern, extracting an initial skeleton line based on the constrained triangulated network, and constructing a topological relationship according to the initial skeleton line to obtain a skeleton line; Preset a width limit threshold, and prune the branch suspension lines of the skeleton line according to the width limit threshold; A triangular face set is obtained based on the intersection of the skeleton lines through the constrained triangulated network identification, and a branch triangular face set is obtained by performing a subtraction operation on the triangular face set according to the in-plane triangles corresponding to the strip element pattern; For each branch of the branch triangular face set, two adjacent points other than the nodes are extracted to form a branch vector, the angle of any branch vector is calculated, and the two vectors with the largest angles are selected as the skeleton line extension direction. If the skeleton line extension direction corresponding to the strip element pattern input by the automatic recognition model is not a quantitative constant, then the strip element pattern is an irregular strip element pattern.

Images