CN106709473B - Voxel-based airborne LIDAR road extraction method - Google Patents
Voxel-based airborne LIDAR road extraction method Download PDFInfo
- Publication number
- CN106709473B CN106709473B CN201710044460.8A CN201710044460A CN106709473B CN 106709473 B CN106709473 B CN 106709473B CN 201710044460 A CN201710044460 A CN 201710044460A CN 106709473 B CN106709473 B CN 106709473B
- Authority
- CN
- China
- Prior art keywords
- voxel
- voxels
- road
- data
- point cloud
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/182—Network patterns, e.g. roads or rivers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T15/00—3D [Three Dimensional] image rendering
- G06T15/08—Volume rendering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Graphics (AREA)
- Multimedia (AREA)
- Image Processing (AREA)
Abstract
A voxel-based airborne LIDAR road extraction method belongs to the technical field of remote sensing data processing; the method comprises the following steps: regularizing original airborne LIDAR point cloud data into a 3D volume metadata set; separating a set of geobody metadata from the 3D voxel data set; searching road seed voxels from the ground voxels by using the laser intensity characteristics of road reflection, and extracting voxels containing data points which are in 3D communication with the road seed voxels and have intensity differences smaller than a set threshold value as road extraction results; the method comprehensively utilizes the intensity information and the implicit elevation information among the voxels, realizes the extraction of the 3D road on the basis of the 3D connectivity construction theory, enables the detection of the target information in the point cloud data to be converted into a search marking mode based on the voxel space neighborhood relationship, well utilizes the implicit neighborhood relationship among the voxels in the 3D voxel data, and is beneficial to the development of the processing and application of airborne LIDAR point cloud data based on the voxel theory.
Description
Technical Field
The invention belongs to the technical field of remote sensing data processing, and particularly relates to a voxel-based airborne LIDAR road extraction method.
Background
Roads are important infrastructure related to national economy, and timely, accurate and efficient acquisition and update of road information are of great significance to traffic management, city planning, automatic vehicle navigation and emergency transaction processing. The airborne laser radar (LIDAR) technology can quickly acquire three-dimensional (3D) point cloud data with high surface precision, And the rapid extraction of 3D information of roads becomes possible due to the appearance of the technology. Classical road extraction methods can be classified into the following categories: a clustering-based method, a mathematical morphology method and a digital image processing method. The data structure form adopted by the method mainly comprises the following steps: point cloud or grid mesh. The data structure of the grid network expresses the 3D point cloud as a 2.5D regular grid network, information loss inevitably exists in the grid network, and road information may exist in each echo, so that the road extraction result based on the 2.5D data structure cannot be complete; the point cloud data structure contains all the information of the original LIDAR point cloud, but is limited by the limitations of the point cloud data structure itself, i.e., it is difficult to utilize the spatial relationship of the data points, resulting in a road extraction method based on the point cloud having low efficiency. Therefore, the data structure adopted by the existing method is not beneficial to repeatedly exerting the technical advantages of airborne LIDAR point cloud true 3D, and a simpler true 3D data structure is needed to be selected and a corresponding data processing technology is designed to complete 3D road extraction based on airborne LIDAR.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides an airborne LIDAR road extraction method based on voxels.
The technical scheme of the invention is as follows:
a voxel-based airborne LIDAR road extraction method comprises the following steps:
step 1: reading original airborne LIDAR point cloud data;
step 2: regularizing original airborne LIDAR point cloud data into a 3D volume metadata set;
step 2.1: removing abnormal data from original airborne LIDAR point cloud data to obtain an abnormal data removing set:
step 2.1.1: dividing the space where the original airborne LIDAR point cloud data is located into M × N × U3D grids, and mapping each original airborne LIDAR point cloud data to each grid unit, wherein the grids containing the original airborne LIDAR point cloud data are called black grids, and the grids not containing the original airborne LIDAR point cloud data are called white grids;
step 2.1.2: positioning black grids with highest elevation and lowest elevation in M x N stand columns in the M x N x U3D grid network as candidate abnormal grid units to obtain candidate abnormal data sets;
step 2.1.3: and comparing the elevation difference of the average elevation of each candidate abnormal grid unit in the candidate abnormal grid data set with the average elevation of the black grids in the peripheral given neighborhood, if the elevation difference is greater than a given threshold Ted, removing the original airborne LIDAR point cloud data contained in the candidate abnormal grid unit as abnormal data, otherwise, retaining the original airborne LIDAR point cloud data contained in the candidate abnormal grid unit, and finally obtaining an abnormal data removing set.
Step 2.2: regularizing the removed anomaly dataset into a 3D volume metadata set:
step 2.2.1: representing the 3D spatial range with an axial bounding box that excludes the anomalous data set;
step 2.2.2: calculating the voxel resolution in the x, y and z directions, namely the voxel size, and determining the voxel resolution (delta x, delta y, delta z) in the x, y and z directions according to the average point distance of the data points in the abnormal data set;
step 2.2.3: dividing the axial bounding box according to the voxel resolution in the x, y and z directions to obtain a 3D voxel grid, wherein each 3D voxel grid unit is called a voxel;
step 2.2.4: and mapping the removed abnormal data set to a 3D voxel grid, judging whether the voxel contains removed abnormal data points, if so, assigning the intensity of the contained data points to the voxel containing the data points, if so, assigning the average value of the intensities of all the data points contained in the voxel, otherwise, assigning 0 to the voxel to obtain a 3D voxel data set, wherein the 0-value voxel and the non-0-value voxel respectively represent a background voxel and a target voxel, and completing the regularization of the removed abnormal data set.
And step 3: separating a set of geobody metadata from the 3D voxel data set;
and 4, step 4: searching road seed voxels from the ground voxels by using the intensity characteristics of laser reflected by the road, extracting target voxels which are in 3D communication with the road seed voxels and have intensity difference smaller than a set threshold value, namely a road body metadata set consisting of voxels containing data points, namely a road extraction result:
step 4.1: searching a road seed voxel Vs from the ground voxel according to the laser intensity information reflected by the road;
step 4.2: starting from target voxels in a given neighborhood scale of Vs in sequence, traversing all adjacent target voxels with a depth priority, and if the adjacent target voxels are communicated with the Vs through paths and the intensity difference between the adjacent target voxels and the seed voxels is smaller than a given threshold value Ti, marking as a road voxel:
step 4.2.1: initialization: setting an initial stack for storing seed voxels, and marking the seed voxels as road voxels;
step 4.2.2: popping an element from the stack top of the initial stack, acquiring unmarked target voxels in a given neighborhood of the element, comparing the intensity difference between the target voxels and the seed voxels, if the intensity difference is less than a given threshold value Ti, marking the target voxels as road voxels and storing the road voxels in the initial stack;
step 4.2.3: judging whether the initial stack is empty or not, if so, marking all target voxels in the 3D voxel grid, which are communicated with the Vs by paths, as road voxels; otherwise, step 4.2.2 is performed.
Has the advantages that: compared with the prior art, the voxel-based airborne LIDAR road extraction method has the following advantages:
the method comprehensively utilizes the intensity information and the elevation information implied among the voxels, realizes the extraction of the 3D road on the basis of the 3D connectivity construction theory, enables the detection of the target information in the point cloud data to be converted into a search marking mode based on the voxel space neighborhood relationship, well utilizes the neighborhood relationship implied among the voxels in the 3D voxel data, and is beneficial to the development of the processing and application of the airborne LIDAR point cloud data based on the voxel theory.
Drawings
Fig. 1 is a flowchart of a voxel-based airborne LIDAR road extraction method according to an embodiment of the present invention;
FIG. 2 is a flowchart illustrating step 2 of an embodiment of the present invention;
fig. 3 is a schematic diagram of calculating the side length of the grid in step 2 according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of the calculation of the average dot spacing in step 2 according to an embodiment of the present invention;
FIG. 5 is a flowchart illustrating step 4 in accordance with an embodiment of the present invention;
FIG. 6 is an experimental data diagram according to an embodiment of the present invention, wherein (a) is a point cloud top view rendering diagram, and (b) is an image data diagram corresponding to the experimental data captured from Google Earth;
fig. 7 is a diagram of a road extraction result in an embodiment of the invention, where (a) is a top view, white is a road, and black is a background, and (b) is a diagram of a 3D road extraction result at a certain viewing angle, black is a road, and white is a background.
Detailed Description
An embodiment of the present invention will be described in detail below with reference to the accompanying drawings.
The following examples are intended to illustrate the invention but are not intended to limit the scope of the invention.
As shown in fig. 1, a voxel-based airborne LIDAR road extraction method includes the following steps:
step 1: reading original airborne LIDAR point cloud data;
in this embodiment, an original airborne LIDAR point cloud dataset P is defined as { P ═ Pi(xi,yi,zi) Where i is an index of the original airborne LIDAR point cloud data, i is 1, 2, …, n, n is the number of original airborne LIDAR point cloud data, piIs the ith original airborne LIDAR point cloud data with coordinates (x)i,yi,zi)。
Step 2: as shown in fig. 2, the raw airborne LIDAR point cloud data is regularized into a 3D volume data set:
step 2.1: removing abnormal data from original airborne LIDAR point cloud data to obtain an abnormal data removing set:
step 2.1.1: dividing original airborne LIDAR point cloud data into an mxnxu 3D grid, and mapping each original airborne LIDAR point cloud data to each grid unit, wherein the grid containing the original airborne LIDAR point cloud data is called a black grid, and the grid not containing the original airborne LIDAR point cloud data is called a white grid;
(1) an axial bounding box of the original airborne LIDAR point cloud dataset P is computed. The axial bounding box consists of its lower left corner coordinates (x)min,ymin,zmin) And coordinates of upper right corner (x)max,ymax,zmax) Determining, wherein (x)max,ymax,zmax)＝max{(xi,yi,zi) And (x)min,ymin,zmin)＝min{(xi,yi,zi) Represents the maximum and minimum values of the x, y and z coordinates in the data set P, respectively.
(2) Divide the axial bounding box into the 3D graticule mesh, wherein, graticule mesh length is got original airborne LIDAR point cloud data average point interval D:
as shown in FIG. 3, a set of points Sxy＝{(xi,yi) Is the data set P ═ Pi(xi,yi,zi) Projection on XOY plane, where C (S)xy) Is a set of points SxyConvex hull of, A (C (S)xy) Is a convex shell C (S)xy) The area of (a). From this, M × N × U3D mesh cells are obtained, M, N, U being determined by equation (2).
Wherein the content of the first and second substances,
is a round-down operator.
(3) The original airborne LIDAR point cloud data are mapped to each grid unit, grid indexes of the original airborne LIDAR point cloud data are obtained, grids containing the original airborne LIDAR point cloud data are called black grids, and grids not containing the original airborne LIDAR point cloud data are called white grids.
Mapping data points to individual grid cells:
wherein (m)i,ni,ui) Representing original airborne LIDAR point cloud data piThe index of the grid cell in which it is located.
Step 2.1.2: positioning black grids with the highest elevation and the lowest elevation in M x N stand columns in the M x N x U3D grid as candidate abnormal grid units to obtain candidate abnormal data sets;
step 2.1.3: comparing the elevation difference of each candidate abnormal grid unit in the candidate abnormal data set with the average elevation difference of the black grid units in the peripheral given neighborhood, and if the elevation difference is larger than a given threshold value TedIf the original airborne LIDAR point cloud data contained in the candidate abnormal grid unit is abnormal data, removing the abnormal data, otherwise, retaining the internal package of the candidate abnormal grid unitAnd finally obtaining an abnormal data set for removing the original airborne LIDAR point cloud data.
In the present embodiment, TedThe value of the constant (such as 3m) is determined according to the spatial distribution condition of the original airborne LIDAR point cloud data.
Step 2.2: regularizing the removed anomaly dataset into a 3D volume metadata set:
step 2.2.1: representing the 3D spatial range with an axial bounding box that excludes the anomalous data set;
remove abnormal data set denoted as Q ═ Qi'(xi',yi',zi') Where i 'is an index for removing data in the abnormal data set, i' is 1, …, t, t is the number of data in the abnormal data set, q isi'Is to remove the ith' data in the abnormal data set, and the coordinate is (x)i',yi',zi'). The 3D spatial range is represented by the axial bounding box of Q, the determination of which is referred to step 2.1.1;
step 2.2.2: calculating voxel resolution in x, y, z directions, i.e. voxel size, the voxel resolution in x, y, z directions (Δ x, Δ y, Δ z) being determined in dependence on the average point spacing of the data points in the removed outlier data set:
wherein, S'xy＝{(xi',yi')}，S'xz＝{(xi',zi')}，S'yz＝{(yi',zi')}. As shown in FIG. 4, C (S'xz) Is point set S'xzConvex shell of, C (S'yz) Is point set S'yzWherein, S'xy、S'xz、S'yzA (C (S'xz))、A(C(S'yz) Is a convex shell C (S'xz)、C(S'yz) The area of (a). The minimum value in equation (4) is because it represents less loss of accuracy between the established 3D voxel grid and the original point cloud.
Step 2.2.3: dividing the axial bounding box according to the voxel resolution in the x, y and z directions to obtain a 3D voxel grid, wherein each 3D voxel grid unit is called a voxel;
the axial bounding box can be divided into a 3D mesh of voxels, represented by a 3D array of voxels, based on the voxel resolution (deltax, deltay, deltaz). Let V be the set of voxels in the 3D voxel array:
V＝{vj(rj,cj,lj)}, (5)
where j is the voxel index, j is 1, …, m; m is the voxel number; v. ofjIs the voxel value of the jth voxel; (r)j,cj,lj) Is the coordinate (row, column and layer number) of the jth voxel in the voxel array. The number of voxels in the X direction is R, the number of voxels in the Y direction is C, and the number of voxels in the Z direction is L, wherein R, C, L is obtained by equation (6):
wherein the content of the first and second substances,
is a round-up operator.
From this, it can be derived that the voxel number m is:
m＝R*C*L (7)
step 2.2.4: and mapping the removed abnormal data set to a 3D voxel grid, and judging whether the 3D voxel grid contains removed abnormal data points or not, wherein the intensity of the contained data points is assigned to the voxel containing the data points, if a plurality of data points exist, the average value of the intensities of all the data points contained in the voxel is assigned, otherwise, the voxel is assigned with 0, and the 3D voxel data set is obtained, wherein the 0-value voxel and the non-0-value voxel respectively represent a background voxel and a target voxel, and the regularization of the removed abnormal data set is completed.
And step 3: separating a set of geobody metadata from the 3D voxel data set; in this embodiment, a voxel-based 3D filtering algorithm is used to directly separate the ground and non-ground voxel sets from the target voxel.
The algorithm firstly selects the ground seed voxel according to the principle of 'lowest elevation', and then marks the target voxel which is communicated with the ground seed voxel in 3D. After filtering, all target voxels are labeled as ground voxels (e.g., labeled "G") and non-ground voxels (e.g., labeled "NG"), respectively.
And 4, step 4: as shown in fig. 5, the laser intensity characteristics of the road reflection are used to search the road seed voxel from the ground voxel, and the target voxel which is 3D connected with the road seed voxel and has an intensity difference smaller than a set threshold, i.e. the voxel containing a data point, is extracted as the road voxel to form a road voxel metadata set, so as to complete the voxel-based airborne LIDAR road extraction, and obtain the road extraction result:
step 4.1: searching road seed voxel V from ground voxel according to laser intensity characteristic of road reflections：
The intensity characteristics of the road points are: for the same measuring area, the flight conditions are relatively close, so that the laser echo intensity can be approximately considered to be only related to the surface medium, and the aim of distinguishing roads can be achieved by establishing a corresponding relation between a group of intensities and the road medium. Establishing a distinguishing scheme for searching road seed voxels from ground voxels according to the characteristics:
step 4.1.1: projecting the ground voxel to an XY plane to obtain a two-dimensional intensity image I1；
Step 4.1.2: intensity information preprocessing: including removing anomalous intensity values and intensity value range transformations, first removing apparently erroneous anomalous intensity values (intensity values)>3000) The dynamic distribution of intensity values is then transformed to linearly stretch the distribution of intensity values to between 0 and 255. Linear stretching is the transformation of original intensity values to new intensity values using a linear relationship. Obtaining a two-dimensional intensity image I after pretreatment2；
Step 4.1.3: display I2In I2Selecting road samples of different materials, and calculating the sample mean value of the roads of different materials;
step 4.1.4: searching the voxel with the same intensity value as the sample mean value in the ground voxel set as the road seed voxel Vs。
Step 4.2: in turn from VsStarting from the target voxel in the given neighborhood scale, traversing all adjacent target voxels with depth first, if the adjacent target voxel and VsThe paths are communicated, and the intensity difference between the adjacent target voxel and the seed voxel is less than a given threshold value TiThen, the label is road voxel:
step 4.2.1: initialization: setting an initial stack storing seed voxels, and marking these seed voxels as road voxels (e.g., marked as "R");
step 4.2.2: popping an element from the top of the initial stack, obtaining unmarked target voxels in the neighborhood of 26, comparing the intensity difference between the target voxels and the seed voxels, and if the intensity difference is less than a given threshold TiIf yes, the target voxel is marked as a road voxel and stored in an initial stack;
step 4.2.3: and judging whether the initial stack is empty or not, if so, marking all target voxels in the 3D voxel grid, which are communicated with the Vs by paths, as road voxels, and otherwise, executing the step 4.2.2.
In this embodiment, TiIs constant (e.g., 2).
The embodiment also optimizes the road extraction result. Some non-road voxel data may still exist in voxel-based onboard LIDAR road extraction results due to the close proximity of intensity and elevation of certain non-road areas (e.g., parking lots, patios, etc.) to road areas. The characteristics of good connectivity of the road area, large coverage area and relatively small area of the non-road area can be utilized to remove the non-road voxels so as to optimize the road extraction result. Namely, 3D connected regions are constructed from the road voxel data set obtained in step 4, and the connected regions with the areas smaller than the threshold are deleted by setting the area threshold.
If remember
A road body metadata set obtained in step 4, wherein k is 1, …, l, flagk1(0) for voxel vkHas been (not) traversed. First, the flag is initializedk0(k is 1, …, l), then pair
And flagk0, with depth-first strategy
The 3D connected set V of which 26 neighborhoods are constructedkc＝{vk′(rk′,ck′,lk′) J, k' ═ 1, …, z, if VkcThe area of the area occupied by the projection to the XY plane is less than or equal to a set area threshold value TaThen from
Deletion in VkcAll voxels contained within and marking those voxels as traversed; otherwise, mark VkcAll voxels within have traversed.
In this example, TaThe value of the constant is determined according to the spatial distribution condition of the original point cloud.
The invention can realize software programming simulation by using MATLAB 7.11.0 platform programming on CPU Core (TM) i5-24003.10GHz, memory 4GB and Windows 7 flagship edition systems.
In the embodiment, urban area data CSite2 provided by a third working group (http:// www.itc.nl/isprsswgIII-3/filtertest /) of International Society for Photogrammetry and Remote Sensing (ISPRS) is used as experimental data to check the effectiveness and feasibility of the method. The data is obtained by an Optech ALTM onboard laser scanner, and the average scanning point density is about 1 point/m2The lengths of the east, west and north directions of the flight band are about 380m and 526m respectively. As shown in fig. 6, (a) is a point cloud top view rendering, and (b) is image data corresponding to experimental data captured from Google Earth.
Fig. 7 is a map of voxel-based onboard LIDAR road extraction results in this embodiment, (a) is a top view, white is a road, black is a background, and (b) is a map of 3D road extraction results from a certain viewing angle. From fig. 7, it is seen that all roads in the experimental data are successfully extracted, thereby verifying the effectiveness of the method proposed by the present invention.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; such modifications and substitutions do not depart from the spirit of the corresponding technical solutions and scope of the present invention as defined in the appended claims.
Claims (2)
1. A voxel-based airborne LIDAR road extraction method is characterized by comprising the following steps:
step 1: reading original airborne LIDAR point cloud data;
step 2: regularizing original airborne LIDAR point cloud data into a 3D volume metadata set;
step 2.1: removing abnormal data from original airborne LIDAR point cloud data to obtain a removed abnormal data set;
step 2.1.1: dividing the space where the original airborne LIDAR point cloud data is located into an mxnxu 3D grid, and mapping each original airborne LIDAR point cloud data to each grid unit, wherein the grid containing the original airborne LIDAR point cloud data is called a black grid, and the grid not containing the original airborne LIDAR point cloud data is called a white grid;
step 2.1.2: positioning black grids with the highest elevation and the lowest elevation in M x N stand columns in the M x N x U3D grid as candidate abnormal grid units to obtain candidate abnormal data sets;
step 2.1.3: comparing the elevation difference of each candidate abnormal grid unit in the candidate abnormal data set with the average elevation difference of the black grids in the peripheral given neighborhood, and if the elevation difference is larger than a given threshold value TedIf so, the original airborne LIDAR point cloud data contained in the candidate abnormal grid unit is taken as abnormal data and removed, otherwise, the original airborne LIDAR point cloud data contained in the candidate abnormal grid unit is retained, and finally an abnormal data removing set is obtained;
step 2.2: regularizing the removed anomaly data set into an intensity 3D volume metadata set;
and step 3: separating a set of geobody metadata from the 3D voxel data set;
and 4, step 4: searching road seed voxels from the ground voxels by using the laser intensity characteristics of road reflection, and extracting data point-containing voxels which are in 3D communication with the road seed voxels and have intensity differences smaller than a set threshold value as road voxels to obtain road extraction results;
step 4.1: searching road seed voxel V from ground voxel according to laser intensity information reflected by roads；
Step 4.2: in turn from VsStarting from the target voxel in the given neighborhood scale, i.e. the voxel containing the data point, traversing all neighboring target voxels with depth first, if the neighboring target voxels and V aresThe paths are communicated, and the intensity difference between the adjacent target voxel and the seed voxel is less than a given threshold value TiMarking as road voxel;
step 4.2.1: initialization: setting an initial stack for storing seed voxels, and marking the seed voxels as road voxels;
step 4.2.2: popping an element from the top of the initial stack, acquiring unmarked target voxels in a given neighborhood, comparing the intensity difference between the target voxels and the seed voxels, and if the intensity difference is less than a given threshold TiIf yes, the target voxel is marked as a road voxel and stored in an initial stack;
step 4.2.3: and judging whether the initial stack is empty or not, if so, marking all target voxels in the 3D voxel grid, which are communicated with the Vs by paths, as road voxels, and otherwise, executing the step 4.2.2.
2. The voxel-based airborne LIDAR road extraction method of claim 1, wherein the step 2.2 comprises in particular the steps of:
step 2.2.1: representing the 3D spatial range with an axial bounding box that excludes the anomalous data set;
step 2.2.2: calculating the voxel resolution in the x, y and z directions, namely the voxel size, and determining the voxel resolution (delta x, delta y, delta z) in the x, y and z directions according to the average point distance of the data points in the abnormal data set;
step 2.2.3: dividing the axial bounding box according to the voxel resolution in the x, y and z directions to obtain a 3D voxel grid, wherein each 3D voxel grid unit is called a voxel;
step 2.2.4: and mapping the removed abnormal data set to a 3D voxel grid, judging whether the voxel contains removed abnormal data points, if so, assigning the intensity of the contained data points to the voxel containing the data points, if so, assigning the average value of the intensities of all the data points contained in the voxel, otherwise, assigning 0 to the voxel to obtain a 3D voxel data set, wherein the 0-value voxel and the non-0-value voxel respectively represent a background voxel and a target voxel.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710044460.8A CN106709473B (en) | 2017-01-19 | 2017-01-19 | Voxel-based airborne LIDAR road extraction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710044460.8A CN106709473B (en) | 2017-01-19 | 2017-01-19 | Voxel-based airborne LIDAR road extraction method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106709473A CN106709473A (en) | 2017-05-24 |
| CN106709473B true CN106709473B (en) | 2020-03-24 |
Family
ID=58909299
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710044460.8A Active CN106709473B (en) | 2017-01-19 | 2017-01-19 | Voxel-based airborne LIDAR road extraction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106709473B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109407115B (en) * | 2018-12-25 | 2022-12-27 | 中山大学 | Laser radar-based pavement extraction system and extraction method thereof |
| CN110516653A (en) * | 2019-09-03 | 2019-11-29 | 武汉天擎空间信息技术有限公司 | A kind of method for extracting roads based on multispectral airborne laser radar point cloud data |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101625414A (en) * | 2009-08-11 | 2010-01-13 | 中国气象局北京城市气象研究所 | Method and device for integrating radar and geographic information data, and weather monitoring system |
| CN105844602A (en) * | 2016-04-01 | 2016-08-10 | 辽宁工程技术大学 | Airborne LIDAR point cloud 3D filtering method based on volume elements |
| CN106324581A (en) * | 2016-08-25 | 2017-01-11 | 辽宁工程技术大学 | Body-element-based airborne LIDAR building detection method |
-
2017
- 2017-01-19 CN CN201710044460.8A patent/CN106709473B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101625414A (en) * | 2009-08-11 | 2010-01-13 | 中国气象局北京城市气象研究所 | Method and device for integrating radar and geographic information data, and weather monitoring system |
| CN105844602A (en) * | 2016-04-01 | 2016-08-10 | 辽宁工程技术大学 | Airborne LIDAR point cloud 3D filtering method based on volume elements |
| CN106324581A (en) * | 2016-08-25 | 2017-01-11 | 辽宁工程技术大学 | Body-element-based airborne LIDAR building detection method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106709473A (en) | 2017-05-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108109139B (en) | Airborne LIDAR three-dimensional building detection method based on gray voxel model | |
| CN110969624A (en) | Laser radar three-dimensional point cloud segmentation method | |
| CN113706480B (en) | Point cloud 3D target detection method based on key point multi-scale feature fusion | |
| CN112613378B (en) | 3D target detection method, system, medium and terminal | |
| CN111325138B (en) | Road boundary real-time detection method based on point cloud local concave-convex characteristics | |
| CN112099046B (en) | Airborne LIDAR three-dimensional plane detection method based on multi-value voxel model | |
| Shirowzhan et al. | Building classification from lidar data for spatio-temporal assessment of 3D urban developments | |
| Zeybek | Classification of UAV point clouds by random forest machine learning algorithm | |
| CN112132108A (en) | Ground point cloud data extraction method, device, equipment and storage medium | |
| CN108074232B (en) | Voxel segmentation-based airborne LIDAR building detection method | |
| Yadav et al. | Identification of trees and their trunks from mobile laser scanning data of roadway scenes | |
| Hu et al. | Research on a single-tree point cloud segmentation method based on UAV tilt photography and deep learning algorithm | |
| CN115063555A (en) | Method for extracting vehicle-mounted LiDAR point cloud street tree growing in Gaussian distribution area | |
| CN113205548A (en) | Automatic registration method and system for forest unmanned aerial vehicle and foundation point cloud | |
| CN106709473B (en) | Voxel-based airborne LIDAR road extraction method | |
| Yadav | A multi-constraint combined method for road extraction from airborne laser scanning data | |
| CN112017219A (en) | Laser point cloud registration method | |
| CN114663485A (en) | Processing method and system for power transmission line image and point cloud data | |
| CN107993242B (en) | Method for extracting boundary of missing area based on airborne LiDAR point cloud data | |
| CN111861946B (en) | Adaptive multi-scale vehicle-mounted laser radar dense point cloud data filtering method | |
| CN112581511A (en) | Three-dimensional reconstruction method and system based on approximate vertical scanning point cloud rapid registration | |
| CN110174115B (en) | Method and device for automatically generating high-precision positioning map based on perception data | |
| CN116994029A (en) | Fusion classification method and system for multi-source data | |
| CN117253205A (en) | Road surface point cloud rapid extraction method based on mobile measurement system | |
| CN115861566A (en) | LiDAR point cloud building integrity enhancement method based on intensity voxel model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |