CN117607896B

CN117607896B - Coastline broken part point property identification method and device based on multiband point cloud data

Info

Publication number: CN117607896B
Application number: CN202311638807.3A
Authority: CN
Inventors: 姜怀刚; 邢侃; 陈阳; 衣宁; 侯丽; 高姗; 解新玉; 赵佳宁; 王晓宇; 张利奎
Original assignee: Chinese People's Liberation Army Naval Staff Chart Information Center
Current assignee: Chinese People's Liberation Army Naval Staff Chart Information Center
Priority date: 2023-12-01
Filing date: 2023-12-01
Publication date: 2024-06-14
Anticipated expiration: 2043-12-01
Also published as: CN117607896A

Abstract

The embodiment of the application discloses a coastline broken point property identification method and device based on multiband point cloud data, wherein the method comprises the following steps: acquiring multiband LiDAR point cloud data of a coast to be measured; based on the coordinate information of each laser point, respectively determining a geometric characteristic value corresponding to each laser point; acquiring adjacent laser points which are adjacent to each laser point in the multiband LiDAR point cloud data and have different wave bands corresponding to the laser points, and carrying out fusion normalization processing on the echo intensity values of the laser points based on the echo intensity values of the laser points and the echo intensity values of the adjacent laser points to obtain fusion echo intensity values corresponding to the laser points; extracting a coastline broken point of the coast to be measured from the multiband LiDAR point cloud data; and carrying out property identification processing on the shoreline crushed points based on a pre-trained classification model to obtain the property types of the shoreline crushed points.

Description

Coastline broken part point property identification method and device based on multiband point cloud data

Technical Field

The application relates to the technical field of coastline extraction, in particular to a method and a device for identifying coastline broken point properties based on multiband point cloud data.

Background

In practical production and construction practice, the cognitive requirements on the landline quality are often greater than those on the landline position. For example, for mariculture industry, the topography of different coastal zone is suitable for the marine economy animal and plant of the cultivation, so the topography of coastal zone is considered first when the cultivation planning. As another example, in terms of military operations, it is desirable to avoid landing sites at the rock steeplands deep at the shoreside cliffs and at the flat, muddy coasts where the mud is spread over, difficult to travel, and scarce in shades during large-scale landing operations. Therefore, related geographic information products are required to simultaneously meet the two requirements of identifying the property of the coastline and measuring the position of the coastline, but no mature scheme can solve the problem of automatically identifying the property of the coastline at present.

Disclosure of Invention

The embodiment of the application aims to provide a coastline broken point property identification method and device based on multiband point cloud data, which are used for solving the problem that the coastline property is difficult to automatically identify in the prior art.

In order to solve the technical problems, the embodiment of the application is realized as follows:

In one aspect, an embodiment of the present application provides a method for identifying a coastline broken point property based on multiband point cloud data, including:

Acquiring multiband LiDAR point cloud data of a coast to be measured, wherein the multiband LiDAR point cloud data comprises coordinate information and return intensity values of each laser point in the multiband LiDAR point cloud data;

Based on the coordinate information of each laser point, respectively determining a geometric characteristic value corresponding to each laser point;

acquiring adjacent laser points which are adjacent to each laser point in the multiband LiDAR point cloud data and have different wave bands corresponding to the laser points, and carrying out fusion normalization processing on the echo intensity values of the laser points based on the echo intensity values of the laser points and the echo intensity values of the adjacent laser points to obtain fusion echo intensity values corresponding to the laser points;

Extracting the multiband LiDAR point cloud data to extract the coastline broken points of the coast to be measured from the multiband LiDAR point cloud data;

Based on a pre-trained classification model, performing property identification processing on the coastline broken part points according to the geometric characteristic values and the fusion echo intensity values of the coastline broken part points to obtain property types of the coastline broken part points, wherein the property types of the coastline broken part points are used for determining the property of the coastline to be measured, which is constructed by the coastline broken part points, and the classification model is a model which is constructed based on a preset machine learning algorithm and is used for determining the property types of the coastline broken part points.

On the other hand, the embodiment of the application provides a coastline broken point property identification device based on multiband point cloud data, which comprises the following components:

The system comprises a point cloud acquisition module, a point cloud measurement module and a point cloud measurement module, wherein the point cloud acquisition module is used for acquiring multiband LiDAR point cloud data of a coast to be measured, and the multiband LiDAR point cloud data comprises coordinate information and return intensity values of each laser point in the multiband LiDAR point cloud data;

The characteristic determining module is used for respectively determining the geometric characteristic value corresponding to each laser point based on the coordinate information of each laser point;

The intensity fusion module is used for acquiring adjacent laser points which are adjacent to each laser point in the multiband LiDAR point cloud data and have different wave bands corresponding to the laser points, and carrying out fusion normalization processing on the echo intensity values of the laser points based on the echo intensity values of the laser points and the echo intensity values of the adjacent laser points to obtain fusion echo intensity values corresponding to the laser points;

the broken point extraction module is used for extracting the multiband LiDAR point cloud data so as to extract the coastline broken points of the coast to be measured from the multiband LiDAR point cloud data;

the property recognition module is used for carrying out property recognition processing on the coastline broken part points according to the geometric characteristic values and the fusion echo intensity values of the coastline broken part points based on a pre-trained classification model to obtain property types of the coastline broken part points, wherein the property types of the coastline broken part points are used for determining the property of the coastline of the coast to be measured, which is constructed by the coastline broken part points, and the classification model is a model which is constructed based on a preset machine learning algorithm and is used for determining the property types of the coastline broken part points.

In still another aspect, an embodiment of the present application provides a shoreline fragment point property identification apparatus based on multiband point cloud data, including a processor and a memory electrically connected to the processor, the memory storing a computer program, the processor being configured to call and execute the computer program from the memory to implement the above-mentioned shoreline fragment point property identification method based on multiband point cloud data.

In yet another aspect, an embodiment of the present application provides a storage medium storing a computer program executable by a processor to implement the above-described coastline fragment point property identification method based on multi-band point cloud data.

By adopting the technical scheme of the embodiment of the application, the multiband LiDAR point cloud data of the coast to be measured can be obtained, the multiband LiDAR point cloud data can comprise the coordinate information and the return intensity value of each laser point in the multiband LiDAR point cloud data, the geometrical characteristic value corresponding to each laser point is respectively determined based on the coordinate information of each laser point, the adjacent laser points which are adjacent to each laser point and have different wave bands corresponding to the laser points in the multiband LiDAR point cloud data are obtained, the return intensity values of the laser points and the return intensity values of the adjacent laser points are used for fusion normalization processing, the fusion return intensity values corresponding to the laser points are obtained, the multiband LiDAR point cloud data is extracted, the coastline crushed parts of the coast to be measured are extracted from the multiband LiDAR point cloud data, the coastline crushed parts are subjected to property identification processing according to the geometrical characteristic values and the fusion intensity values of the coastline crushed parts of the coastline are trained in advance, the coastline crushed parts are obtained, and the coastline crushed parts are determined to be the machine type model based on the preset type of the coastline crushed parts are established by the machine type model. In this way, the property identification processing can be carried out on the coastline broken points according to the geometric characteristic values and the fusion echo intensity values of the coastline broken points through a pre-trained classification model, so that the property types of the coastline broken points are obtained, the property of the coastline of the coast to be measured constructed by the coastline broken points is determined based on the property types of the coastline broken points, the automatic identification of the property of the coastline is realized, and the property identification accuracy of the coastline of the coast to be measured constructed by the coastline broken points is improved through improving the property identification accuracy of the coastline broken points.

Drawings

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments described in the present application, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.

Fig. 1 is a schematic flow chart of a shoreline fragment point property identification method based on multi-band point cloud data according to an embodiment of the present application;

fig. 2 is a schematic flow chart of a shoreline fragment point property identification method based on multi-band point cloud data according to another embodiment of the present application;

fig. 3 is a schematic flow chart of a shoreline fragment point property identification method based on multi-band point cloud data according to another embodiment of the present application;

FIG. 4 is a schematic diagram of adjacent laser points according to an embodiment of the present application;

FIG. 5 is a schematic diagram of a cube for counting second neighborhood features of point cloud data of a coastal zone in accordance with an embodiment of the application;

fig. 6 is a schematic flow chart diagram of a shoreline fragment point property identification method based on multi-band point cloud data according to another embodiment of the present application;

fig. 7 is a schematic structural diagram of a shoreline fragment point property recognition device based on multiband point cloud data according to an embodiment of the present application;

fig. 8 is a schematic hardware structure of a shoreline crushed point property identification device based on multiband point cloud data according to an embodiment of the application.

Detailed Description

In order to make the technical solution of the present application better understood by those skilled in the art, the technical solution of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application, and it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, shall fall within the scope of the application.

Fig. 1 is a schematic flow chart of a method for identifying characteristics of a crushed coastline point based on multiband point cloud data according to an embodiment of the present application, as shown in fig. 1, the method includes:

in S102, multi-band LiDAR point cloud data of a coast to be measured is acquired.

In this embodiment, in order to avoid repeated identification of coastline properties of a coastline with known properties, whether the coast to be measured is in a training area may be first determined before the multi-band LiDAR point cloud data is acquired, and if it is determined that the coast to be measured is not in the training area, the LiDAR point cloud data of the coast to be measured may be acquired by a plurality of laser radars with different wavebands.

In S104, based on the coordinate information of each laser point, the geometric feature value corresponding to each laser point is determined.

In an implementation, the server may determine a geometric feature value corresponding to each laser point based on the three-dimensional coordinate data corresponding to the laser point.

In S106, adjacent laser points adjacent to each laser point in the multiband LiDAR point cloud data and having different wave bands corresponding to the laser points are obtained, and fusion normalization processing is performed on the echo intensity values of the laser points based on the echo intensity values of the laser points and the echo intensity values of the adjacent laser points, so as to obtain fusion echo intensity values corresponding to the laser points.

In S108, the multiband LiDAR point cloud data is extracted to extract a shoreline fragment point of the coast to be measured from the multiband LiDAR point cloud data.

The coastline broken point may be used to describe a characteristic point of the coastline, and a plurality of coastline broken points may be included in multi-band Light Detection AND RANGING (laser radar) point cloud data.

In S110, based on the pre-trained classification model, the property identification process is performed on the shoreline crushed points according to the geometric feature values and the fused echo intensity values of the shoreline crushed points, so as to obtain the property type of the shoreline crushed points.

The property type of the coastline broken point can be used for determining the property of the coastline to be measured, which is constructed by the coastline broken point, and the classification model is a model which is constructed based on a preset machine learning algorithm and is used for determining the property type of the coastline broken point.

In an implementation, the server may train the classification model constructed by the preset machine learning algorithm based on a large number of training samples corresponding to the training area to obtain a trained classification model. For example, the server may obtain coastal zone point cloud data for a large number of sample coasts, determining a training area for training the classification model.

The sample types corresponding to the laser points in the multi-band LiDAR point cloud data of the coastal zone in the training area can be obtained based on manual marking, and accuracy of the determined sample types is improved. After multi-band LiDAR point cloud data of a coastal zone of a large number of sample coasts are acquired, an area with more types of point supplementing properties of coastlines is used as a training area, and sample types corresponding to each point are marked manually, so that the completeness and effectiveness of a trained classification model are improved.

In addition, in the model training stage or the model using stage, the server can firstly perform noise removal processing on the multiband LiDAR point cloud data, namely, some isolated points with too high or too low elevation (such as electronic noise generated by laser radar equipment and/or noise generated by aerial flying objects and the like) in the data can be removed.

The server can determine the geometrical characteristic value and the fusion echo intensity value corresponding to each coastline broken point respectively through the geometrical characteristic value and the fusion echo intensity value corresponding to each laser point in the multiband LiDAR point cloud data, and then the geometrical characteristic value and the fusion echo intensity value of the coastline broken point can be input into a trained classification model to obtain the property type of the coastline broken point.

After the property type of the coastline broken point is obtained, the server can construct the coastline of the coast to be measured through the coastline broken point, and further determine the coastline property of the coast to be measured through the property type of the coastline broken point. For example, the server may perform smoothing processing on the three-dimensional coordinate data of each coastline broken point according to a preset smoothing manner, determine a coastline broken point corresponding to a start point of the coastline according to a coastline trend, and then connect each coastline broken point after the smoothing processing according to the coastline trend based on the coastline broken point corresponding to the start point to generate a coastline to be measured. In this way, by smoothing the three-dimensional coordinate data of the coastline crushed points and connecting the coastline crushed points after the smoothing according to the coastline trend, a relatively smooth coastline can be generated.

In one embodiment, as shown in fig. 2, the above step S106 may specifically include the following processing of steps S1062 to S1064.

In S1062, multi-band LiDAR point cloud data is partitioned into multiple point cloud data sets based on differences in laser bands.

In S1064, adjacent laser points corresponding to any one of the laser points in the first point cloud data set are searched for in the point cloud data sets other than the first point cloud data set based on the preset search range.

The first point cloud data set may be any one point cloud data set in the point cloud data set, and the preset search range may be determined based on a point cloud density of multiband LiDAR point cloud data.

In an implementation, the server may divide the multiband LiDAR point cloud data into a plurality of point cloud data sets according to different laser bands, and the server may search for one or more adjacent laser points corresponding to any one of the first point cloud data sets from among the point cloud data sets other than the first point cloud data set based on a preset search range with any one of the laser points in the first point cloud data set as a reference laser point. And traversing the laser points in all cloud data sets by the method until one or more adjacent laser points corresponding to each laser point are determined.

In one embodiment, in the case where the multi-band LiDAR point cloud data is dual-band LiDAR point cloud data, correspondingly, as shown in fig. 2, step S106 may specifically further include the following processing of step S1066, that is, after step S1064, step S1066 may further be performed continuously.

In S1066, a sum of the echo intensity value of the laser spot and the echo intensity value of the adjacent laser spot is obtained, and a ratio of the echo intensity value of the laser spot to the sum is determined as a fused echo intensity value corresponding to the laser spot.

In implementation, the dual-band LiDAR point cloud data may be divided into 2 point cloud data sets according to the laser bands, the laser point of any one of the 2 point cloud data sets may be used as a reference laser point, the laser point of another one of the 2 point cloud data sets may be used as a reference laser point, all the laser points in the point cloud data sets corresponding to the two bands are traversed by selecting a certain preset search range, and the laser point closest to each of the reference laser points (i.e., the adjacent laser point corresponding to the reference laser point) is found in the reference laser point.

The server may assign echo intensity values of adjacent laser points to the reference laser point, and calculate the normalized echo intensity values (i.e., fused echo intensity values) of the laser points corresponding to different wavebands through the following formula (1) and formula (2), respectively.

Wherein NDFI _C1 is the fused echo intensity value of a certain reference laser spot under the C1 wave band, gamma ₁ is the echo intensity value of the reference laser spot under the C1 wave band, and gamma ₂ is the echo intensity value of the adjacent laser spot corresponding to the reference laser spot under the C2 wave band.

NDFI _C2 is the fused echo intensity value of a certain reference laser spot under the C2 wave band, gamma ₂ is the echo intensity value of the reference laser spot under the C2 wave band, and gamma ₁ is the echo intensity value of the adjacent laser spot corresponding to the reference laser spot under the C1 wave band.

If the adjacent laser spot corresponding to the reference laser spot is not searched within the preset search range, the echo intensity value of the adjacent laser spot corresponding to the reference laser spot may be assigned to 0, that is, the echo intensity value of the reference laser spot in other bands may be assigned to 0. Thus, most laser points after data fusion can have echo intensity values of two wavebands.

In one embodiment, in the case where the multi-band LiDAR point cloud data is multi-band LiDAR point cloud data greater than two bands, correspondingly, as shown in fig. 3, step S106 may specifically further include the following steps S1068 to S10610, that is, after S1064, S1068 to S10610 may be further executed.

In S1068, a first sum value between echo intensity values of adjacent laser points is acquired, and a difference value between the first sum value and the echo intensity value of the laser point is acquired.

In S10610, a second sum of the echo intensity values of the adjacent laser points and the echo intensity value of the laser point is obtained, and a ratio between the difference and the second sum is determined as a fused echo intensity value corresponding to the laser point.

In an implementation, taking multi-band LiDAR point cloud data as three-band LiDAR point cloud data as an example, the server can divide the point cloud data into 3 point cloud data sets according to different laser bands. The server may take the point cloud data of any one of the 3 point cloud data sets as reference data, and the point cloud data of the other two wave bands as reference data, traverse all points in the reference data and the reference data by selecting a certain preset search range, and find a laser point closest to each laser point in the reference data as a reference laser point.

For example, assuming that the three-band LiDAR point cloud data includes point cloud data in the C1 band, point cloud data in the C2 band, and point cloud data in the C3 band, the preset search range is a range configured with r as a search radius, and as shown in fig. 4, a search range configured with r as a search radius may include a plurality of laser points in the C1 band, a plurality of laser points in the C2 band, and a plurality of laser points in the C3 band.

And traversing the point cloud data in the C2 wave band in the preset searching range by taking the laser point a in the C1 wave band as reference data, and finding a laser point b closest to the laser point a, namely the laser point b can be a reference laser point corresponding to the laser point a in the C2 wave band, and correspondingly, the laser point C in the C3 wave band is a reference laser point corresponding to the laser point a in the C3 wave band, namely the laser point b and the laser point C are adjacent laser points of the laser point a.

The server may assign the echo intensity value of the reference laser spot to the reference laser spot, and calculate the normalized echo intensity values (i.e., fused echo intensity values) of the laser spots corresponding to different bands through the following formulas (3) to (5).

Wherein NDFI _C1 is the fused echo intensity value of a certain reference laser spot under the C1 wave band, gamma ₁ is the echo intensity value of the reference laser spot under the C1 wave band, gamma ₂ is the echo intensity value of the adjacent laser spot of the reference laser spot under the C2 wave band, and gamma ₃ is the echo intensity value of the adjacent laser spot of the reference laser spot under the C3 wave band.

NDFI _C2 is the fused echo intensity value of a reference laser spot in the C2 band, γ ₂ is the echo intensity value of the reference laser spot in the C2 band, γ ₁ is the echo intensity value of an adjacent laser spot of the reference laser spot in the C1 band, and γ ₃ is the echo intensity value of an adjacent laser spot of the reference laser spot in the C3 band.

NDFI _C3 is the fused echo intensity value of a certain reference laser spot under the C3 wave band, gamma ₃ is the echo intensity value of the reference laser spot under the C1 wave band, gamma ₁ is the echo intensity value of the adjacent laser spot of the reference laser spot under the C1 wave band, and gamma ₂ is the echo intensity value of the reference laser spot under the C2 wave band.

In addition, taking multiband LiDAR point cloud data as n-band LiDAR point cloud data greater than three bands as an example, n is a positive integer greater than 3. The server can divide the point cloud data into n point cloud data sets according to different laser wave bands, takes the point cloud data of any one wave band in any one point cloud data set as reference data and the point cloud data of other wave bands as reference data, traverses all laser points in the reference data and the reference data by selecting a certain preset search range, and searches the laser point closest to each laser point in the reference data as a reference laser point.

The server may assign the echo intensity value of the reference laser spot to the reference laser spot, and calculate the echo intensity values (i.e., fused echo intensity values) of the laser spots corresponding to different bands after normalization by the following formulas (6) to (8).

Wherein NDFI _C1 is the fusion echo intensity value of a certain reference laser spot under the C1 wave band, gamma ₁ is the echo intensity value of the reference laser spot under the C1 wave band, gamma _n is the echo intensity value (n is more than or equal to 2) of the adjacent laser spot of the reference laser spot under the Cn wave band.

Wherein NDFI _C2 is the fusion echo intensity value of a certain reference laser spot under the C2 wave band, gamma ₂ is the echo intensity value of the reference laser spot under the C2 wave band, gamma ₁ is the echo intensity value of the adjacent laser spot of the reference laser spot under the C1 wave band, gamma _n is the echo intensity value of the adjacent laser spot of the reference laser spot under the Cn wave band (n is more than or equal to 3).

Wherein NDFI _cCn is the fused echo intensity value of a certain reference laser spot under Cn wave band, gamma ₁ is the echo intensity value of the reference laser spot under C1 wave band, and gamma _n-1 is the echo intensity value of the adjacent laser spot of the reference laser spot under Cn-1 wave band.

In one embodiment, the geometric feature values corresponding to the laser points may include 26-dimensional geometric feature values determined based on coordinate information of the laser points, wherein the 26-dimensional features may include 16 three-dimensional geometric feature values, 6 two-dimensional geometric features, and 4 cube neighborhood feature values.

The server can determine adaptive areas of sample characteristic values corresponding to the multiband LiDAR point cloud data respectively, wherein the adaptive areas comprise a certain number of points. The number of points n _opt can be calculated from shannon entropy. Assuming that the number of neighborhood points is n, the covariance matrix and its eigenvalue λ ₁,λ₂,λ₃ can be calculated from the three-dimensional coordinates (x, y, z) of the n nearest neighbors. The shannon entropy is calculated as follows:

E_λ＝-e₁lne₁-e₂lne₂-e₃lne₃， (9)

e_i＝λ_i/Σ_λ ， (10)

Σ_λ＝λ₁+λ₂+λ₃， (11)

Where e ₁,e₂,e₃ is the normalized eigenvalue and Σ _λ is the eigenvalue sum. When Σ _λ is minimum, the value of n is the optimal value of n _opt, and the sample characteristic value corresponding to each laser point in the unit area can be calculated according to the three-dimensional coordinate data of each laser point in the unit area containing n _opt points.

The range of the adaptive area can be determined according to any one of the point cloud density of the coastal zone point cloud data and the designated shape area. If the range of the self-adaptive area is determined according to the point cloud density of the point cloud data of the coastal zone, the point cloud density threshold value can be set in advance, and the area with the point cloud density greater than or equal to the point cloud density threshold value is determined to be the self-adaptive area. If the range of the adaptive area is determined according to the specified shape area, the specified shape can be preset, and the specified shape is determined as the adaptive area in the area where the coastal zone point cloud data belongs.

The server can calculate geometrical characteristic values corresponding to the laser points according to the three-dimensional coordinate data of the laser points.

Wherein, the three-dimensional geometric features comprise: normalized eigenvalue E ₁, normalized eigenvalue E ₂, normalized eigenvalue E ₃, and linear index L _λ, flatness index P _λ, scattered index S _λ, total variance O _λ, anisotropy index a _λ, characteristic entropy E _λ, curvature change rate C _λ, vertical index V _λ calculated from the above three normalized eigenvalues, the calculation formulas are as follows:

L_λ＝(e₁-e₂)/e₁， (12)

P_λ＝(e₂-e₃)/e₁， (13)

S_λ＝e₃/e₁， (14)

A_λ＝(e₁-e₃)/e₁， (16)

C_λ＝e₁/(e₁+e₂+e₃)， (18)

V_λ＝1-e₃。 (19)

In addition, the local point density D, the local point cloud radius r, the elevation z, the elevation change range Δh, and the elevation standard deviation H _std also belong to three-dimensional geometric features.

In a coastal zone scene, as the point clouds of ground features such as rocks, epitaxial stones and seawater have vertical spatial distribution characteristics, the same laser beam may have multiple echo intensity values when the coastal zone point cloud data are acquired, so that the point density formed by the object in a local range after horizontal projection is larger. Two-dimensional geometric features tend to be complementary to three-dimensional features, and are therefore also characterized to some extent. According to the plane coordinates of the coastal zone point cloud data, the following 6 two-dimensional geometric features can be calculated: neighborhood point density D _2D, neighborhood point radius r _2D, two-dimensional eigenvalue sum Σ _λ2D, eigenvalue ratio μ, normalized eigenvalue e _1-2D, and normalized eigenvalue e _2-2D.

As shown in fig. 5, a cube with a height not limited and a length and a width d is constructed with (x, y) as the center. Calculating three-dimensional coordinate data of all coast point cloud data to be measured in the cube, and obtaining cube neighborhood characteristics corresponding to the coast point cloud data to be measured: elevation median H, elevation standard deviation H _cube-std, number of points n _cube-std, elevation range Δh _p,

ΔH_P＝H_p90-H_p10， (20)

Where H _p90 and H _p10 are the 90 and 10 percentile points, respectively, of all point elevations within the cube. Assuming 100 points in the cube, all points are ordered from low to high, the first point is 1m (meter) high, the second point is 2m high, the third point is 3 m..third point is 10 m..third point is 90 m..third point is 100m of the first hundred point, then the tenth point is 10m and the nineteenth point is 90m, respectively, the 10 percentile and the 90 percentile of all point elevations in the cube, and the elevation range is 90-10=80 m.

In addition to the 26-dimensional sample geometric feature values, the geometric feature values corresponding to the laser points may further include waveform parameter features: the echo intensity value gamma, the echo times k and the waveform parameter characteristics are recognized as being very characteristic, and are used as the characteristic values of another two-dimensional sample.

In one embodiment, as shown in fig. 6, the above step S108 may specifically further include the following processing of steps S1082 to S1088.

In an implementation, the server may perform coordinate transformation on the LiDAR point cloud data, and transform three-dimensional coordinate data corresponding to the LiDAR point cloud data into a coordinate system (for example, 2000 national geodetic coordinate system) and an elevation reference (for example, 1985 national elevation reference) according to the current application requirement.

Secondly, calculating according to a formula (21) to obtain an average high tide level elevation value H _MHWS,

Wherein,For the coastline elevation of the coast to be measured (from the average sea surface), ζ is the average sea surface elevation of the location where the coast to be measured is located in the 1985 national elevation reference.

Finally, discarding the LiDAR point cloud data with the elevation value smaller than the elevation threshold value according to the LiDAR point cloud data after coordinate conversion. Wherein H _MHWS may be set as an elevation threshold, points contained in the LiDAR point cloud data less than the elevation threshold may be considered points in the sea, and points contained in the LiDAR point cloud data greater than or equal to the elevation threshold may be considered points on the sea to be measured.

In S1082, a size of the coarse mesh is determined based on the preset chart mapping scale information and the point cloud density of the multi-band LiDAR point cloud data.

In S1084, the coastline broken points in the multiband LiDAR point cloud data are coarsely extracted using a plurality of coarse grids.

In practice, the server may first determine the size of the coarse grid according to the "airborne LiDAR data acquisition specification," which specifies grid spacing and point cloud density when interpolating DEMs (Digital Elevation Model ) using LiDAR point cloud data. Based on the point cloud density information in the multi-band LiDAR point cloud data, the size of the coarse grid may be selected with reference to the specification. The correspondence between the chart mapping scale information, the point cloud density information, and the size of the coarse mesh is shown in table 1.

TABLE 1DEM grid spacing

Drawing scale	DEM grid spacing/m	Point cloud Density/(Point/m ²)
			1:500	0.5	≥16
1:1 000	1.0	≥4
			1:2 000	2.0	≥1
1:5 000	2.5	≥1
			1:10 000	5.0	≥0.25

In table 1, the scale of the chart is chart scale information, the distance between DEM grids is the size of coarse grids, and the density of the point cloud is point cloud density information.

Secondly, according to three-dimensional coordinate data corresponding to the LiDAR point cloud data, counting an outsourcing rectangle (X _max,Y_max;X_min,Y_min) of the LiDAR point cloud data, and according to the selected grid side length a and plane coordinates (X, y) of the LiDAR point cloud data, calculating an outsourcing rectangle rank value M, N and grid rank values m and n of each point through a formula (22).

Wherein [ ] in the formula represents rounding, M is an outsourcing rectangle row value, N is an outsourcing rectangle column value, M is a dot row value, N is a dot column value, and a is a grid side length.

After the grid organization of the multiband LiDAR point cloud data is realized, in order to avoid the extraction of the non-outer side edges, filling processing is required to be carried out on grid holes caused by a low-altitude area existing in a land part, edge detection is carried out on the filled data, edge grid data is obtained through extraction, then the maximum communication area detection is carried out on the edge grid, the position of a coastline coarse grid is determined, and finally LiDAR point cloud data in the coarse grid where the coastline is located is obtained.

In the embodiment, the coarse extraction result of the broken points of the coastline is obtained by carrying out regularized coarse grid management on disordered LiDAR point cloud data and extracting the edge grid where the coastline is positioned by utilizing an edge detection algorithm, so that the accuracy of the coarse extraction of the broken points of the coastline is improved.

In S1086, the size of the fine mesh is determined based on a predetermined fine mesh size determination method corresponding to the point cloud density and chart mapping scale information of the multiband LiDAR point cloud data.

In S1088, the obtained coastline crushed portion point rough extraction result is subjected to fine extraction by using a plurality of fine mesh networks, so as to obtain a plurality of coastline crushed portion points of the coast to be measured.

In implementations, the server may determine a point cloud density size from the point cloud density information. In this embodiment, the following three fine mesh size determining modes corresponding to the point cloud density information and the chart mapping scale information are specifically included:

In the first mode, if the point cloud density is high, the maximum resolution (0.1 mm) of the human eye on the chart of the chart is determined as the side length of the fine mesh.

In the second mode, if the density of the point cloud is relatively sparse, the width (0.2 mm) of a common shoreline in sea chart imaging can be used as the side length of the fine grid.

The third mode has higher requirements on the density distribution of the point clouds along with the increase of the chart forming scale, but the two modes for determining the size of the fine mesh are not applicable any more because of the existence of the area with over-dense or over-sparse density of the point clouds, and the determination of the size of the fine mesh is required to be based on the fact that the fine mesh of the continuous coastline can be formed.

After determining the size of the fine grid, the outsourcing rectangle (X _max,Y_max;X_min,Y_min) of the multiband LiDAR point cloud data can be counted according to the three-dimensional coordinate data corresponding to the multiband LiDAR point cloud data, and the outsourcing rectangle rank value M, N and the grid rank value m and n where each point is located are calculated according to the selected grid side length and the plane coordinates (X, y) of the LiDAR point cloud data through the formula (20).

After fine grid division is performed on multi-band LiDAR point cloud data, an expansion algorithm can be adopted to process fine grid data break points caused by uneven point cloud density, holes inside the fine grid are filled to avoid the edge of an inland low-altitude area from being extracted, then corrosion treatment is performed on the fine grid data to recover expanded data, and finally fine edge grid extraction is performed.

When fine extraction is performed on the coastline crushed points after coarse extraction by using a plurality of fine grids, the fine grids without LiDAR point cloud data are firstly removed, and then, among at least one point contained in each fine grid, a point with an elevation value closest to H _MHWS is taken as the coastline crushed point, and unreasonable points in the fine grids are removed by setting a difference threshold delta between the point and H _MHWS.

In the embodiment, the coastline broken point in the LiDAR point cloud data is extracted through the thick and thin grid, and the regularized grid management of disordered LiDAR point cloud data is realized, so that the coastline broken point data is obtained, and a data basis is provided for automatic identification of coastline properties in the follow-up process.

In summary, particular embodiments of the present subject matter have been described. Other embodiments are within the scope of the following claims. In some cases, the actions recited in the claims can be performed in a different order and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may be advantageous.

The method for identifying the coastline broken point property based on the multiband point cloud data provided by the embodiment of the application is based on the same thought, and the embodiment of the application also provides a device for identifying the coastline broken point property based on the multiband point cloud data.

Fig. 7 is a schematic structural diagram of a shoreline crushed point property identification device based on multiband point cloud data according to an embodiment of the application, and as shown in fig. 7, the shoreline crushed point property identification device based on multiband point cloud data includes:

The point cloud acquisition module 701 is configured to acquire multiband LiDAR point cloud data of a coast to be measured, where the multiband LiDAR point cloud data includes coordinate information and a return intensity value of each laser point in the multiband LiDAR point cloud data;

The feature determining module 702 is configured to determine a geometric feature value corresponding to each laser point based on the coordinate information of each laser point;

The intensity fusion module 703 is configured to obtain adjacent laser points adjacent to each laser point in the multi-band LiDAR point cloud data and having different wave bands corresponding to the laser points, and perform fusion normalization processing on the echo intensity values of the laser points based on the echo intensity values of the laser points and the echo intensity values of the adjacent laser points, so as to obtain fusion echo intensity values corresponding to the laser points;

The broken point extraction module 704 is configured to perform extraction processing on the multi-band LiDAR point cloud data, so as to extract a coastline broken point of the coast to be measured from the multi-band LiDAR point cloud data;

The property recognition module 705 is configured to perform property recognition processing on the shoreline broken portion point according to the geometric feature value and the fused echo intensity value of the shoreline broken portion point based on a pre-trained classification model, so as to obtain a property type of the shoreline broken portion point, where the property type of the shoreline broken portion point is used for determining a property of a shoreline to be measured, which is constructed by the shoreline broken portion point, and the classification model is a model constructed based on a preset machine learning algorithm and used for determining the property type of the shoreline broken portion point.

In one embodiment, the intensity fusion module 703 is configured to:

dividing the multiband LiDAR point cloud data into a plurality of point cloud data sets based on the difference of laser wave bands;

Based on a preset searching range, searching adjacent laser points corresponding to any one laser point in a first point cloud data set in point cloud data sets except the first point cloud data set, wherein the first point cloud data set is any one point cloud data set in the point cloud data sets.

In one embodiment, the preset search range is determined based on a point cloud density of the multi-band LiDAR point cloud data.

In one embodiment, the multi-band LiDAR point cloud data is dual-band LiDAR point cloud data, and the intensity fusion module 703 is configured to:

and acquiring the sum value of the echo intensity value of the laser point and the echo intensity value of the adjacent laser point, and determining the ratio of the echo intensity value of the laser point to the sum value as a fusion echo intensity value corresponding to the laser point.

In one embodiment, the multi-band LiDAR point cloud data is multi-band LiDAR point cloud data greater than two bands, and the intensity fusion module 703 is configured to:

acquiring a first sum value between echo intensity values of the adjacent laser points, and acquiring a difference value between the first sum value and the echo intensity value of the laser point;

and acquiring a second sum value between the echo intensity value of the adjacent laser point and the echo intensity value of the laser point, and determining the ratio of the difference value to the second sum value as a fused echo intensity value corresponding to the laser point.

In one embodiment, the geometric feature values corresponding to the laser points include 26-dimensional geometric feature values determined based on coordinate information of the laser points, wherein the 26-dimensional features include 16 three-dimensional geometric feature values, 6 two-dimensional geometric features, and 4 cube neighborhood feature values.

In one embodiment, the crushed point extraction module 704 is configured to:

Determining the size of a coarse grid based on preset chart scale information and the point cloud density of the multiband LiDAR point cloud data;

coarsely extracting coastline broken points in the multiband LiDAR point cloud data by utilizing a plurality of coarse grids;

determining the size of a fine grid based on a preset fine grid size determination mode corresponding to the point cloud density of the multiband LiDAR point cloud data and the chart scale information;

And carrying out fine extraction on the obtained rough extraction result of the shoreline broken points by using a plurality of fine grids to obtain a plurality of shoreline broken points of the coast to be measured.

By adopting the device of the embodiment of the application, the multiband LiDAR point cloud data of the coast to be measured can be obtained, the multiband LiDAR point cloud data can comprise the coordinate information and the return intensity value of each laser point in the multiband LiDAR point cloud data, the geometrical characteristic value corresponding to each laser point is respectively determined based on the coordinate information of each laser point, the adjacent laser points which are adjacent to each laser point and have different wave bands corresponding to the laser points in the multiband LiDAR point cloud data are obtained, the return intensity value of the adjacent laser points is based on the return intensity value of the laser points, fusion normalization processing is carried out on the return intensity value of the laser points to obtain the fusion return intensity value corresponding to the laser points, the multiband LiDAR point cloud data is extracted to obtain the point of the crushed coastline of the coast to be measured, the point of the coastline is subjected to the characteristic recognition processing according to the geometrical characteristic value and the return intensity value of the point of the crushed coastline of the coastline on the basis of a pre-trained classification model, and the type of the crushed coastline is determined based on the model of the type of the crushed coastline of the coastline is based on the machine learning algorithm. In this way, the property identification processing can be carried out on the coastline broken points according to the geometric characteristic values and the fusion echo intensity values of the coastline broken points through a pre-trained classification model, so that the property types of the coastline broken points are obtained, the property of the coastline of the coast to be measured constructed by the coastline broken points is determined based on the property types of the coastline broken points, the automatic identification of the property of the coastline is realized, and the property identification accuracy of the coastline of the coast to be measured constructed by the coastline broken points is improved through improving the property identification accuracy of the coastline broken points.

It should be understood by those skilled in the art that the coastline broken point property recognition device based on the multiband point cloud data in fig. 7 can be used to implement the coastline broken point property recognition method based on the multiband point cloud data described above, and the detailed description thereof should be similar to that of the method section described above, so that the detailed description thereof is omitted herein for avoiding complexity.

Based on the same thought, the embodiment of the application also provides a coastline broken point property identification device based on multi-band point cloud data, as shown in fig. 8. The shoreline fragment point property identification device based on the multiband point cloud data may have a relatively large difference due to different configurations or performances, may include one or more processors 801 and a memory 802, and may have one or more storage applications or data stored in the memory 802. Wherein the memory 802 may be transient storage or persistent storage. The application stored in memory 802 may include one or more modules (not shown in the figures), each of which may include a series of computer-executable instructions for identifying coastline point properties in a device based on multi-band point cloud data. Still further, the processor 801 may be configured to execute a series of computer executable instructions in the memory 802 on a shoreline fragment point property identification device based on multi-band point cloud data in communication with the memory 802. The shoreline fragment point property identification device based on the multi-band point cloud data may also include one or more power sources 803, one or more wired or wireless network interfaces 804, one or more input/output interfaces 805, one or more keyboards 806.

In particular, in this embodiment, a shoreline point property identification device based on multi-band point cloud data includes a memory, and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer-executable instructions in the shoreline point property identification device based on multi-band point cloud data, and configured to be executed by one or more processors, the one or more programs including computer-executable instructions for:

By adopting the device provided by the embodiment of the application, the multiband LiDAR point cloud data of the coast to be measured can be obtained, the multiband LiDAR point cloud data can comprise the coordinate information and the return intensity value of each laser point in the multiband LiDAR point cloud data, the geometrical characteristic value corresponding to each laser point is respectively determined based on the coordinate information of each laser point, the adjacent laser points which are adjacent to each laser point and are different in wave bands corresponding to the laser points in the multiband LiDAR point cloud data are obtained, the return intensity value of the adjacent laser points is based on the return intensity value of the laser points, fusion normalization processing is carried out on the return intensity value of the laser points to obtain the fusion return intensity value corresponding to the laser points, the multiband LiDAR point cloud data is extracted to obtain the point of the crushed coastline of the coast to be measured, the point of the crushed coastline is subjected to the characteristic recognition processing according to the geometrical characteristic value and the return intensity value of the crushed coastline of the point of the coastline based on the pre-trained classification model, and the type of the crushed coastline is determined based on the model of the type of the crushed coastline of the coastline is based on the machine learning algorithm. In this way, the property identification processing can be carried out on the coastline broken points according to the geometric characteristic values and the fusion echo intensity values of the coastline broken points through a pre-trained classification model, so that the property types of the coastline broken points are obtained, the property of the coastline of the coast to be measured constructed by the coastline broken points is determined based on the property types of the coastline broken points, the automatic identification of the property of the coastline is realized, and the property identification accuracy of the coastline of the coast to be measured constructed by the coastline broken points is improved through improving the property identification accuracy of the coastline broken points.

The embodiment of the application also provides a storage medium, which stores one or more computer programs, the one or more computer programs comprise instructions, when the instructions are executed by an electronic device comprising a plurality of application programs, the electronic device can be caused to execute the processes of the shoreline fragment point property identification method embodiment based on the multiband point cloud data, and the same technical effects can be achieved, so that repetition is avoided, and no further description is provided herein.

The system, apparatus, module or unit set forth in the above embodiments may be implemented in particular by a computer chip or entity, or by a product having a certain function. One typical implementation is a computer. In particular, the computer may be, for example, a personal computer, a laptop computer, a cellular telephone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

For convenience of description, the above devices are described as being functionally divided into various units, respectively. Of course, the functions of each element may be implemented in the same piece or pieces of software and/or hardware when implementing the present application.

It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

In one typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include volatile memory in a computer-readable medium, random Access Memory (RAM) and/or nonvolatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of computer-readable media.

Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, does not include transitory computer-readable media (transmission media), such as modulated data signals and carrier waves.

It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises the element.

The application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.

In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments. In particular, for system embodiments, since they are substantially similar to method embodiments, the description is relatively simple, as relevant to see a section of the description of method embodiments.

The foregoing is merely exemplary of the present application and is not intended to limit the present application. Various modifications and variations of the present application will be apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. which come within the spirit and principles of the application are to be included in the scope of the claims of the present application.

Claims

1. The coastline broken point property identification method based on the multiband point cloud data is characterized by comprising the following steps of:

Based on a pre-trained classification model, performing property identification processing on the coastline broken part points according to the geometric characteristic values and the fusion echo intensity values of the coastline broken part points to obtain property types of the coastline broken part points, wherein the property types of the coastline broken part points are used for determining the property of the coastline to be measured, which is constructed by the coastline broken part points, and the classification model is a model which is constructed based on a preset machine learning algorithm and is used for determining the property types of the coastline broken part points;

the geometric characteristic value comprises a three-dimensional geometric characteristic value, a two-dimensional geometric characteristic and a cube neighborhood characteristic value, wherein the three-dimensional geometric characteristic value is characterized, and the two-dimensional geometric characteristic can form complementarity with the three-dimensional geometric characteristic value and has characterization.

2. The method of claim 1, wherein the acquiring adjacent laser points in the multi-band LiDAR point cloud data that are adjacent to each laser point and that differ in the band corresponding to the laser point comprises:

3. The method of claim 2, the preset search range being determined based on a point cloud density of the multi-band LiDAR point cloud data.

4. The method of claim 3, wherein the multi-band LiDAR point cloud data is dual-band LiDAR point cloud data, and the performing fusion normalization processing on the return intensity values of the laser points based on the return intensity values of the laser points and the return intensity values of the neighboring laser points to obtain fusion return intensity values corresponding to the laser points comprises:

5. The method of claim 3, wherein the multi-band LiDAR point cloud data is multi-band LiDAR point cloud data greater than two bands, and the performing fusion normalization on the return intensity values of the laser points based on the return intensity values of the laser points and the return intensity values of the neighboring laser points to obtain the fused return intensity values corresponding to the laser points comprises:

6. The method of claim 1, wherein the geometric feature values corresponding to the laser points comprise 26-dimensional geometric feature values determined based on coordinate information of the laser points, wherein the 26-dimensional geometric features comprise 16 three-dimensional geometric feature values, 6 two-dimensional geometric features, and 4 cube neighborhood feature values.

7. The method of claim 1, wherein the extracting the multi-band LiDAR point cloud data to extract the coastline break point of the coast to be measured from the multi-band LiDAR point cloud data comprises:

8. Coastline broken section point nature recognition device based on multiband point cloud data, characterized by comprising:

The property recognition module is used for carrying out property recognition processing on the coastline broken part points according to the geometric characteristic values and the fusion echo intensity values of the coastline broken part points based on a pre-trained classification model to obtain property types of the coastline broken part points, wherein the property types of the coastline broken part points are used for determining the property of the coastline of the coast to be measured, which is constructed by the coastline broken part points, and the classification model is a model which is constructed based on a preset machine learning algorithm and is used for determining the property types of the coastline broken part points;

9. A shoreline fragment point property identification device based on multi-band point cloud data, characterized by comprising a processor and a memory electrically connected to the processor, the memory storing a computer program, the processor being configured to invoke and execute the computer program from the memory to implement the multi-band point cloud data-based shoreline fragment point property identification method of any one of claims 1-7.

10. A storage medium storing a computer program for execution by a processor to implement the shoreline fragment point property identification method based on multiband point cloud data of any one of claims 1-7.