Mountain city split-level building DEM optimization method and system based on confidence coefficient

Abstract

The invention relates to the technical field of digital elevation model processing, in particular to a confidence-based mountain city staggered-floor building DEM optimization method and system, and the method comprises the steps: carrying out the grouping frequency number statistics of a buffer region, generating a frequency number histogram, recognizing a significant wave crest in the frequency number histogram, calculating a staggered-floor confidence score, and carrying out the calculation of the frequency number histogram according to the significant wave crest and the staggered-floor confidence score. The house type is judged, if the house type is a common house, the minimum elevation value in the significant wave crests is taken as the terrace height, and if the house type is a split-level building, a binding relation set of the house side edge and the terrace height is established; and taking the house contour line as a terrain fracture line, generating a constraint point in combination with the terrace height, and generating a DEM by adopting a constraint triangulated irregular network method. According to the scheme, manual intervention is not needed, the split-level building in the mountain city can be automatically, effectively and accurately recognized, high-fidelity modeling is conducted on the multi-level terrace height of the split-level building, and therefore the high-precision DEM conforming to semantic authenticity is generated, and the precision and authenticity of the DEM are improved.

Description

Technical Field

[0001] This invention relates to the field of digital elevation model processing technology, specifically to a confidence-based method and system for optimizing the DEM of staggered buildings in mountainous cities. Background Technology

[0002] Digital elevation models (DEMs) are fundamental data for fields such as terrain analysis, urban planning, and flood simulation. However, existing technologies face significant challenges in generating DEMs for mountainous cities. Traditional methods, particularly automatic filtering algorithms, struggle to accurately remove building points, resulting in a lack of real ground points in building areas. Furthermore, traditional methods often rely on manually drawing feature lines to determine building heights, which is inefficient and highly subjective. While some machine learning methods have shown potential to address these issues, they depend on large amounts of labeled data, have limited model generalization capabilities, and suffer from poor engineering applicability.

[0003] In particular, mountainous cities often contain complex buildings such as split-level buildings and stilted houses, with different sides situated on platforms at different heights. Existing automated methods typically assign a uniform elevation value to the entire building, resulting in severely distorted smooth slopes around the building in the DEM, failing to reproduce the true split-level terrain and greatly affecting the accuracy and reliability of the DEM.

[0004] Therefore, there is an urgent need for a confidence-based DEM optimization method and system for staggered buildings in mountainous cities. This system should be able to automatically, effectively, and accurately identify staggered buildings in mountainous cities without human intervention, and perform high-fidelity modeling of their multi-level ground heights, thereby generating a high-precision DEM that conforms to semantic realism and improving the accuracy and realism of the DEM. Summary of the Invention

[0005] One of the objectives of this invention is to provide a confidence-based DEM optimization method for staggered-level buildings in mountainous cities. This method can automatically, effectively, and accurately identify staggered-level buildings in mountainous cities without human intervention, and perform high-fidelity modeling of their multi-level ground heights, thereby generating a high-precision DEM that conforms to semantic realism and improving the accuracy and realism of the DEM.

[0006] The basic solution provided by this invention is a confidence-based DEM optimization method for staggered-level buildings in mountainous cities, comprising the following: Acquisition steps: Acquire building vector outline surface data and initial DEM data; Statistical steps: Using the building vector outline as a reference, a buffer area is generated by expanding outward, and the DEM elevation values ​​within the buffer area are grouped and frequency counted to generate a frequency histogram; Analysis steps: Identify significant peaks in the frequency histogram, calculate the split-level confidence score, and determine the building type based on the significant peaks and the split-level confidence score. If the building type is a regular building, take the minimum elevation value among the significant peaks as the uniform ground height. If the building type is a split-level building, establish a set of binding relationships between the building side and the ground height. Generation steps: Use the building outline as the terrain break line and combine it with the ground height to generate constraint points. Based on the terrain break line and constraint points, use the constrained irregular triangular mesh method to generate the DEM.

[0007] The second objective of this invention is to provide a confidence-based DEM optimization system for staggered-level buildings in mountainous cities.

[0008] This invention provides a second basic solution: a confidence-based DEM optimization system for staggered-level buildings in mountainous cities, used to implement the aforementioned confidence-based DEM optimization method for staggered-level buildings in mountainous cities, comprising: The data input module is used to acquire building vector outline data and initial DEM data; The frequency and confidence analysis module is used to perform buffer zone generation, elevation frequency statistics, significant peak identification, and cross-level confidence score calculation. The intelligent decision-making module is used to determine the house type and select the path to generate the ground height based on the house type. The semantic DEM modeling module is used to perform terrain fault line definition, constraint point generation, constraint irregular triangular network construction, and DEM generation and output.

[0009] Beneficial effects: This scheme first acquires the vector outline data of the buildings and the initial DEM data. Using the building vector outline as a reference, a buffer area is generated by expanding outward. The DEM elevation values ​​within the buffer area are grouped and frequency-counted to generate a frequency histogram. Second, significant peaks in the frequency histogram are identified, and the split-level confidence score is calculated. Based on the significant peaks and the split-level confidence score, the building type is determined. If the building type is a regular building, the minimum elevation value among the significant peaks is taken as the unified ground height. If the building type is a split-level building, a set of binding relationships between the building side and the ground height is established. Finally, the building outline is used as the terrain fault line, and constraint points are generated in combination with the ground height. Based on the terrain fault line and constraint points, the constrained irregular triangular mesh method is used to generate the DEM. The entire solution requires no manual intervention or training data, overcoming the shortcomings of traditional methods such as low efficiency and strong subjectivity, making it suitable for large-scale engineering applications. Furthermore, it adapts to different types of buildings, generating corresponding ground heights, and thus generating more accurate constraint points, making the DEM generated by the constrained irregular triangular mesh method based on terrain break lines and constraint points more accurate. In particular, this solution distinguishes between different types of buildings by constructing a split-level confidence score, which quantitatively and intelligently differentiates between ordinary buildings and split-level buildings, enabling adaptive selection of processing paths and improving the robustness and accuracy of the method. Furthermore, for split-level buildings, a set of binding relationships between the building's side and the ground height is constructed, enabling semantic-level modeling in the subsequent DEM generation. Specifically, through the binding relationship set between the building's side and the ground height and the constraint of terrain fault lines, the DEM output is upgraded from a simple elevation field to a model that can express the semantics of the building's ground structure, significantly improving accuracy.

[0010] In summary, this solution requires no manual intervention and can automatically, effectively, and accurately identify staggered-level buildings in mountainous cities. It then performs high-fidelity modeling of their multi-level ground elevations, generating a high-precision DEM that conforms to semantic realism, thus improving the accuracy and realism of the DEM. This solution solves the problems of strong reliance on manual intervention and poor adaptability to complex building terrain, achieving fully automated, high-precision DEM optimization. Attached Figure Description

[0011] Figure 1 This is a flowchart illustrating an embodiment of the confidence-based DEM optimization method for staggered-level buildings in mountainous cities according to the present invention. Detailed Implementation

[0012] The following detailed description illustrates the specific implementation method: Example 1 This embodiment is basically as shown in the appendix. Figure 1 As shown, a confidence-based DEM optimization method for staggered-level buildings in mountainous cities is provided, including the following: Acquisition steps: Acquire building vector outline surface data and initial DEM data; The specific process includes: S101. Obtain the vector outline data of the house to determine its location and shape; S102. Obtain initial DEM data; the initial DEM data should be consistent with the data source used in subsequent optimization, for example, it can be generated by interpolation after filtering non-ground points from point cloud data; its resolution is not equal to the preset minimum threshold, and the grid size is greater than or equal to the preset grid size to ensure that it can meet the analysis needs of complex terrain in mountainous cities. The smaller the grid size, the more refined the DEM and the more detailed the data it can provide; the preset minimum threshold and preset grid size are set according to the requirements. The preset minimum threshold should not be too low, and the preset grid size is generally set to 2 meters.

[0013] Statistical steps: Using the building vector outline as a reference, a buffer area is generated by expanding outward, and the DEM elevation values ​​within the buffer area are grouped and frequency counted to generate a frequency histogram; The specific process includes: S201. Obtain the outward expansion distance based on the house area; preset the outward expansion distance corresponding to different house areas, i.e., the outward expansion distance relationship; determine the outward expansion distance based on the outward expansion distance relationship to which the house area belongs; in this embodiment, when the house area is less than or equal to 200 square meters, a 1.0 meter outward expansion distance is used, and when the house area is greater than 200 square meters, a 1.5 meter outward expansion distance is used. S202. Based on the outward expansion distance, a buffer area is generated by expanding outward from the building vector outline surface as the reference. The frequency of all DEM elevation values ​​in the buffer area is counted in groups, and a frequency histogram is generated. In this embodiment, based on the outward expansion distance, a buffer area is generated by expanding outward from the building vector outline surface as the reference. The elevation values ​​of all DEM grid points in the buffer area are grouped at a preset grouping distance (0.1 meters in this embodiment). The frequency of each elevation value group is counted. The median value of each group is used as the representative elevation value of the group. For example, 178.1 meters represents the value range [178.05, 178.15] meters. After traversing all grid points to complete the frequency count, a frequency histogram is generated with the elevation group as the horizontal axis and the frequency as the vertical axis.

[0014] Analysis steps: Identify significant peaks in the frequency histogram, calculate the split-level confidence score, and determine the building type based on the significant peaks and the split-level confidence score. If the building type is a regular building, take the minimum elevation value among the significant peaks as the uniform ground height. If the building type is a split-level building, establish a set of binding relationships between the building side and the ground height. The specific process is as follows: S301. Use a sliding window to identify significant peaks in the frequency histogram; In this embodiment, a sliding window with a width of 1 meter is set, and it slides along the horizontal axis with a step size of 0.1 meters. At the current position of the sliding window, if the center elevation group simultaneously meets the following preset conditions, it is determined to be a significant peak. The preset conditions include: local extremum conditions, significance conditions, and dominant conditions; Local extremum condition: frequency of elevation group The frequency is greater than the frequency of each of the n adjacent elevation groups before and after it; in this embodiment, n is 2. Significance condition: frequency of elevation groups The proportion of the total frequency within the current window exceeds a first preset proportion; in this embodiment, the first preset proportion is 20%. Dominant condition: Frequency of elevation groups Frequency of the second highest peak within the current window The difference is greater than the second preset percentage of the total frequency of the window; in this embodiment, the second preset percentage is 10%.

[0015] S302. Determine the number N of identified significant peaks. If the number N of identified significant peaks is 1, then the house is determined to be an ordinary house, and the elevation value corresponding to the significant peak is directly assigned. As a uniform ground level, it is also the minimum elevation value among the significant peaks that serves as the uniform ground level. If N≥2, then calculate the layered confidence score. And determine the confidence score of the misalignment. If the value is greater than or equal to a preset threshold, the house is determined to be a split-level building and S303 is executed; otherwise, the house is determined to be a regular house, and a conservative strategy is adopted, taking the minimum elevation value among the significant peaks as the uniform ground height; in this embodiment, the preset threshold is 0.7. The calculation of the stratified confidence score uses a stratified confidence scoring model, specifically: ; in, The average elevation difference between wave crests. The sharpness of each peak (which can be measured by the kurtosis coefficient); As the weighting coefficient, in this embodiment ; S303. For split-level buildings, the elevation value of each significant peak is correlated with the orientation of the side of the building to form a set of binding relationships between the side and the ground height. Specifically, the average azimuth of grid points belonging to each peak elevation group is calculated and bound to the house side closest to the average azimuth, forming a set of binding relationships between side and ground height, such as: {East: H1, South: H1, West: H2, North: H2}.

[0016] Generation steps: Use the building outline as the terrain break line, and combine it with the ground height to generate constraint points. Use the constrained triangular mesh (TIN) method to generate the DEM.

[0017] The specific process of generating a DEM using the constrained triangular network (TIN) method is as follows: S401. Perform data preprocessing, marking the ground point cloud data within the building outline area as non-ground points, which will not participate in subsequent DEM generation; S402. Generate constraint features, including: Define the break line: Define the house outline (in this embodiment, the house vector outline) as the terrain break line that must be followed during the TIN construction process, i.e., the house outline break line; Generate elevation constraint points: Take a constraint point every preset distance along the building outline; in this embodiment, the preset distance is 0.5 meters, that is, take a constraint point every 0.5 meters along the building outline. Different ground level height values ​​are assigned to constraint points for different house types, including: For ordinary houses, all constraint points are assigned a uniform ground height. ; For split-level buildings, based on the set of binding relationships between the side and the ground height obtained in S303, the constraint point segments located on different sides are assigned their bound ground height values ​​respectively. ; S403. Constructing constrained TINs and generating DEMs, including: The constrained Delaunay triangulation (CDT) algorithm is used to construct a constrained triangular network (TIN) by taking the ground point cloud data retained in S401, the constrained points generated in S402, and the terrain fault lines as inputs. The constrained Delaunay triangulation (CDT) algorithm forces the terrain fault lines to become the edges of the triangular network, ensuring that the terrain can change abruptly at these points. Based on the edges of the triangular network, TIN uses a linear interpolation algorithm to generate the DEM, which is the final DEM grid data.

[0018] Through the above implementation methods, this solution can effectively identify and accurately reconstruct the multi-level terrain of staggered building areas in mountainous cities, significantly improving the accuracy and realism of the DEM.

[0019] This embodiment also provides a confidence-based DEM optimization system for staggered-level buildings in mountainous cities, used to implement the aforementioned confidence-based DEM optimization method for staggered-level buildings in mountainous cities, including: The data input module is used to acquire building vector outline data and initial DEM data; The frequency and confidence analysis module is used to perform buffer zone generation, elevation frequency statistics, significant peak identification, and staggered floor confidence score calculation. Specifically, it uses the building vector outline as a reference to expand outward to generate a buffer zone, performs grouped frequency statistics on the DEM elevation values ​​within the buffer zone, generates a frequency histogram, identifies significant peaks in the frequency histogram, and calculates the staggered floor confidence score. The frequency and confidence analysis module includes: a frequency statistics unit, a peak identification unit, and a confidence calculation unit; The frequency statistics unit is used to expand outward from the building vector outline surface to generate a buffer area, and to perform grouped frequency statistics on the DEM elevation values ​​within the buffer area to generate a frequency histogram. Peak identification unit, used to identify significant peaks in the frequency histogram; The confidence score calculation unit is used to calculate the confidence score for the different levels of stratification. The intelligent decision-making module is used to determine the house type and select the path to generate the ground height based on the house type. The intelligent decision-making module includes: a path selection unit and a side binding unit; The path selection unit is used to determine the house type based on the significant peak and the confidence score of the split floor, and select the processing path according to the house type. If the house type is a normal house, the minimum elevation value among the significant peaks is taken as the uniform ground height. If the house type is a split floor building, the side binding unit is called to establish a set of binding relationships between the house side and the ground height. Side binding unit, used to establish a set of binding relationships between the side of the house and the ground level; The semantic DEM modeling module is used to perform terrain fault line definition, constraint point generation, constraint irregular triangular network construction, and DEM generation and output.

[0020] The semantic DEM modeling module includes: constraint feature generation unit and constraint TIN construction unit.

[0021] The constraint element generation unit is used to use the building outline as the terrain break line and combine it with the ground height to generate constraint points. Constrained TIN building blocks are used to generate DEMs based on terrain fault lines and constraint points using the constrained irregular triangular mesh method.

[0022] Example 2 This embodiment is basically the same as the above embodiment, except that: in order to further optimize the terrain representation inside the staggered-level building, the following steps can be performed after S303: Internal planarization forced steps: Logical partitioning: For split-level buildings, the area enclosed by consecutive sides bound to the same elevation value is divided into a logical sub-region; Internal constraint generation: Generate internal break lines between different logical sub-regions; add a set of elevation auxiliary points within each logical sub-region, and force all elevation auxiliary points to be assigned a uniform elevation value for that logical sub-region; Integrated mesh construction: The internal break lines and elevation auxiliary points are added to the TIN construction process of S403, thereby forcing each logical sub-region to appear as a horizontal plane in the DEM, and forming a steep transition between logical sub-regions.

[0023] The above descriptions are merely embodiments of the present invention. Commonly known structures and characteristics are not described in detail here. Those skilled in the art are aware of all common technical knowledge in the field prior to the application date or priority date, are aware of all existing technologies in that field, and have the ability to apply conventional experimental methods prior to that date. Those skilled in the art can, under the guidance of this application, improve and implement this solution in combination with their own capabilities. Some typical known structures or methods should not be obstacles for those skilled in the art to implement this application. It should be noted that those skilled in the art can make several modifications and improvements without departing from the structure of the present invention. These should also be considered within the scope of protection of the present invention, and will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A confidence-based DEM optimization method for staggered-level buildings in mountainous cities, characterized in that, Includes the following: Acquisition steps: Acquire building vector outline surface data and initial DEM data; Statistical steps: Using the building vector outline as a reference, a buffer area is generated by expanding outward, and the DEM elevation values ​​within the buffer area are grouped and frequency counted to generate a frequency histogram; Analysis steps: Identify significant peaks in the frequency histogram, calculate the split-level confidence score, and determine the building type based on the significant peaks and the split-level confidence score. If the building type is a regular building, take the minimum elevation value among the significant peaks as the uniform ground height. If the building type is a split-level building, establish a set of binding relationships between the building side and the ground height. Generation steps: Use the building outline as the terrain break line and combine it with the ground height to generate constraint points. Based on the terrain break line and constraint points, use the constrained irregular triangular mesh method to generate the DEM.

2. The confidence-based DEM optimization method for staggered-level buildings in mountainous cities according to claim 1, characterized in that, The statistical steps include: S201. Obtain the outward expansion distance based on the house area; S202. Based on the outward expansion distance, take the building vector outline as the reference, expand outward to generate a buffer area, and perform group frequency statistics on all DEM elevation values ​​within the buffer area to generate a frequency histogram.

3. The confidence-based DEM optimization method for staggered-level buildings in mountainous cities according to claim 2, characterized in that, S202 includes: based on the outward expansion distance, using the building vector outline as a reference, expanding outward to generate a buffer area; grouping the elevation values ​​of all DEM grid points in the buffer area at preset grouping intervals; counting the frequency of each elevation value group; and using the median value of each group as the representative elevation value of that group.

4. The confidence-based DEM optimization method for staggered-level buildings in mountainous cities according to claim 1, characterized in that, The analysis steps include: S301. Use a sliding window to identify significant peaks in the frequency histogram; S302. Determine the number N of identified significant peaks. If the number N of identified significant peaks is 1, then the house is determined to be an ordinary house, and the elevation value corresponding to the significant peak is directly assigned. As a uniform ground level; If N≥2, then calculate the layered confidence score. And determine the confidence score of the misalignment. If the value is greater than or equal to a preset threshold, the house is determined to be a split-level building and S303 is executed; otherwise, the house is determined to be a regular house and the minimum elevation value among the significant peaks is taken as the uniform ground height. S303. For split-level buildings, the elevation value of each significant peak is correlated with the orientation of the side of the building to form a set of binding relationships between the side and the ground height.

5. The confidence-based DEM optimization method for staggered-level buildings in mountainous cities according to claim 4, characterized in that, S301 includes: Set up a sliding window; At the current sliding window position, if the central elevation group simultaneously meets the following preset conditions, it is determined to be a significant peak; The preset conditions include: local extremum conditions, significance conditions, and dominant conditions; Local extremum condition: frequency of elevation group The frequency is greater than the frequency of each of the n adjacent elevation groups before and after it; Significance condition: frequency of elevation groups The proportion of the total frequency within the current window exceeds the first preset proportion; Dominant condition: Frequency of elevation groups Frequency of the second highest peak within the current window The difference is greater than the second preset percentage of the total frequency of the window.

6. The confidence-based DEM optimization method for staggered-level buildings in mountainous cities according to claim 4, characterized in that, The calculation of the stratified confidence score uses a stratified confidence scoring model: ； in, The average elevation difference between wave crests. The sharpness of each wave peak; These are the weighting coefficients.

7. The confidence-based DEM optimization method for staggered-level buildings in mountainous cities according to claim 4, characterized in that, S303 includes: calculating the average azimuth of grid points belonging to each peak elevation group, and binding it to the house side closest to the average azimuth, forming a set of binding relationships between side and ground height.

8. The confidence-based DEM optimization method for staggered-level buildings in mountainous cities according to claim 1, characterized in that, The method of generating a DEM using constrained irregular triangular meshes includes: S401. Perform data preprocessing, marking the ground point cloud data within the building outline area as non-ground points; S402. Generate constraint features, including: Define break lines: Define the building outline as a terrain break line; Generate elevation constraint points: Take a constraint point at preset intervals along the building outline; Different ground height values ​​are assigned to constraint points of different building types. For ordinary buildings, all constraint points are assigned a uniform ground height. For split-level buildings, the set of binding relationships between the side and the ground height is used, and the constraint point segments located on different sides are assigned their bound ground height values ​​respectively. S403. Constructing a constrained TIN and generating a DEM, including: using the constrained Delaunay triangulation algorithm, taking the retained ground point cloud data, the generated constraint points and terrain break lines as input, constructing a constrained irregular triangular network, with the building outline break lines as the edges of the triangular network; Based on the edges of the triangular network, TIN uses a linear interpolation algorithm to generate the DEM.

9. The confidence-based DEM optimization method for staggered-level buildings in mountainous cities according to claim 8, characterized in that, It also includes an internal planarization forced step; Internal planarization enforcement steps: For split-level buildings, the area enclosed by continuous sides bound to the same elevation value is divided into a logical sub-region; internal break lines are generated between different logical sub-regions; Add a set of elevation auxiliary points within each logical sub-region, and force all elevation auxiliary points to be assigned a uniform elevation value for that logical sub-region; add the internal break lines and elevation auxiliary points to the TIN construction process of S403, so that each logical sub-region appears as a horizontal plane in the DEM, and a steep transition is formed between logical sub-regions.

10. A confidence-based DEM optimization system for staggered-level buildings in mountainous cities, characterized in that, The method for implementing the confidence-based DEM optimization method for staggered-level buildings in mountainous cities as described in any one of claims 1-9 includes: The data input module is used to acquire building vector outline data and initial DEM data; The frequency and confidence analysis module is used to perform buffer zone generation, elevation frequency statistics, significant peak identification, and cross-level confidence score calculation. The intelligent decision-making module is used to determine the house type and select the path to generate the ground height based on the house type. The semantic DEM modeling module is used to perform terrain fault line definition, constraint point generation, constraint irregular triangular network construction, and DEM generation and output.

Images