A surface model generation method suitable for low-altitude airspace demarcation

Abstract

The present invention relates to a method for generating a surface model suitable for low-altitude airspace demarcation, comprising the following steps: Step 1, obtaining laser point cloud data within a target range, and screening out effective surface point clouds from the obtained laser point cloud data; Step 2, extracting the spatial range of flight obstacles within the target range, and performing elevation transformation on the flight obstacles within the target range to obtain the transformed point cloud data within the target range; Step 3, iteratively gridding the transformed point cloud data within the target range, and assigning each grid the maximum value of the point cloud elevation within the current grid, that is, generating a surface model. The advantages are: the surface model generated by this method can significantly characterize low-altitude flight obstacles, and the generated surface model data volume is small, suitable for high-frequency calls on the Internet, and has the advantages of low operational difficulty, high efficiency, and strong universality.

Description

Technical Field

[0001] The present invention relates to the technical field of low-altitude airspace demarcation, and in particular to a method for generating a surface model suitable for low-altitude airspace demarcation. Background Art

[0002] With the development of the low-altitude economy, many cities in China have begun to explore low-altitude applications such as low-altitude air traffic, drone logistics, and security inspection cruises. The complex space utilization represented by low-altitude air traffic has broken through the traditional airspace and intervened in the urban surface space. The low-altitude space close to the surface is a natural space resource that needs to be developed in the city and is a rigid demand for the development of the low-altitude economy. The low-altitude airspace is characterized by spatial coordinates and has a significant geographical gene. The undulating terrain, high-rise buildings, high towers, windmills, power lines and other surface structures have a significant impact on the demarcation of low-altitude flight airspace and route planning.

[0003] The current surface models are mainly digital elevation models (DEM), digital terrain models (DTM) and digital surface models (DSM). The generation of grid-type DEM, DTM and DSM, which is commonly used today, can be simply called mechanical modeling. At the rectangular (or square) grid spacing corresponding to the specified resolution, whether calculated based on known points or directly measured, even if the elevations and elevation programs at various grid points are correct, the elevation features (points) will inevitably be lost in large quantities, and the larger the grid, the more serious it is. The phenomenon of "cutting off the head and feet" of the commanding heights and low points is inevitable. Unless the grid is small enough, the data generated thereby is large in volume and too accurate to meet the needs of lightweight applications on the Internet. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a surface model generation method suitable for low-altitude airspace demarcation with small data volume and high precision in response to the above-mentioned prior art.

[0005] The technical solution adopted by the present invention to solve the above technical problems is: a method for generating a surface model suitable for low-altitude airspace demarcation, characterized by comprising the following steps:

[0006] Step 1: Acquire laser point cloud data within the target range, and select valid surface point clouds from the acquired laser point cloud data;

[0007] Step 2: extract the spatial range of the flight obstacles within the target range, and perform elevation transformation on the flight obstacles within the target range to obtain the transformed point cloud data within the target range;

[0008] Step 3, iteratively grid the transformed point cloud data within the target range, and assign each grid the maximum point cloud elevation value in the current grid, that is, generate a surface model;

[0009] The specific steps for generating the surface model in step 3 are:

[0010] Step 3-1, performing a first grid division on the transformed point cloud data within the target range in step 2 to obtain a plurality of first grids;

[0011] Step 3-2: Calculate the entropy of the elevation data in each first grid according to the following formula: E ;

[0012]

[0013] in, N is the number of equally spaced discretized intervals within the distribution range of all elevation data under the current grid size, The elevation data in the current first grid falls in n The probability of an interval;

[0014] Step 3-3, arranging the entropies of all elevation data in the first grid in descending order, and performing a second grid division on the first grid corresponding to the first M entropies to obtain multiple second grids;

[0015] Step 3-4: Process the second grid in the same manner as in step 3-2 and step 3-3 until the divided first grid is a grid of a preset size, thus completing the iterative grid division and generating a surface model.

[0016] Preferably, the specific steps of screening the effective surface point cloud in step 1 are:

[0017] Step 1-1, using coordinate conversion parameters to convert the plane coordinates and elevation benchmark of the laser point cloud data within the target range into the result coordinate system;

[0018] Step 1-2, the original geodetic height of the laser point cloud data in the result coordinate system is corrected by using the existing quasi-geoid refinement result to obtain the laser point cloud data after elevation correction;

[0019] Step 1-3: Filter out noise points in the elevation-corrected laser point cloud data to obtain the screened valid surface point cloud.

[0020] Specifically, the specific process of filtering out noise points in the laser point cloud data after elevation correction in steps 1-3 is as follows:

[0021] Step 1-3-1: Use the elevation-corrected laser point cloud data i Point cloud data points As the center of the sphere, find the neighborhood set of point cloud data in the area where the sphere with radius r is located ;in, i∈[1, Q], Q is the total number of point cloud data points in the elevation-corrected laser point cloud data, is a 3×1 vector consisting of two-dimensional position coordinates and elevation values;

[0022] Step 1-3-2, calculation The mean of all cloud data points in and covariance ,in, is a 3×1 vector, is a 3×3 matrix;

[0023] Step 1-3-3, calculate the i Point cloud data points Local linear transformation matrix with respect to its own neighborhood and bias :

[0024]

[0025] in, is a hyperparameter that controls the degree of smoothing, and is a 1×1 scalar. I is a 3×3 identity matrix, is a 3×3 matrix, is a 3×1 vector;

[0026] Step 1-3-4: Traverse the neighborhood of all point cloud data points and obtain The set of all neighbors of ;

[0027] Steps 1-3-5, according to The calculation formula for The result after filtering out noise points ;

[0028] in, , are the local linear transformation matrix and the bias in each Neighborhood set of The average value on , , Representation calculation The number of elements in ;

[0029] Steps 1-3-6, use i= 1, 2, ...Q, and process according to steps 1-3-1 to 1-3-5 to finally obtain the screened valid surface point cloud.

[0030] Preferably, step 2 further includes supplementing the spatial range of the flight obstacles within the target range to generate transformed point cloud data within the target range.

[0031] Specifically, the flight obstacles include power towers and power lines. If the point clouds of power towers and power lines are completely missing, the specific process of supplementing the spatial range of the flight obstacles within the target range is to collect point cloud data of power lines along the power towers.

[0032] Specifically, if there are power tower and power line point clouds, but the density of the power tower and power line point clouds is too low, resulting in missing and interrupted power line reflections, the power line point cloud data is fitted, and the fitted power line point cloud data is used to complete the existing power line point cloud data.

[0033] Specifically, the specific process of fitting the power line point cloud data is as follows:

[0034] Step a: Use a catenary model to simulate the natural drooping shape of the power line under the action of gravity. The calculation formula of the catenary model is:

[0035]

[0036] in, S is the arc parameter along the electric line starting from the reference point, are the three-dimensional coordinate points predicted by the catenary model, is the coordinate of the reference point, is the radius of curvature of the catenary;

[0037] Step b: Calculate the optimal catenary curvature radius according to the following calculation formula: , and its calculation formula is:

[0038]

[0039] in, The first point cloud data collected j The actual three-dimensional coordinates of the points on the power line, The three-dimensional coordinate points predicted by the catenary model;

[0040] Step c: Set the optimal catenary curvature radius Substituting it into the calculation formula of the catenary model, the optimal catenary model is obtained, that is, the power line point cloud data is calculated according to the optimal catenary model.

[0041] Preferably, the flight obstacle also includes at least one of construction equipment, lightning protection equipment, buildings, chimneys, and water towers. The elevations of the construction equipment, lightning protection equipment, buildings, chimneys, and water towers are modified according to the following calculation formula. The specific calculation formula is:

[0042] H=h+b;

[0043] Among them, H is the height of the facility after elevation modification, h is the existing point cloud height of the facility, and b is the height of the facility before modification.

[0044] In order to keep the low-altitude surface model current, the following steps are also included after step 3:

[0045] Step 4: Obtain the spatial range and elevation changes of the flight obstacles, and update the surface model in the same way as in steps 1 to 3.

[0046] Compared with the prior art, the advantages of the present invention are: by performing elevation transformation on the flight obstacles within the target range, and iteratively gridding the transformed point cloud data within the target range, each grid is assigned the maximum value of the point cloud elevation in the current grid, so that the surface model generated by this method can significantly characterize low-altitude flight obstacles, and the generated surface model data has a small volume, is suitable for high-frequency calls on the Internet, and has the advantages of low operating difficulty, high efficiency, and strong universality. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 The figure is a flow chart of a method for generating a ground model in an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The present invention is further described in detail below with reference to the accompanying drawings.

[0049] like Figure 1 As shown, the surface model generation method applicable to low-altitude airspace demarcation in this embodiment includes the following steps:

[0050] Step 1: Acquire laser point cloud data within the target range, and select valid surface point clouds from the acquired laser point cloud data;

[0051] The specific steps for screening effective surface point clouds are:

[0052] Step 1-1, using coordinate conversion parameters to convert the plane coordinates and elevation benchmark of the laser point cloud data within the target range into the result coordinate system;

[0053] Step 1-2, the original geodetic height of the laser point cloud data in the result coordinate system is corrected by using the existing quasi-geoid refinement result to obtain the laser point cloud data after elevation correction;

[0054] Step 1-3, filtering out noise points in the elevation-corrected laser point cloud data to obtain a screened valid surface point cloud;

[0055] In this embodiment, the specific process of filtering out noise points in the elevation-corrected laser point cloud data is as follows:

[0056] Step 1-3-1: Use the elevation-corrected laser point cloud data i Point cloud data points As the center of the sphere, find the neighborhood set of point cloud data in the area where the sphere with radius r is located ;in, i∈ [1, Q], Q is the total number of point cloud data points in the elevation-corrected laser point cloud data, is a 3×1 vector consisting of two-dimensional position coordinates and elevation values;

[0057] Step 1-3-2, calculation The mean of all cloud data points in and covariance ,in, is a 3×1 vector, is a 3×3 matrix;

[0058] Step 1-3-3, calculate the i Point cloud data points Local linear transformation matrix with respect to its own neighborhood and bias :

[0059]

[0060] in, is a hyperparameter that controls the degree of smoothing, and is a 1×1 scalar. I is a 3×3 identity matrix, is a 3×3 matrix, is a 3×1 vector;

[0061] Step 1-3-4: Traverse the neighborhood of all point cloud data points and obtain The set of all neighbors of ;

[0062] Steps 1-3-5, according to The calculation formula for The result after filtering out noise points ;

[0063] in, , are the local linear transformation matrix and the bias in each Neighborhood set of The average value on , , Representation calculation The number of elements in ;

[0064] Steps 1-3-6, use i= 1, 2, ...Q, and process according to steps 1-3-1 to 1-3-5, and finally obtain the screened effective surface point cloud;

[0065] The above-mentioned noise point filtering method not only retains important edge information of the three-dimensional point cloud data, but also performs local smoothing processing to achieve overall noise reduction, which is beneficial to subsequent meshing;

[0066] Step 2: extract the spatial range of the flight obstacles within the target range, and perform elevation transformation on the flight obstacles within the target range to obtain the transformed point cloud data within the target range;

[0067] In this embodiment, step 2 also includes supplementing the spatial range of the flight obstacles within the target range to generate transformed point cloud data within the target range;

[0068] Specifically, flight obstacles include power towers and power lines. When collecting large-scale point cloud data, the point cloud density is generally within 4 points per square meter. Power lines are very thin, and it is generally impossible to extract complete power line information through point clouds. However, power lines are very important low-altitude flight obstacles. If you hit this type of obstacle, it is not only dangerous to property but also easy to cause personal danger, so special consideration should be given; power towers can generally be located. Due to the relationship between power towers and power lines, combined with sporadic power line point cloud data, the complete information of power lines can be inverted through the model;

[0069] In this embodiment, if the point clouds of power towers and power lines are completely missing, the specific process of supplementing the spatial range of flight obstacles within the target range is: collect point cloud data of power lines along the power towers; if there are point clouds of power towers and power lines, but the density of the point clouds of power towers and power lines is too low to cause the reflection of power lines to be missing and interrupted, fit the point cloud data of power lines, and use the fitted point cloud data of power lines to complete the existing point cloud data of power lines;

[0070] In this embodiment, the specific process of fitting the power line point cloud data is as follows:

[0071] Step a: Use a catenary model to simulate the natural drooping shape of the power line under the action of gravity. The calculation formula of the catenary model is:

[0072]

[0073] in, S is the arc parameter along the electric line starting from the reference point, are the three-dimensional coordinate points predicted by the catenary model, is the coordinate of the reference point, is the radius of curvature of the catenary;

[0074] Step b: Calculate the optimal catenary curvature radius according to the following calculation formula: , and its calculation formula is:

[0075]

[0076] in, The first point cloud data collected j The actual three-dimensional coordinates of the points on the power line, The three-dimensional coordinate points predicted by the catenary model;

[0077] Step c: Set the optimal catenary curvature radius Substitute it into the calculation formula of the catenary model to obtain the optimal catenary model, that is, calculate the power line point cloud data according to the optimal catenary model;

[0078] In this embodiment, the flight obstacle also includes at least one of construction equipment, lightning protection equipment, buildings, chimneys, and water towers. The elevations of the construction equipment, lightning protection equipment, buildings, chimneys, and water towers are modified according to the following calculation formula. The specific calculation formula is:

[0079] H=h+b;

[0080] Among them, H is the height of the facility after elevation modification, h is the existing point cloud height of the facility, and b is the height of the facility before modification;

[0081] Step 3, iteratively grid the transformed point cloud data within the target range, and assign each grid the maximum point cloud elevation value in the current grid, that is, generate a surface model;

[0082] The specific steps for generating the surface model in step 3 are:

[0083] Step 3-1, performing a first grid division on the transformed point cloud data within the target range in step 2 to obtain a plurality of first grids;

[0084] Step 3-2: Calculate the entropy of the elevation data in each first grid according to the following formula: E ;

[0085]

[0086] in, N is the number of equally spaced discretized intervals within the distribution range of all elevation data under the current grid size, The elevation data in the current first grid falls in n The probability of an interval;

[0087] Step 3-3, the entropies of all elevation data in the first grid are arranged in descending order, and the first grid corresponding to the first M entropies is divided into a second grid to obtain multiple second grids; the M value in this embodiment can be determined based on experience or experiments;

[0088] Step 3-4, processing the second grid in the same manner as in step 3-2 and step 3-3 until the divided first grid is a grid of a preset size, thus completing the iterative grid division and generating a surface model;

[0089] Step 4: Obtain the spatial range and elevation changes of the flight obstacles, and update the surface model in the same way as in steps 1 to 3.

[0090] In order to facilitate the understanding of step 3-2, the following specific example is used in this embodiment to illustrate it. Assume that the distribution range of all elevation data under the current grid size after the first grid division is 0-20. Assume that N=4, that is, it is divided into 4 equally spaced intervals of 5. For each first grid, the elevation information entropy of the grid is calculated. Assume that the elevation data in the first first grid is [1, 2, 7, 8, 12], then the corresponding probabilities of the first grid elevation data falling in the 4 intervals are [0.4, 0.4, 0.2, 0], and the corresponding entropy is E=-(0.4*log(0.4)+0.4*log(0.4 )+0.2*log(0.2))=1.0549; assuming that the elevation data in the second first grid is [1, 2, 3, 7, 8], then the corresponding probability that the elevation data of the first grid falls in the four intervals is [0.6, 0.4, 0, 0], and the corresponding entropy is E=-(0.6*log(0.6)+0.4*log(0.4))=0.6730. Since the entropy value corresponding to the first first grid is greater than the entropy value corresponding to the second first grid, if the number of first grids is 2, it means that the elevation data distribution of the first first grid is more complex, and the first first grid needs to be iteratively refined.

[0091] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for generating a surface model suitable for low-altitude airspace demarcation, characterized in that The steps include: Step 1: Acquire laser point cloud data within the target range, and select valid surface point clouds from the acquired laser point cloud data; Step 2: extract the spatial range of the flight obstacles within the target range, and perform elevation transformation on the flight obstacles within the target range to obtain the transformed point cloud data within the target range; Step 3, iteratively grid the transformed point cloud data within the target range, and assign each grid the maximum point cloud elevation value in the current grid, that is, generate a surface model; The specific steps for generating the surface model in step 3 are: Step 3-1, performing a first grid division on the transformed point cloud data within the target range in step 2 to obtain a plurality of first grids; Step 3-2, calculating the entropy of the elevation data in each first grid; Step 3-3, arranging the entropies of all elevation data in the first grid in descending order, and performing a second grid division on the first grid corresponding to the first M entropies to obtain multiple second grids; Step 3-4: Process the second grid in the same manner as in step 3-2 and step 3-3 until the divided first grid is a grid of a preset size, thus completing the iterative grid division and generating a surface model.

2. The method for generating a surface model according to claim 1, characterized in that: The specific steps of screening the effective surface point cloud in step 1 are: Step 1-1, using coordinate conversion parameters to convert the plane coordinates and elevation benchmark of the laser point cloud data within the target range into the result coordinate system; Step 1-2, the original geodetic height of the laser point cloud data in the result coordinate system is corrected by using the existing quasi-geoid refinement result to obtain the laser point cloud data after elevation correction; Step 1-3: Filter out noise points in the elevation-corrected laser point cloud data to obtain a screened valid surface point cloud.

3. The method for generating a surface model according to claim 2, characterized in that: The specific process of filtering out noise points in the laser point cloud data after elevation correction in step 1-3 is as follows: Step 1-3-1: Use the elevation-corrected laser point cloud data i Point cloud data points As the center of the sphere, find the neighborhood set of point cloud data in the area where the sphere with radius r is located ;in, i∈ [1, Q], Q is the total number of point cloud data points in the laser point cloud data after elevation correction, is a 3×1 vector consisting of two-dimensional position coordinates and elevation values; Step 1-3-2, calculation The mean and covariance of all point cloud data points in ; Step 1-3-3, calculate the i Point cloud data points Local linear transformation matrix with respect to its own neighborhood and bias : Step 1-3-4: Traverse the neighborhood of all point cloud data points and obtain The set of all neighbors of ; Steps 1-3-5, according to The calculation formula for The result after filtering out noise points ; in, , are the local linear transformation matrix and the bias in each Neighborhood set of The average value on Steps 1-3-6, use i= 1, 2, ...Q, and process according to steps 1-3-1 to 1-3-5 to finally obtain the screened valid surface point cloud.

4. The method for generating a surface model according to any one of claims 1 to 3, characterized in that: The step 2 also includes supplementing the spatial range of the flight obstacles within the target range to generate transformed point cloud data within the target range.

5. The method for generating a surface model according to claim 4, characterized in that: The flight obstacles include power towers and power lines. If the point clouds of power towers and power lines are completely missing, the specific process of supplementing the spatial range of the flight obstacles within the target range is to collect point cloud data of power lines along the power towers.

6. The method for generating a surface model according to claim 5, characterized in that: If there are power tower and power line point clouds, but the density of the power tower and power line point clouds is too low to cause the power line reflection to be missing and interrupted, the power line point cloud data is fitted, and the fitted power line point cloud data is used to complete the existing power line point cloud data.

7. The method for generating a surface model according to claim 6, characterized in that: The specific process of fitting the power line point cloud data is as follows: Step a, using a catenary model to simulate the natural drooping shape of the power line under the action of gravity; Step b: Calculate the optimal catenary curvature radius ; Step c: Set the optimal catenary curvature radius Substituting it into the calculation formula of the catenary model, the optimal catenary model is obtained, that is, the power line point cloud data is calculated according to the optimal catenary model.

8. The method for generating a surface model according to claim 4, characterized in that: If the flight obstacles also include at least one of construction equipment, lightning protection equipment, buildings, chimneys, and water towers, the elevations of the construction equipment, lightning protection equipment, buildings, chimneys, and water towers are modified by adding the existing point cloud height of the facility to the height of the facility before modification.

9. The method for generating a ground surface model according to claim 4, characterized in that: The step 3 also includes the following steps: Step 4: Obtain the spatial range and elevation changes of the flight obstacles, and update the surface model in the same way as in steps 1 to 3.

Images